@@painless4785 It's a community meme. According to Wirtual, it means clutching a close scenario in his favor, according to all other people it means failing in the last crucial moment.
@@SnowyRedstone well gpu drivers often come with security updates and improvements to the code making the gpu more reliable and sometimes faster. When a big game releases, nvidia at least idk about amd or intel, also release drivers which have improvements to that specific game making it run better on their gpu's. So while updating a driver every 2 weeks isnt necessary strictly, updating it every once in a while is important even if only for the security updates.
22:4523:00 These are both times that Trackmana gave the no internet warning on the lower right of the screen. The 23:00 time is right when Wirtual crossed the finish, leading to the lagging theory being incredibly plausible.
I thought the whole point of Trackmania was that the replays are deterministic. The server shouldn't even care what time it sees on the packages; the inputs should be timestamped in the *round*. So if a lagging packet arrives late, it should be correcting the ghosts on other people's screens, sure, but on the simulation where it's counting the victor, it should stop running the game until it gets affirmative confirmation that there's been *no" input up to a point. Simple lockstep with prediction.
@@larion2336 Faked records are driven offline so there's no server verification happening, or at least not until the time of submission. This is how people are able to drive with slo-mo etc. while not having their times inflated - because there's no server running a timer in the background. Tournaments are driven on live servers to maintain competitive integrity. It's a completely different system. Faking a record in a live tournament is anything but effortless, but removing the server time and relying entirely on client information would make it easy.
@@SpiceChThey implemented anti cheats into TMNF which included playing in slomo, because runs were done offline, you're saying that they've lost the ability to do this?
It’s insane that whats supposed to be a 100% deterministic game can at the same time have bad internet make you slower lol Love the content Mr. Wirtual keep it up :)
Well, if the inputs were taken and it gets replayed does it give out the same time 🤔? From my point of view it shouldn't as the time occured because of package loss and repeating the inputs with stable connection would result in the "real" time. So the time wouldn't be validated and the run therefore can't be counted at all.
@@SinclairWest its not tmnf. doing that for thousands of players every cup would take some time and make servers even laggier. akso switching in the last rounds would be difficult but maybe the only way
@@privatjetconnaisseur could just send the replays to each players in the match and verify by majority. it can't be too much work given that it's doing the same thing when showing ghosts.
@@privatjetconnaisseuryou are correct about the replaying but the verification can be done in other ways. For example collecting some anticheat info to verify no cheats are present with some verification of the aplication. The data can be send via encripted conection in which only TM issued certificates are accepted and each user will have thair own certificate. And send times and keypress info in the comunication. The encription can be handled on TM side on dedicated network infrustructure to offload any aditional load to servers and the serwers would only need to verivy that the retrieved data is in correct format and that all data arrived. And after this share the data with other users. So solutions are there for verification whiteout doing any hard work on servers. Ofcorse some other info would need to be collected but the info for times would not be delayed due to packetloss (for mitigating issues with packetloss TCP can be used to ensure all data came trough)
As a (former, by now I guess?) game developer, I can tell you: netcode is hell upon earth and there's a million things that could explain this. My guess is that the server overrides the time, if the time between packets is too large, even if it could validate them
Netcode is only relevant if action on one player affects other players action. Trackmania is effectively a single-player game, so other than visual glitch there is no excuse on validating inputs in real-time.
Yep, netcode is hell. Though for things like the cup-of-the-day, I feel like the client could record the inputs, send to server at end of a round, and server could verify that the reported time matches a simulated run and client is not mis-reporting. The fact that the rounds being so short, the physics consistent (hah!), and the gameplay-loop being so discontinuous (clearly segmented into rounds with several seconds between each), means it should not be too difficult to do seamlessly. Combine that with how it only matters for the top2 players (or rather, the 2 players surrounding the cut-off mark. Or a few more, for the few cases where ppl are within thousands around it - either way, only need to verify runs where server and client time differ *AND* the cut-off segment differs correspondingly) and there aren't even many simulated runs that have to be run. So should not be infeasible. That way there's no netcode required! Only "post-processing" with plenty of latency allowed! Hell, if it does eat too sizeable a chunk of time, it could even be further optimized thanks to the checkpoint system. Only have to simulate the splits where time, position, angle and speed differ between client and server, cutting down the number of inputs (and simulated time) greatly, potentially. Depending on how they do things, the server might even be rather idle during the actual round, meaning some of that verification (between checkpoints) could be done _during_ the round (though the more you divide such a system, while also making it immediate, the closer you are racing towards doing netcode again :p).
I'm surprized there's a difference in replay considering deterministic physics. My guess that client and server keyboard input is different for some reason. It's surely possible in case of real time replay, I guess server tries to interpolate input in case of late/missing packets from the client. So my theory is that near finish Wirtual's input packets were delayed/dropped and server had to show not quite accurate position in the replay for everyone else. But this doesn't explain the time difference between client and server, hard to tell.
@@hanifarroisimukhlis5989 you're over simplifying. In a perfect world the game could just be running locally and each client just sends their times to the server. But people are people and this is very exploitable. Cheaters would just intercept their packets and send whatever time they wanted A realistic solution might be to capture the player inputs when there are less then 12 players left or something, then validate them by running them on the server and gettting accurate timings from that. Even this is susepible to cheating, but it would be better than just relying on internet connetions I don't think it will be realistic to use this method for every single round. It would just add too much overhead and wouldn't achieve much
There are couple common ways of doing multiplayer: -*client is right;* server only cares about position and time. Times are as registered on your screen; very prone to cheating -*server is right;* server cares about inputs you make Resistant to cheating and lasting desync; prone to rubberbanding as server does not register inputs and backs you or even does not let you move It seems that Trackmania uses a mix of those, where client is right about position, but server is right about timing, which is prone to something as little as ping fluctuation in such competitive setting. What I think *SHOULD* happen, is that your game sends precise timings of your inputs not "now left" but "left at 20.45" so server can then save a demo/replay that it then can simulate as a TAS since game is deterministic, at least for those spots that matter (all those that KO and top few), and use that as true time.
as far as I understand, replays are saved exactly in that format (key press/release + milliseconds from start), so it should be simple enough to register it like that. The only problem is, this probably uses A LOT of server CPU time, so it would be reasonable for Nadeo to implement a hybrid approach where this can be turned on only for key cases (like div1 COTD, and maybe only the final round for other divs)
One of the reasons I like playing TM, and not other multiplayer games, is that I don't have to worry about ping, or any other problems that other online games have. If that's not the case, then TM loses a substantial amount of value for me.
Bad connection has always led to problems playing online. It’s not new in tm2020. You would have an invalid run that wouldn’t even be finishable in tmnf.
@@KaptainKerlmy first run on Detroit 2 online... 3 hours in my internet got wonky for a second and then "this time will be ignored" popped up... I was not amused
"When will NVIDIA make a graphics card that doesn't need an update every week" For context, I'm a programmer with a graphical computing specialization. The frequent updates aren't needed per se, but a nice to have. 1- the most impactful reason, NVIDIA has added a metric ton of software features, both for users (like DLSS, reflex, rtx voice, video up scaling, etc) as well as for developers (optix, mesh shading, VRS, etc) that are essentially just apps that run in your GPU. The same way Adobe premier or some other piece of software you may use update, so do these apps... Everything in computing is becoming more and more software defined as we go on. For stuff less relevant, but that still have an impact... 2- GPU's are just... Weird to code for. The overarching frameworks are just no where near as mature as CPU's, and the massive parallelization introduces many quirks. If you code something in C, you can be pretty sure it's going to work as intended. With a GPU not so much, weird behaviors are essentially unavoidable. Vulkan only came out in 2016 for example. Stuff is being patched all the time. 3- On the same vein, CPU code is basically stagnant in terms of patterns. For GPU devs are still constantly coming up sneaky ways to optimize the code, exacerbating the first problem. 4- bringing the first two points together, you'll notice most updates come tied to a big game release. These make sure the game words properly as the devs work directly with NVIDIA.
At 25:14 if you slow the video down and go frame by frame, you can see that right before Wirtual crosses the finish line (ahead of Yannex) he suddenly stops midair, allowing Yannex to overtake. He then suddenly jumps to in front of Yannex, but by that time, since Yannex had already passed the finish line, Wirtual had technically crossed second. What Yannex saw is what the server saw, because Yannex is only viewing through the lens of the server, so on the server side, it registered the time Wirtual crossed the finish line on the server. It's important to note that this kind of thing is generally done to eliminate cheaters. If the client is allowed to transmit their time to the server, then there is the opportunity for a player to cheat their time. You might say "Well, but the server has to verify that time, so you couldn't get a cheated time on the leaderboard anyway." But, that only applies to replays. In a live setting like COTD, a cheater could theoretically turn off their internet mid-match, hack themselves to the end, then send that data up to the server and claim "Oh yeah, I just drove that. Must be bad packets." It's unfortunate that it happened right at the end of Wirtual's race, but the only solution here is really for Wirt to get better internet DinkDonk
@@Psychomaniac14 That's true. Ultimately, someone is going to have better ping just by the nature of where the server is. Call it pay to win, I guess :P
They just need to actually use the inputs/replay in case of packet loss. Probably a time out of 5 seconds or so, such that a bad actor can't easily exploit it. It wouldn't be the client telling the server what their time is, it would be the client telling the server what they did and when, such that the server can check the time.
I think the server's timer keeps running until there's a response from the client, and then the server verifies the client's time. But if the discrepancy between the client's time and the server's time is too big, then the server's time is prioritized (possibly to prevent cheating). But this also means that unstable internet can lead to slower times if there's packet losses happening at the time of finishing. I don't really know that there's much Nadeo can do to fix this unless they ditch the server timer and rely entirely on client times - but that could make it a lot easier for people to cheat times. Manual review might be an option for certain parts of the pro scene, but track of the day has far too many participants for such a thing to be remotely feasible.
@@GriziDaWiz it wasnt that wirtual was a twisted person, trackmania just relies on server to receive time instead of the client, theres a reason for this(when someone cheats runs offline in slowmo it is because they arent connected to the server meaning their time isnt verified) however its still trackmania's fault, they should just make the client not tell the server the time but rather what the player did insteas during competitions, anyways there are so many people spectating these guys in competitive matches so theres no way they try to cheat and even if they do and make it to the finals someone will notice it for sure 100% basically making it practically impossible to cheat while not loosing time due to lag
This was the first cup of the day I caught on stream, and on the last round I hard gambl3d all my points that he would win, and that was the fastest I went from triumph to defeat ever I was truly SCAMMED
Thank you for explaining Elensar, i hope you guys are able to track down the issue and prevent these laggs from happening ! Also nice shirt btw, i love Architects and listened to their absolute banger songs just today 🤘🏼
it also happened on an other cotd where he got 59th last second before the timer ran out but it didnt register and he got 65th (which was from his time before his last finish)
If I had a nickel for every time something controversial happened with a COTD directly involving Wirtual, I'd have 2 nickels Which isn't a lot, but it's weird that it's happened twice
@@hotshot0079Wirtual was accused of cheating his author time on his Cotd map Midori due to using external programs (analogue keyboard) to automatically steer 34% when he pressed a certain button. (In the game, you can only bind 20, 40, 60, 80, and 100)
@Wirtual, regarding the drivers, let me know whether this makes sense or if not, where I should explain further. Nvidia has something called "GeForce Game Ready Drivers" which means they work with developers to fine tune their drivers to work optimally across a wide range of games and hardware. As new hardware and new games get released there are updates to the drivers to ensure that the drivers work for the new hardware/games to the standard Nvidia expects. Of course, these micro-adjustments can lead to small bugs which will also get fixed in further updates as they become apparent. In essence, this means that not updating your drivers can lead to sub-par performance with the newest hardware/games and you might have small bugs that don't get resolved, leading to e.g. too much resources being used (which can lead to system degradation if you keep your computer running for long periods of time) or subsystem crashes (that too can lead to system degradation). On the flipside, it also means that if your system works fine, it'll work fine without updating unless you start playing new games or install new hardware.
TM servers do indeed check the run, that is exactly what was done at the finish line, during the check the time was updated to 35.518, which is the 5th checkpoint time, then to the final 40.850 and hence the correction. Data is often lost over the internet, it is normal. Games use input bulking or report-correcting to counter this issue. Something absolutely stoopid happened just before the finish line. Maybe an input discrepancy could not be corrected after the finish was crossed due to a faulty algorithm. But Nadeo has robbed the man nonetheless.
You're right that not updating drivers shouldn't cause performance degradation which means there's likely something else going on which happened to have gotten resolved when you updated your drivers last. I'd recommend going into task manager when TM is super laggy and investigating if there's anything hogging system resources when it doesn't need to be.
He also has his pc turned on for weeks if not months at a time, so something getting fixed when he updates drivers might actually just be because he actually restarts his pc when he does that
@@piknos3381 It *has* been a problem before, though? Even in this video he references the last time that he was lagging and people told him to update drivers, which he did and his problems were fixed
@@T0NI_ exactly, the reboot required while updating gpu drivers was most likely causing the positive impact on the performance. as long he stays a notorious "never shutdown" guy there is nothing to troubleshoot about this issue - as it could be every single service interaction running on his computer
@@lux2248 He also randomingly mentioning between all of that, that CODT servers are laggy sometimes? Like bruv are you actually frame dropping or is it just bad connection, cause lowering draw distance for objects certainly won't help you that much, if the connection is just shitty.
12:30 Banana Question. Then you have a mini intro line like "James forgot to lock up the monkeys..." Also "So I got an F in my math test because James bought 52 Bananas"
Wirtual the man who, never shuts off his PC, never updates drivers, plugins, anything, and doesn't close TM (and for some reason TM runs hot, even on my high end PC) when something goes wrong with his game. Listen it *maybe* had nothing to do with that but still man you're just stacking debuffs at this point
he could just type 3 words into google 'why restart PC' and would get result over result telling him that leaving his PC running for days causes performance degradation. But comparing all possible PC configurations and resulting quirks to consoles seems to be easier.
Yeah, like obviously it's sad that this happened in the very final round on the very final checkpoint with just enough time to fall behind, but there's also something poetically ironic about the fact that a good chunk of this video is spent complaining about how graphics driver updates are not worth doing
It took me and hour and stress, but I finally found the song between 5:05-7:57. It’s called infestation I think from Terraria, and the artist is DM DOKURO. I have many songs on my playlist that Wirtual plays, and it’s pretty cool.
Yeah it is funny how he undertands nearly nothing about tech but refuses to accept that he needs to reboot his PC, to fix memory leaks for example. It is not like he has hints of that happening because his performance gets worse over time, but can't argue with somebody that thinks they are right without knowing anything.
@@user-to7ds6sc3p you have no clue what you're talking about. Not only are memory leaks not a given, but also they don't impact anything unless you're actually running out of RAM. Modern operating systems are designed to run 24/7 for months no problem.
The reason drives need constant updates is because Windows will have updates and games will have updates that the current driver isn't optimized for so they have to update it so that your computer won't be slow when playing said games
okay but how is this not a bug being reported/talked about en masse? this feels like an incredibly dangerous thing to have in a game that counts down to the tenth of a second.
I think it's the game time display in the game who isn't exact. When you display images on the screen there is a delay. The timer you get it's from the beginning of the computation, it ask to be render. Then it's render. The display time can be many frames late. A CPU have many task in parallel, get the time isn't the most important thing and that why it can jump some frame
wait does the server not double check if it got the right package? the server should wait a few seconds at least for the player to send time packages. a few seconds between rounds after finishing isn't gonna tick players off
Imagine a new graphics processor is like a new map. You might be able to drive it pretty well when it's first made, but the more you drive it, the more you realize where you are losing performance and how to make it better. And you find ways to keep from clipping edges and losing speed from minor errors. That's what display drivers are like. They are constantly improving as they learn more and apply it. Yes, there will be times where they missed something and it causes an issue with certain games or programs, but they usually work to fix those quickly when they realize they exist. So overall, the drivers get better and better over time to where the drivers towards the end of a card's life cycle are going to give very noticeable improvements
Have you ever considered a video or series where the premise is "If you are at world rank X, you should make sure you understand these techniques to improve by (50, 100, 1000, 10000, whatever) ranks"? I think that would be useful for any newcomers to know what sort of knowledge is common at different rank tiers, so that they can hear about these high level techniques, but understand that they are unlikely to benefit from it until they're closer to whatever rank.
12:36 i work in a grocery store, our online orders get multiple people a week that order 60 or more bananas, on Monday someone ordered like 150ish lmao
i would have assumed trackmania would be doing a live rollback similar to how hit detection works in certain games, the idea being, since the game runs "deterministically" the game must be using a loop that registers inputs at a certain intervals. those inputs, "time stamped" with the interval they were recorded in would then be sent to the server, the server simulating the vehicle physics and collsions independantly, would, on packet loss, store the last valid position, provide a prediction for others to see, and when the missing packets come back in, make sure they're in the right sequence, before picking up the simulation from that last valid position simulating till it gets to the latest timestamp or another missing packet. if that were the case, the times shown when going passed checkpoints should always stay consistant between the players game and the server, since the server is simulating it the same as the actual game on a players computer, trackmania being "deterministic" any calculations made for both the player or server would be evaluated the same and get the same result, providing the same final time. why isn't this how it works?
there is no need to update drivers, however updates can fix bugs (you might not be experiencing) or security issues. for example right now AMD has a bug in their driver that allows privilege escalations from normal user to admin. in that case you should immediately update
The only one to blame here is Wirtual for not getting business contract with his ISP into his house and taking legal action against the ISP for loss of revenue. When you pay for consumer grade internet you get consumer grade service and consumer grade routing. His streams have been lagging ever since he moved into that house and he has done nothing to change the situation, just the other day on stream he said that he hasn't even tried getting a new router, ISP supplied routers can get overheated from streaming due to them being complete dog . Imagine being Nadeo and paying Wirtual to cast TMGL when his stream crashes and lags out on the regular. If any of the Norway based editors sees this, please have an IRL chat with him and do everything for him because Wirtual is completely incompetent when it comes to technology. Personally if I made the money he is making I would get the business contract with the ISP, a brand new non-refurbished trash router, and change all the ethernet cables. He doesn't live in a third world country, I had better internet stability playing Unreal tournament in the early 2000s.
Bro chill, Norway might not be a third world country but I can tell you internet is expensive when you have to run cables through mountains and fjords. Internet and mobile services are a lot more expensive in Norway than Sweden and Denmark.
The reason the game trusts the server time over the client's time is that if it were ro do the opposite, it would be opening the door to all manner of cheat methods. Because the only real way for the game to reliably check if a client's time is legit would be by referencing the server time and seeing how they line up. Were the game to trust the client over the server, players could just figure out how to send in a fake time to the game without any way to check for authenticity, you could theoretically drive a 10 minute lap but tell the game it took you a minute, and the game would have no choice but to believe you. For live events like CotD, what they might be able to do would be to save the inputs of each lap and have those checked when client and server times dont line up, since cheats like slowmo arent exactly viable during a live competition. That might be a potential solution, though its possible that too could open the door for other exploits. tldr: videogame code is weird and complicated, and anything online gets even worse, while anything offline is free real estate for cheaters.
I'm on my way to Croatia for vacation and I'm currently watching this video with my girlfriend who just fell asleep to it. Wirtual really has the superpower of sleep
The Constant graphics driver updates are caused by constant hardware updates and constant software updates. For a PC, a driver needs to support almost all the hardware out there, where many companies utilize the chipset and report bugs and issues with the latest and greatest hardware as well as with the old chipsets. Part of this may be due to the custom hardware implementations. Second, Operating systems are always being updated. I have a very old computer yet graphics performance degraded greatly after a certain version of windows came out. In my case my video card became end of life and windows stopped supporting it even though I had the drivers for it. Nintendo, had fixed hardware and fixed OS so the game developers don't need to patch for OS and hardware changes.
a big part in the "gb games didn't need updates" and "graphic cards weekly updates" is that developers have become complacent due to resources being so available. The amount of RAM(or equivalent) for gameboys was so small, same limitation for CPU etc. We don't have to think so much about what we do, since computers do it for us. Additionally, improvements are harder to come by due to technology changes. Changing a picture on a 4k monitor is a lot more resource-intensive than doing so on a GB screen. And then the point that they just _couldn't_ update GB cartrigdes without calling them back, or some other dedicated infrastructure. Also: pokemon (for example) is far from bug-free ;)
Sure his stubbornness about updating is little bit funny but in this case its totally reasonable to question why is the driver that 2 weeks ago was good thing that fixed stuff suddenly bad when the game has no major updates in between. In his case the lag is most likely something streaming related rather than nvidia driver related, if I would have to guess in last 50 nvidia driver updates maybe 1 has had something directly related to trackmania and that would probably be something related to raytracing :D
Like with Pokemon G/S, when clever people are given design limitations they'll find creative ways of making things work. Nowadays, the 'crunch culture' also doesn't help. Endless resources make people inefficient imo; an Exec probably said: "why waste time creatively cutting the size of Call of Duty, people will spend hours downloading it anyway, and the cash shop needs to be worked on".
In regards to driver updates, is not one of the things you can do is to update and apply a recommended set for each game you are playing when it is running? Though i have no idea on the weekly general updates.
Wirtual should try nvidia studio drivers where the updates are less frequent instead of game ready drivers which updates every time a new game comes out
Maybe they should use TCP for syncing lap times but UDP/websockets or whatever they use for ghosts. That way they will always know if a packet is lost wrt critical data, this probably happens a lot to the hundreths and thousands secons in a match. Just went unnoticed
I don't know if Wirt reads these, but here goes: The reason why you have to update software is that it is very hard/impossible to create bug-free code. Essentially, everyone writing code is very bad at it, and so either you spend an insane amount of effort and money writing the most perfect software ever, or you just say "screw ot we're doing it live". So at some point, someone goes "huh weird, the driver crashes if I try to draw two green triangles in a row", tells nVidia, and they issue a fix in the form of a software update. I'd be happy to tell you more if you want :)
You would think that since Trackmania is a deterministic game, the client would be sending input diffs over to the server while the client runs the game locally. This means that the client and the server play the same game, and the server is just there to check that the inputs actually lead to the car crossing the finish line at the correct time. But from this it looks like the server counts the player as crossing the finish line whenever the client reports such. This would mean that a player's internet latency, usually measured in thousandths of a second, are included in the run time in the best case scenario. We have had cup-of-the-days where players finished within a thousandth of a second of each other, and the name chose the winner. How many of those had been a clear win for one player or another? How many where a couple of dropped packets resulted in a player going out? As far as I'm concerned, Trackmania is not that competitive of a game anymore, at least in live matches. Having extremely fast and reliable internet as a requirement to be competivitve in a skill-based game is nothing new, but I thought Trackmania wasn't like that. It has the ability to not be that way. What I imagine is that at the beginning of the match, there are 64 people playing Trackmania locally on their own clients. The inputs get sent to the server for replay validation. The local client is authoritative on what the inputs were, the server is authoritative on what the finish time is. Client checkpoint times get sent out to everyone with the logic being that it's close enough. This depends on the servers being able to compute the physics at 10x-100x realtime, which seems reasonable.
how many track of the days have been the mapper's first map? there was this one and the one by sneakyy (desert one with that one plastic lizard on the corner & the end cut) are there any others?
I hope someone explained about the game boy and pc being two completely different things. A console, a game only developed for said console, back when alpha and beta was only tested in house before shipping at a full release state v1.0 would never need a patch as it would only ever be played on that device. Where as PC'S run different specs, different OS versions xp, vista, 7, 8 - 8.1, 10 and 11. Plus different states of windows updates on said OS. Not to mention game porting to PC from console, too many variables, game studios try to address those issues in the dx dll's. Then video card manufacturers have to try and take those considerations also and ensure that all new drivers work in compatibility with every old driver which takes in all those other considerations also 😮
The issue aren't even his not up to date drivers. Performance doesn't get worse from that and I myself only check for graphics drivers once a month or when I see a potential driver issue in a game. The problem is him not restarting his PC. Memory leaks, plus other things I don't have a professional name to through around for, making ressources unavailable causing performance to degrade. As you said consoles have software made for that hardware and that hardware only, that causes those bugs to be way less common, and nobody leaves there console running for weeks.
@@user-to7ds6sc3p all good was just explaining to the interwebs hoping he would even check his comments, to give him somewhat of an understanding of the complexity of drivers. Not saying he needs to update, just saying why they exist the way they do. And totally agree with what you are saying and understand what you mean.
I think usualy this is what happens 24:17 because there isn't time corrections each time. But I think when someone is too laggy the server takes over and penalizes the time
Updating your drivers isn’t necessary, but extremely recommended. If you don’t update you leave yourself at a competitive disadvantage because your software is no longer optimized. Updating drivers is just maximizing the capabilities of your gpu. The reason we have so many driver updates is the same reason new, better gpus come out. Research and new technologies allow developers to provide us with the benefits of such, and we get more efficient gpus as time goes on without having to buy the latest one. It’s why my RX 580 4gb (worse than a 1050) can play R6 or the new cod titles or apex and I don’t need a new gpu. Game shaders help of course, that’s the games’ way of optimizing to your gpu. But as with gpu driver updates, games get new shaders frequently and help you run the title more efficiently. The combination of both helps keep your hardware not only relevant, but also gives you better performance.
rocket league has a similar issue with "ghost touches" sometimes the server dont detect a cosission with the ball and your car and the ball will visuell glitch trough you
Your layman understanding of drivers is funny. There is one game console from each major brand. There are infinite hardware combos on PC. And many orders of magnitude more software (that must be compatible with each other as they run at the same time and interact). They don't just work together. They have to be told how.
This bug happened to me when I was playing yesterday. I don't think I'll be playing this game anymore. I improved my time by one second in the last minute, but it registered after more than a second, so I ended up in division 17. By the way, there's no chance of sending packets that long because of the internet. Another thing is that the position of the player would probably have to change as well. This would mean that if the packets are indeed sent this late, players in spectator mode would also see Wirtual behind. So I personally think the problem is in Trackmania, but I have no insight into the game and I could be wrong, the problem could be somewhere else.
the problem was not that packet was late, but rather that it was lost making him stop in place before teleporting behind the finish line from spectator, opponent, and timing perspective. Now it seems that either they use UDP or they don't use any late information given by TCP caring only about position that is "now", but whatever transmit method they use they also should use simulation on their end. In essence if a packet is lost, server should keep moving the player as if no change in input was made, but when the lost packet with that change in input and a timestamp is resend the simulation goes back in time to apply for this input. If done only to verify after the fact your time would be as you seen, then go up a second, and then go back a second on final result screen. But if done live you would never notice that you lost a packet.
I also very rarely restart or update drivers, but at least I understand the problems I'm giving myself... I've just decided that the annoyances from restarting are more significant to me than the ones from not updating. If something more significant becomes a factor then I reconsider updating (like recently, there was a bug in the driver for my 3060 card that was causing increasing problems in games that got fixed).
There's an ex dev in here saying net code is hell. Would I be correct in saying that they (Nadeo) have their current model of the server time being the determining factor in who wins as an anti cheat measure? Just speculating as Trackmania as a franchise is know for having players alter the speed of their games locally. So in theory if they were to change the current model. They would take the local game time from the system as the determining time. If I was to be a cheater there's a whole plethora of alterations I could do to mess with the timer or the speed of the game. It makes for a very interesting case study either way
well nadeo has weird system in place determing who wins at ties: based on the name at alphabetical order and apparently later they have updated it to based on user id which is randomly generated.
From what I can tell, as a computer engineering student, the graphics driver updates are mostly for optimizing to new popular games, with some bug patches, stability improvements, and the rare feature addition. It's much more worth your time to stay updated in the windows settings because those bring security updates against the latest hacking techniques. If you're also playing games that were released relatively recently, then graphics driver updates are very useful. Of course, you should keep updated anyways. Just imagine the computer scientists on their knees begging you to let them help you.
The reason nvidia keeps pushing updates is because new games continue to utilize new technologies that may run faster with updated driver for older hardware. Now you might say, well still why would they need to updated it THIS often then? Because change in drivers may cause older software for unkown reasons not to work as expected or straight up crash. Nvidia receives customer mail complaining about that problem causing them to release a bugfix. Every bugfix (or multiple) is adressed as new update for the driver so that with the newest driver the likelyhood of it working as optimal as possible for the widest varity of software is the highest. Sometimes bugfixes create even new problems that first go unnoticed and need to be fixed later with a new bugfix for the bugfix... and so on... Also... the driver updates are basically free advertisement place = extra monetarization
As many here suggested it would be best if the servers took the inputs and replayed them quickly to determine a time, however this would take immense resources. A better solution would be to compare the time recorded on the client to the time recorded on the server and if they diverge then the time from the input should be calculated. But yeah, overall this is almost a beginners mistake as racing games shouldn't depend on internet speeds or package loss at all. Such a weird oversight and I am wondering how this issue got this far without being fixed. I mean 0.150 seconds is a HUGE margin in a game like this, that's sometimes the difference between WR and Top 1000
Gpu drivers get updated regularly because of bugfixes and performance updates in (new) games. Aswell as issues with new monitors or new software updates with obs for example. Playing trackmania or rocketleague or tetris on old drivers isnt worse than on newer ones. But for best stability and newer games/software its always recommended to update. Windows updates are useless. Their "security updates" break more than they fix or prevent.
Well i updated my drivers for my 6900xt and the system started hanging and sometimes freezing when just launching a 3d accelerated app or game. had to freaking roll back to march drivers because those actually worked. but then, that has some visual glitches with ratchet & clank that were fixed in newer releases . good stuff
I don't know how this would be fixed. If it takes client time it's too easy to cheat. It has to take the server time, and then there is always a slight inconsistency. Try playing any modern shooter game, after a while you start to notice that there is always some lag or lag compensation going on in every single shootout, sometimes in your advantage, sometimes disadvantage.
Congrats on the first ever 2nd place cup of the day win. Truly a groundbreaking content creator.
He was technically first, so would we say he came second or first
Cause from yannex’s pov, wirtual won
@@zhanglihussey8222 I mean he did say "2nd place cup of the day win" (2nd place, win)
You must be fun in parties
Its actually the first ever 1st place cup of the day loss
When Wirtual forgets to pull a wirtual, the game does it for him
underrated comment
No matter what side of the Wirtual line you stand we can all agree that this time he definitely pulled a Wirtual
@@painless4785 choking under pressure which the community memed into being his signature move
@@painless4785 It's a community meme. According to Wirtual, it means clutching a close scenario in his favor, according to all other people it means failing in the last crucial moment.
For close races, the server really should upload the replay file from contestants and quickly replay them on server - kinda like a photo-finish thing.
that is very easily exploitable for cheats
@@AlbertoTuber It can validate the inputs too, so I think it's not a problem.
And who shall pay for the extra traffic and computational costs of doing that?
@@un1kum42 erm nando or ubisoft? at the end of the day they are the devs and the ones responsible of maintaining it
@@un1kum421 race a day wont cost much and if it makes tournament times reliable then the pros outweigh the costs, the cost will be
what i love most about watching wirtual lately is he reminds me to update my graphics card
exactly
I mean, I completely agree with him. It's working then an update come out and it is no longer working??
Maybe this is why he lost, the new driver :D
@@SnowyRedstone well gpu drivers often come with security updates and improvements to the code making the gpu more reliable and sometimes faster. When a big game releases, nvidia at least idk about amd or intel, also release drivers which have improvements to that specific game making it run better on their gpu's. So while updating a driver every 2 weeks isnt necessary strictly, updating it every once in a while is important even if only for the security updates.
I was wondering why the map looked like it'd gotten schrecked
22:45 23:00
These are both times that Trackmana gave the no internet warning on the lower right of the screen. The 23:00 time is right when Wirtual crossed the finish, leading to the lagging theory being incredibly plausible.
nice find wow, never would've noticed
I thought the whole point of Trackmania was that the replays are deterministic. The server shouldn't even care what time it sees on the packages; the inputs should be timestamped in the *round*. So if a lagging packet arrives late, it should be correcting the ghosts on other people's screens, sure, but on the simulation where it's counting the victor, it should stop running the game until it gets affirmative confirmation that there's been *no" input up to a point. Simple lockstep with prediction.
if you were to implement that, you would be very easily expoitable by cheats. that is likey the reason why it is how it is right now
@@zockersfreund It's already effortless to fake records in trackmania, so that makes no sense.
@@larion2336Don't think you understand what "effortless" means. The right word you're looking for is "practically impossible".
@@larion2336 Faked records are driven offline so there's no server verification happening, or at least not until the time of submission. This is how people are able to drive with slo-mo etc. while not having their times inflated - because there's no server running a timer in the background. Tournaments are driven on live servers to maintain competitive integrity. It's a completely different system. Faking a record in a live tournament is anything but effortless, but removing the server time and relying entirely on client information would make it easy.
@@SpiceChThey implemented anti cheats into TMNF which included playing in slomo, because runs were done offline, you're saying that they've lost the ability to do this?
It’s insane that whats supposed to be a 100% deterministic game can at the same time have bad internet make you slower lol
Love the content Mr. Wirtual keep it up :)
Well, if the inputs were taken and it gets replayed does it give out the same time 🤔? From my point of view it shouldn't as the time occured because of package loss and repeating the inputs with stable connection would result in the "real" time. So the time wouldn't be validated and the run therefore can't be counted at all.
@@SinclairWest its not tmnf. doing that for thousands of players every cup would take some time and make servers even laggier. akso switching in the last rounds would be difficult but maybe the only way
@@privatjetconnaisseur could just send the replays to each players in the match and verify by majority. it can't be too much work given that it's doing the same thing when showing ghosts.
devs need to fix this bug asap
@@privatjetconnaisseuryou are correct about the replaying but the verification can be done in other ways. For example collecting some anticheat info to verify no cheats are present with some verification of the aplication. The data can be send via encripted conection in which only TM issued certificates are accepted and each user will have thair own certificate. And send times and keypress info in the comunication. The encription can be handled on TM side on dedicated network infrustructure to offload any aditional load to servers and the serwers would only need to verivy that the retrieved data is in correct format and that all data arrived. And after this share the data with other users. So solutions are there for verification whiteout doing any hard work on servers. Ofcorse some other info would need to be collected but the info for times would not be delayed due to packetloss (for mitigating issues with packetloss TCP can be used to ensure all data came trough)
As a (former, by now I guess?) game developer, I can tell you: netcode is hell upon earth and there's a million things that could explain this. My guess is that the server overrides the time, if the time between packets is too large, even if it could validate them
Netcode is only relevant if action on one player affects other players action. Trackmania is effectively a single-player game, so other than visual glitch there is no excuse on validating inputs in real-time.
Yep, netcode is hell.
Though for things like the cup-of-the-day, I feel like the client could record the inputs, send to server at end of a round, and server could verify that the reported time matches a simulated run and client is not mis-reporting.
The fact that the rounds being so short, the physics consistent (hah!), and the gameplay-loop being so discontinuous (clearly segmented into rounds with several seconds between each), means it should not be too difficult to do seamlessly. Combine that with how it only matters for the top2 players (or rather, the 2 players surrounding the cut-off mark. Or a few more, for the few cases where ppl are within thousands around it - either way, only need to verify runs where server and client time differ *AND* the cut-off segment differs correspondingly) and there aren't even many simulated runs that have to be run. So should not be infeasible.
That way there's no netcode required! Only "post-processing" with plenty of latency allowed!
Hell, if it does eat too sizeable a chunk of time, it could even be further optimized thanks to the checkpoint system. Only have to simulate the splits where time, position, angle and speed differ between client and server, cutting down the number of inputs (and simulated time) greatly, potentially. Depending on how they do things, the server might even be rather idle during the actual round, meaning some of that verification (between checkpoints) could be done _during_ the round (though the more you divide such a system, while also making it immediate, the closer you are racing towards doing netcode again :p).
Yep, I also think this is some override, or slightly overeager cheat prevention, fallback in case of missing packets. Still, unacceptable.
I'm surprized there's a difference in replay considering deterministic physics. My guess that client and server keyboard input is different for some reason. It's surely possible in case of real time replay, I guess server tries to interpolate input in case of late/missing packets from the client.
So my theory is that near finish Wirtual's input packets were delayed/dropped and server had to show not quite accurate position in the replay for everyone else. But this doesn't explain the time difference between client and server, hard to tell.
@@hanifarroisimukhlis5989 you're over simplifying. In a perfect world the game could just be running locally and each client just sends their times to the server. But people are people and this is very exploitable. Cheaters would just intercept their packets and send whatever time they wanted
A realistic solution might be to capture the player inputs when there are less then 12 players left or something, then validate them by running them on the server and gettting accurate timings from that. Even this is susepible to cheating, but it would be better than just relying on internet connetions
I don't think it will be realistic to use this method for every single round. It would just add too much overhead and wouldn't achieve much
The fact that this problem is so widespread that it's even in pro matches where money is involved, shows that it needs to be addressed.
There are couple common ways of doing multiplayer:
-*client is right;* server only cares about position and time.
Times are as registered on your screen; very prone to cheating
-*server is right;* server cares about inputs you make
Resistant to cheating and lasting desync; prone to rubberbanding as server does not register inputs and backs you or even does not let you move
It seems that Trackmania uses a mix of those, where client is right about position, but server is right about timing, which is prone to something as little as ping fluctuation in such competitive setting.
What I think *SHOULD* happen, is that your game sends precise timings of your inputs not "now left" but "left at 20.45" so server can then save a demo/replay that it then can simulate as a TAS since game is deterministic, at least for those spots that matter (all those that KO and top few), and use that as true time.
as far as I understand, replays are saved exactly in that format (key press/release + milliseconds from start), so it should be simple enough to register it like that.
The only problem is, this probably uses A LOT of server CPU time, so it would be reasonable for Nadeo to implement a hybrid approach where this can be turned on only for key cases (like div1 COTD, and maybe only the final round for other divs)
@@PronteCoI completely agree with this approach
One of the reasons I like playing TM, and not other multiplayer games, is that I don't have to worry about ping, or any other problems that other online games have. If that's not the case, then TM loses a substantial amount of value for me.
this only happens on online competitions, not on regular solo play.
@@JuanEstradaGiuria trackmania isnt a regular solo play.
Bad connection has always led to problems playing online. It’s not new in tm2020. You would have an invalid run that wouldn’t even be finishable in tmnf.
@@KaptainKerlmy first run on Detroit 2 online... 3 hours in my internet got wonky for a second and then "this time will be ignored" popped up... I was not amused
"When will NVIDIA make a graphics card that doesn't need an update every week"
For context, I'm a programmer with a graphical computing specialization.
The frequent updates aren't needed per se, but a nice to have.
1- the most impactful reason, NVIDIA has added a metric ton of software features, both for users (like DLSS, reflex, rtx voice, video up scaling, etc) as well as for developers (optix, mesh shading, VRS, etc) that are essentially just apps that run in your GPU. The same way Adobe premier or some other piece of software you may use update, so do these apps... Everything in computing is becoming more and more software defined as we go on.
For stuff less relevant, but that still have an impact...
2- GPU's are just... Weird to code for. The overarching frameworks are just no where near as mature as CPU's, and the massive parallelization introduces many quirks. If you code something in C, you can be pretty sure it's going to work as intended. With a GPU not so much, weird behaviors are essentially unavoidable. Vulkan only came out in 2016 for example. Stuff is being patched all the time.
3- On the same vein, CPU code is basically stagnant in terms of patterns. For GPU devs are still constantly coming up sneaky ways to optimize the code, exacerbating the first problem.
4- bringing the first two points together, you'll notice most updates come tied to a big game release. These make sure the game words properly as the devs work directly with NVIDIA.
At 25:14 if you slow the video down and go frame by frame, you can see that right before Wirtual crosses the finish line (ahead of Yannex) he suddenly stops midair, allowing Yannex to overtake. He then suddenly jumps to in front of Yannex, but by that time, since Yannex had already passed the finish line, Wirtual had technically crossed second. What Yannex saw is what the server saw, because Yannex is only viewing through the lens of the server, so on the server side, it registered the time Wirtual crossed the finish line on the server.
It's important to note that this kind of thing is generally done to eliminate cheaters. If the client is allowed to transmit their time to the server, then there is the opportunity for a player to cheat their time. You might say "Well, but the server has to verify that time, so you couldn't get a cheated time on the leaderboard anyway." But, that only applies to replays. In a live setting like COTD, a cheater could theoretically turn off their internet mid-match, hack themselves to the end, then send that data up to the server and claim "Oh yeah, I just drove that. Must be bad packets." It's unfortunate that it happened right at the end of Wirtual's race, but the only solution here is really for Wirt to get better internet DinkDonk
could take like 4-5 seconds to run the physics at hyperspeed to verify the replay right after the match ends.
in some places, getting better internet literally isn't possible, like in Australia where I live
@@Psychomaniac14 That's true. Ultimately, someone is going to have better ping just by the nature of where the server is. Call it pay to win, I guess :P
They just need to actually use the inputs/replay in case of packet loss. Probably a time out of 5 seconds or so, such that a bad actor can't easily exploit it. It wouldn't be the client telling the server what their time is, it would be the client telling the server what they did and when, such that the server can check the time.
@@Psychomaniac14 I'm aus too, I wholeheartedly recommend launtel residential, check them out
I think the server's timer keeps running until there's a response from the client, and then the server verifies the client's time. But if the discrepancy between the client's time and the server's time is too big, then the server's time is prioritized (possibly to prevent cheating). But this also means that unstable internet can lead to slower times if there's packet losses happening at the time of finishing. I don't really know that there's much Nadeo can do to fix this unless they ditch the server timer and rely entirely on client times - but that could make it a lot easier for people to cheat times. Manual review might be an option for certain parts of the pro scene, but track of the day has far too many participants for such a thing to be remotely feasible.
guys i think this cup of the day mightve been controversial... idk though, just a guess...
I was thinking this too
I don’t believe you
if you read it carefully the video title actually alludes to that so I think you might be right
Wirtual controversial? That has never happened before.
@@DanielWilson-d7i dude he OBVIOUSLY cheated the midori AT. he cheated so hard that he beat it without his cheater keyboard
I love trackmania videos! I hope there's no controversy in this one
wait until you find out, this shits tragic :(
There isnt
I didnt think Wirtual was this kind of twisted person.
@@GriziDaWiz it wasnt that wirtual was a twisted person, trackmania just relies on server to receive time instead of the client, theres a reason for this(when someone cheats runs offline in slowmo it is because they arent connected to the server meaning their time isnt verified)
however its still trackmania's fault, they should just make the client not tell the server the time but rather what the player did insteas during competitions, anyways there are so many people spectating these guys in competitive matches so theres no way they try to cheat and even if they do and make it to the finals someone will notice it for sure 100% basically making it practically impossible to cheat while not loosing time due to lag
This was the first cup of the day I caught on stream, and on the last round I hard gambl3d all my points that he would win, and that was the fastest I went from triumph to defeat ever
I was truly SCAMMED
LOL RIP
@@jornhertsig1398same as having your sports team lose unfairly, fans can feel scammed along with the one they root for.
@@jornhertsig1398 Wirtual lost a total of 0 channel points, I'm the victim here 😓
Censoring the word gambling ☠️
@hello-world-gaming yeah, too many of my accounts have been banned on different apps/sites to stupid stuff like that, I don't want to risk it 😂
I am going to protest in-front of Nadeo HQ until they fix packet loss timing error.
Thank you for explaining Elensar, i hope you guys are able to track down the issue and prevent these laggs from happening ! Also nice shirt btw, i love Architects and listened to their absolute banger songs just today 🤘🏼
prevent lag? its the net code that needs to be fixed not the lag... you cant just "prevent" lag...
@@shadoww7301You could move to a house right next to the server. Basically LAN at that point
yesss go architects!
super appreciated that explanation at the end :D this looked like such an awesome map
Wirtual dodges the pillar irl lmao 8:42
I appreciate the architects shirt in the editors note
🤟
Amazing band
it also happened on an other cotd where he got 59th last second before the timer ran out but it didnt register and he got 65th (which was from his time before his last finish)
By God what could possibly be about to happen on this fateful day?
Not sure, I don't think I've heard this one mentioned.
If I had a nickel for every time something controversial happened with a COTD directly involving Wirtual, I'd have 2 nickels
Which isn't a lot, but it's weird that it's happened twice
What was the previous cotd controversy?
W pfp
@@hotshot0079 Midori
@@hotshot0079 Midori. The number 34.
@@hotshot0079Wirtual was accused of cheating his author time on his Cotd map Midori due to using external programs (analogue keyboard) to automatically steer 34% when he pressed a certain button.
(In the game, you can only bind 20, 40, 60, 80, and 100)
@Wirtual, regarding the drivers, let me know whether this makes sense or if not, where I should explain further. Nvidia has something called "GeForce Game Ready Drivers" which means they work with developers to fine tune their drivers to work optimally across a wide range of games and hardware. As new hardware and new games get released there are updates to the drivers to ensure that the drivers work for the new hardware/games to the standard Nvidia expects. Of course, these micro-adjustments can lead to small bugs which will also get fixed in further updates as they become apparent. In essence, this means that not updating your drivers can lead to sub-par performance with the newest hardware/games and you might have small bugs that don't get resolved, leading to e.g. too much resources being used (which can lead to system degradation if you keep your computer running for long periods of time) or subsystem crashes (that too can lead to system degradation). On the flipside, it also means that if your system works fine, it'll work fine without updating unless you start playing new games or install new hardware.
I mean if he is the type of person who doesn't shut down his computer at all then his system/performance degradation is going to be crazyyyyyyy.
TM servers do indeed check the run, that is exactly what was done at the finish line, during the check the time was updated to 35.518, which is the 5th checkpoint time, then to the final 40.850 and hence the correction.
Data is often lost over the internet, it is normal. Games use input bulking or report-correcting to counter this issue. Something absolutely stoopid happened just before the finish line. Maybe an input discrepancy could not be corrected after the finish was crossed due to a faulty algorithm.
But Nadeo has robbed the man nonetheless.
I like the arrows and as a first map it's dope. As a track of the day... Kinda sus.
You're right that not updating drivers shouldn't cause performance degradation which means there's likely something else going on which happened to have gotten resolved when you updated your drivers last. I'd recommend going into task manager when TM is super laggy and investigating if there's anything hogging system resources when it doesn't need to be.
He also has his pc turned on for weeks if not months at a time, so something getting fixed when he updates drivers might actually just be because he actually restarts his pc when he does that
@@T0NI_ It''s not been a problem before, dont see why it would be one now
@@piknos3381 It *has* been a problem before, though? Even in this video he references the last time that he was lagging and people told him to update drivers, which he did and his problems were fixed
@@T0NI_ exactly, the reboot required while updating gpu drivers was most likely causing the positive impact on the performance. as long he stays a notorious "never shutdown" guy there is nothing to troubleshoot about this issue - as it could be every single service interaction running on his computer
@@lux2248 He also randomingly mentioning between all of that, that CODT servers are laggy sometimes? Like bruv are you actually frame dropping or is it just bad connection, cause lowering draw distance for objects certainly won't help you that much, if the connection is just shitty.
Loved the little explanation from Elensar, very clear and easy to listen to. Good vid
Wirtual anytime he has to touch a computer "Why is technology progressing?"
12:30 Banana Question. Then you have a mini intro line like "James forgot to lock up the monkeys..."
Also "So I got an F in my math test because James bought 52 Bananas"
Wirtual the man who, never shuts off his PC, never updates drivers, plugins, anything, and doesn't close TM (and for some reason TM runs hot, even on my high end PC) when something goes wrong with his game. Listen it *maybe* had nothing to do with that but still man you're just stacking debuffs at this point
he could just type 3 words into google 'why restart PC' and would get result over result telling him that leaving his PC running for days causes performance degradation.
But comparing all possible PC configurations and resulting quirks to consoles seems to be easier.
My machine has been running without a reboot since february. your argument is invalid.
@@MrStonedOne "i have smoked for 30 years and i never got cancer" ass argument.
Yeah, like obviously it's sad that this happened in the very final round on the very final checkpoint with just enough time to fall behind, but there's also something poetically ironic about the fact that a good chunk of this video is spent complaining about how graphics driver updates are not worth doing
would love to see a full videos worth of that issue, especially since it causes so many problems
It took me and hour and stress, but I finally found the song between 5:05-7:57. It’s called infestation I think from Terraria, and the artist is DM DOKURO. I have many songs on my playlist that Wirtual plays, and it’s pretty cool.
The issue is that Wirtual leaves Trackmania running for months at a time.
I don't actually agree with you, I'm just here for ribbing wirtual about his relationship with updates.
Yeah it is funny how he undertands nearly nothing about tech but refuses to accept that he needs to reboot his PC, to fix memory leaks for example.
It is not like he has hints of that happening because his performance gets worse over time, but can't argue with somebody that thinks they are right without knowing anything.
@@user-to7ds6sc3p you have no clue what you're talking about. Not only are memory leaks not a given, but also they don't impact anything unless you're actually running out of RAM. Modern operating systems are designed to run 24/7 for months no problem.
@Sopel997 could certainly be a memory leak though. And yes the os is designed to run for months. Trackmania is not.
@@Sopel997the os is able to be run for months on end tm on the other hand is not
We've seen so much of Buckley, it's fun to see Elensar
The reason drives need constant updates is because Windows will have updates and games will have updates that the current driver isn't optimized for so they have to update it so that your computer won't be slow when playing said games
You dont need to update your graphics drivers every time a new version comes out. Do it only every couple of months.
Or when a new game comes out that you want to play.
okay but how is this not a bug being reported/talked about en masse?
this feels like an incredibly dangerous thing to have in a game that counts down to the tenth of a second.
5:07 I kept wondering if there was a reason wirtual wasn't going on the inside of the turn, just a classic case of "el no sabe" 🤣
I think it's the game time display in the game who isn't exact. When you display images on the screen there is a delay.
The timer you get it's from the beginning of the computation, it ask to be render.
Then it's render.
The display time can be many frames late.
A CPU have many task in parallel, get the time isn't the most important thing and that why it can jump some frame
Finally, a new main channel video in 5 days
Wirtual playing infestation in the background is what sells this stream for me
wait does the server not double check if it got the right package?
the server should wait a few seconds at least for the player to send time packages. a few seconds between rounds after finishing isn't gonna tick players off
Imagine a new graphics processor is like a new map. You might be able to drive it pretty well when it's first made, but the more you drive it, the more you realize where you are losing performance and how to make it better. And you find ways to keep from clipping edges and losing speed from minor errors. That's what display drivers are like. They are constantly improving as they learn more and apply it. Yes, there will be times where they missed something and it causes an issue with certain games or programs, but they usually work to fix those quickly when they realize they exist. So overall, the drivers get better and better over time to where the drivers towards the end of a card's life cycle are going to give very noticeable improvements
Have you ever considered a video or series where the premise is "If you are at world rank X, you should make sure you understand these techniques to improve by (50, 100, 1000, 10000, whatever) ranks"? I think that would be useful for any newcomers to know what sort of knowledge is common at different rank tiers, so that they can hear about these high level techniques, but understand that they are unlikely to benefit from it until they're closer to whatever rank.
12:36 i work in a grocery store, our online orders get multiple people a week that order 60 or more bananas, on Monday someone ordered like 150ish lmao
i would have assumed trackmania would be doing a live rollback similar to how hit detection works in certain games,
the idea being, since the game runs "deterministically" the game must be using a loop that registers inputs at a certain intervals.
those inputs, "time stamped" with the interval they were recorded in would then be sent to the server, the server simulating the vehicle physics and collsions independantly, would, on packet loss, store the last valid position, provide a prediction for others to see, and when the missing packets come back in, make sure they're in the right sequence, before picking up the simulation from that last valid position simulating till it gets to the latest timestamp or another missing packet.
if that were the case, the times shown when going passed checkpoints should always stay consistant between the players game and the server, since the server is simulating it the same as the actual game on a players computer, trackmania being "deterministic" any calculations made for both the player or server would be evaluated the same and get the same result, providing the same final time.
why isn't this how it works?
first thing, im 30 seconds in, I love the signage on this map making it clear where you need to go
there is no need to update drivers, however updates can fix bugs (you might not be experiencing) or security issues. for example right now AMD has a bug in their driver that allows privilege escalations from normal user to admin. in that case you should immediately update
The only one to blame here is Wirtual for not getting business contract with his ISP into his house and taking legal action against the ISP for loss of revenue. When you pay for consumer grade internet you get consumer grade service and consumer grade routing.
His streams have been lagging ever since he moved into that house and he has done nothing to change the situation, just the other day on stream he said that he hasn't even tried getting a new router, ISP supplied routers can get overheated from streaming due to them being complete dog .
Imagine being Nadeo and paying Wirtual to cast TMGL when his stream crashes and lags out on the regular.
If any of the Norway based editors sees this, please have an IRL chat with him and do everything for him because Wirtual is completely incompetent when it comes to technology. Personally if I made the money he is making I would get the business contract with the ISP, a brand new non-refurbished trash router, and change all the ethernet cables.
He doesn't live in a third world country, I had better internet stability playing Unreal tournament in the early 2000s.
Bro chill, Norway might not be a third world country but I can tell you internet is expensive when you have to run cables through mountains and fjords. Internet and mobile services are a lot more expensive in Norway than Sweden and Denmark.
The reason the game trusts the server time over the client's time is that if it were ro do the opposite, it would be opening the door to all manner of cheat methods.
Because the only real way for the game to reliably check if a client's time is legit would be by referencing the server time and seeing how they line up.
Were the game to trust the client over the server, players could just figure out how to send in a fake time to the game without any way to check for authenticity, you could theoretically drive a 10 minute lap but tell the game it took you a minute, and the game would have no choice but to believe you.
For live events like CotD, what they might be able to do would be to save the inputs of each lap and have those checked when client and server times dont line up, since cheats like slowmo arent exactly viable during a live competition. That might be a potential solution, though its possible that too could open the door for other exploits.
tldr: videogame code is weird and complicated, and anything online gets even worse, while anything offline is free real estate for cheaters.
I'm on my way to Croatia for vacation and I'm currently watching this video with my girlfriend who just fell asleep to it. Wirtual really has the superpower of sleep
It's 3am, and I'm definitely sleeping to this.
The Constant graphics driver updates are caused by constant hardware updates and constant software updates. For a PC, a driver needs to support almost all the hardware out there, where many companies utilize the chipset and report bugs and issues with the latest and greatest hardware as well as with the old chipsets. Part of this may be due to the custom hardware implementations. Second, Operating systems are always being updated. I have a very old computer yet graphics performance degraded greatly after a certain version of windows came out. In my case my video card became end of life and windows stopped supporting it even though I had the drivers for it. Nintendo, had fixed hardware and fixed OS so the game developers don't need to patch for OS and hardware changes.
a big part in the "gb games didn't need updates" and "graphic cards weekly updates" is that developers have become complacent due to resources being so available. The amount of RAM(or equivalent) for gameboys was so small, same limitation for CPU etc. We don't have to think so much about what we do, since computers do it for us.
Additionally, improvements are harder to come by due to technology changes. Changing a picture on a 4k monitor is a lot more resource-intensive than doing so on a GB screen.
And then the point that they just _couldn't_ update GB cartrigdes without calling them back, or some other dedicated infrastructure.
Also: pokemon (for example) is far from bug-free ;)
Sure his stubbornness about updating is little bit funny but in this case its totally reasonable to question why is the driver that 2 weeks ago was good thing that fixed stuff suddenly bad when the game has no major updates in between. In his case the lag is most likely something streaming related rather than nvidia driver related, if I would have to guess in last 50 nvidia driver updates maybe 1 has had something directly related to trackmania and that would probably be something related to raytracing :D
Like with Pokemon G/S, when clever people are given design limitations they'll find creative ways of making things work. Nowadays, the 'crunch culture' also doesn't help. Endless resources make people inefficient imo; an Exec probably said: "why waste time creatively cutting the size of Call of Duty, people will spend hours downloading it anyway, and the cash shop needs to be worked on".
In regards to driver updates, is not one of the things you can do is to update and apply a recommended set for each game you are playing when it is running?
Though i have no idea on the weekly general updates.
Your track is chosen as Track of the Day, and on his second run, Wirtual beats your author time. Humbling.
Wirtual should try nvidia studio drivers where the updates are less frequent instead of game ready drivers which updates every time a new game comes out
Maybe they should use TCP for syncing lap times but UDP/websockets or whatever they use for ghosts.
That way they will always know if a packet is lost wrt critical data, this probably happens a lot to the hundreths and thousands secons in a match. Just went unnoticed
Could be confirmed with wireshark, if lan uses the same system it could even be veriefied on the server side
Websockets are just a wrapper around TCP
Man did a Crabulon theme song and thought we wouldn't notice
I don't know if Wirt reads these, but here goes:
The reason why you have to update software is that it is very hard/impossible to create bug-free code.
Essentially, everyone writing code is very bad at it, and so either you spend an insane amount of effort and money writing the most perfect software ever, or you just say "screw ot we're doing it live".
So at some point, someone goes "huh weird, the driver crashes if I try to draw two green triangles in a row", tells nVidia, and they issue a fix in the form of a software update.
I'd be happy to tell you more if you want :)
You would think that since Trackmania is a deterministic game, the client would be sending input diffs over to the server while the client runs the game locally. This means that the client and the server play the same game, and the server is just there to check that the inputs actually lead to the car crossing the finish line at the correct time.
But from this it looks like the server counts the player as crossing the finish line whenever the client reports such. This would mean that a player's internet latency, usually measured in thousandths of a second, are included in the run time in the best case scenario. We have had cup-of-the-days where players finished within a thousandth of a second of each other, and the name chose the winner. How many of those had been a clear win for one player or another? How many where a couple of dropped packets resulted in a player going out?
As far as I'm concerned, Trackmania is not that competitive of a game anymore, at least in live matches. Having extremely fast and reliable internet as a requirement to be competivitve in a skill-based game is nothing new, but I thought Trackmania wasn't like that. It has the ability to not be that way.
What I imagine is that at the beginning of the match, there are 64 people playing Trackmania locally on their own clients. The inputs get sent to the server for replay validation. The local client is authoritative on what the inputs were, the server is authoritative on what the finish time is. Client checkpoint times get sent out to everyone with the logic being that it's close enough. This depends on the servers being able to compute the physics at 10x-100x realtime, which seems reasonable.
how many track of the days have been the mapper's first map? there was this one and the one by sneakyy (desert one with that one plastic lizard on the corner & the end cut)
are there any others?
this. this is the definition of pulling a wirtual. not not choking or choking when the pressure is on but this.
I hope someone explained about the game boy and pc being two completely different things. A console, a game only developed for said console, back when alpha and beta was only tested in house before shipping at a full release state v1.0 would never need a patch as it would only ever be played on that device. Where as PC'S run different specs, different OS versions xp, vista, 7, 8 - 8.1, 10 and 11. Plus different states of windows updates on said OS. Not to mention game porting to PC from console, too many variables, game studios try to address those issues in the dx dll's. Then video card manufacturers have to try and take those considerations also and ensure that all new drivers work in compatibility with every old driver which takes in all those other considerations also 😮
The issue aren't even his not up to date drivers. Performance doesn't get worse from that and I myself only check for graphics drivers once a month or when I see a potential driver issue in a game.
The problem is him not restarting his PC. Memory leaks, plus other things I don't have a professional name to through around for, making ressources unavailable causing performance to degrade.
As you said consoles have software made for that hardware and that hardware only, that causes those bugs to be way less common, and nobody leaves there console running for weeks.
@@user-to7ds6sc3p all good was just explaining to the interwebs hoping he would even check his comments, to give him somewhat of an understanding of the complexity of drivers. Not saying he needs to update, just saying why they exist the way they do. And totally agree with what you are saying and understand what you mean.
I think usualy this is what happens 24:17 because there isn't time corrections each time. But I think when someone is too laggy the server takes over and penalizes the time
not me screaming YESSSSSSS and then the lag split happens and I curse nadeo for 5 minutes
Kinda feels like an extension of the inside area of Banshee Boardwalk
Updating your drivers isn’t necessary, but extremely recommended. If you don’t update you leave yourself at a competitive disadvantage because your software is no longer optimized. Updating drivers is just maximizing the capabilities of your gpu. The reason we have so many driver updates is the same reason new, better gpus come out. Research and new technologies allow developers to provide us with the benefits of such, and we get more efficient gpus as time goes on without having to buy the latest one. It’s why my RX 580 4gb (worse than a 1050) can play R6 or the new cod titles or apex and I don’t need a new gpu.
Game shaders help of course, that’s the games’ way of optimizing to your gpu. But as with gpu driver updates, games get new shaders frequently and help you run the title more efficiently.
The combination of both helps keep your hardware not only relevant, but also gives you better performance.
Bros soul left his body every time he got a landing bug
rocket league has a similar issue with "ghost touches" sometimes the server dont detect a cosission with the ball and your car and the ball will visuell glitch trough you
Your layman understanding of drivers is funny. There is one game console from each major brand. There are infinite hardware combos on PC. And many orders of magnitude more software (that must be compatible with each other as they run at the same time and interact). They don't just work together. They have to be told how.
This bug happened to me when I was playing yesterday. I don't think I'll be playing this game anymore. I improved my time by one second in the last minute, but it registered after more than a second, so I ended up in division 17.
By the way, there's no chance of sending packets that long because of the internet. Another thing is that the position of the player would probably have to change as well. This would mean that if the packets are indeed sent this late, players in spectator mode would also see Wirtual behind.
So I personally think the problem is in Trackmania, but I have no insight into the game and I could be wrong, the problem could be somewhere else.
the problem was not that packet was late, but rather that it was lost making him stop in place before teleporting behind the finish line from spectator, opponent, and timing perspective.
Now it seems that either they use UDP or they don't use any late information given by TCP caring only about position that is "now", but whatever transmit method they use they also should use simulation on their end.
In essence if a packet is lost, server should keep moving the player as if no change in input was made, but when the lost packet with that change in input and a timestamp is resend the simulation goes back in time to apply for this input.
If done only to verify after the fact your time would be as you seen, then go up a second, and then go back a second on final result screen. But if done live you would never notice that you lost a packet.
35 years ago my teacher told me not to brag about my ignorance. WAKE UP!
I also very rarely restart or update drivers, but at least I understand the problems I'm giving myself... I've just decided that the annoyances from restarting are more significant to me than the ones from not updating. If something more significant becomes a factor then I reconsider updating (like recently, there was a bug in the driver for my 3060 card that was causing increasing problems in games that got fixed).
wow been watching wirtual videos all night and heres a new one xD
There's an ex dev in here saying net code is hell. Would I be correct in saying that they (Nadeo) have their current model of the server time being the determining factor in who wins as an anti cheat measure?
Just speculating as Trackmania as a franchise is know for having players alter the speed of their games locally.
So in theory if they were to change the current model. They would take the local game time from the system as the determining time.
If I was to be a cheater there's a whole plethora of alterations I could do to mess with the timer or the speed of the game.
It makes for a very interesting case study either way
well nadeo has weird system in place determing who wins at ties: based on the name at alphabetical order and apparently later they have updated it to based on user id which is randomly generated.
I was wondering to ask for it for a bit now, can you add date of cup of the day in the begining?
Bro I love how every video I have watched recently have all used terraria calamity ost soundtracks
Nice runs! What is the name of the song at your last run?
From what I can tell, as a computer engineering student, the graphics driver updates are mostly for optimizing to new popular games, with some bug patches, stability improvements, and the rare feature addition.
It's much more worth your time to stay updated in the windows settings because those bring security updates against the latest hacking techniques. If you're also playing games that were released relatively recently, then graphics driver updates are very useful.
Of course, you should keep updated anyways. Just imagine the computer scientists on their knees begging you to let them help you.
Even Wirtual's internet pulls Wirtuals
The reason nvidia keeps pushing updates is because new games continue to utilize new technologies that may run faster with updated driver for older hardware.
Now you might say, well still why would they need to updated it THIS often then? Because change in drivers may cause older software for unkown reasons not to work as expected or straight up crash. Nvidia receives customer mail complaining about that problem causing them to release a bugfix. Every bugfix (or multiple) is adressed as new update for the driver so that with the newest driver the likelyhood of it working as optimal as possible for the widest varity of software is the highest. Sometimes bugfixes create even new problems that first go unnoticed and need to be fixed later with a new bugfix for the bugfix... and so on...
Also... the driver updates are basically free advertisement place = extra monetarization
When playing on arcade servers sometimes this issue happens to me too, where I’ll be ahead of my pb ghost but still lose, Nadeo plz fix!!!
i can tell this will be controversial
I don’t think so
The racing is fine and all but everyone knows we are really here for the crabulon theme in the background
Me watching the entire video: Where's the controversy here?
The end: Oh.
As many here suggested it would be best if the servers took the inputs and replayed them quickly to determine a time, however this would take immense resources. A better solution would be to compare the time recorded on the client to the time recorded on the server and if they diverge then the time from the input should be calculated. But yeah, overall this is almost a beginners mistake as racing games shouldn't depend on internet speeds or package loss at all. Such a weird oversight and I am wondering how this issue got this far without being fixed. I mean 0.150 seconds is a HUGE margin in a game like this, that's sometimes the difference between WR and Top 1000
Gpu drivers get updated regularly because of bugfixes and performance updates in (new) games. Aswell as issues with new monitors or new software updates with obs for example.
Playing trackmania or rocketleague or tetris on old drivers isnt worse than on newer ones. But for best stability and newer games/software its always recommended to update.
Windows updates are useless. Their "security updates" break more than they fix or prevent.
One day Wirt is going to learn how computers work. And also learn that gameboy games themselves had patches.
The iceberg had a "tm guesses the times" part
His second run he hit author medal exactly down to the thousandth lol
Oh boy a WIrtual COTD video! I sure do hope this one doesn't have some controversial last round or anything.
Nice shirt man. Where did you get it? (if you don’t mind)
Well i updated my drivers for my 6900xt and the system started hanging and sometimes freezing when just launching a 3d accelerated app or game. had to freaking roll back to march drivers because those actually worked. but then, that has some visual glitches with ratchet & clank that were fixed in newer releases . good stuff
"this is a fullspeeder" he says while drifting
Elensar rocking the Architects Shirt Pog
I don't know how this would be fixed. If it takes client time it's too easy to cheat. It has to take the server time, and then there is always a slight inconsistency. Try playing any modern shooter game, after a while you start to notice that there is always some lag or lag compensation going on in every single shootout, sometimes in your advantage, sometimes disadvantage.