By popular demand, I will continue this series and keep doing these tests for different aircraft at different brs (just not for the next week and the next video). I'll probably start with 3.7, then skip to 4.7 and work my way up to maybe 6.3 However, due to the labor intensive nature of the process, I will be locking the graphs for the vehicles behind a paywall. To unlock each new EM diagrams you must pay into a gacha lootbox system and each pull will cost you 3 doubloons and a sheep liver with a 0.0000% of getting an EM diagram that you haven't gotten yet. In all seriousness though, this does take a lot of time and I do this as a hobby and not a job so I'll get to it when I feel like it but it may be some time until I start making more graphs. It's pretty annoying since my laptop is also limited in performance meaning stutters and crashes and having to go in and prune the data. If you want to try and make graphs yourself, my testing method and code are all shared in the previous video and google drive, though it isnt annotated so you'll have to figure that out on your own.
@@CatWerfer honestly Id be happy to pay for these in future (not a lot Im poor lmao), but would also consider contributing if there are things I could do without risk of backseat driving or making you wait on a rando for something haha
@@CatWerfer the simplest thing I can say is it wouldn’t be unreasonable to release more stuff on patreon than you do publicly, people aren’t entitled to your work. It will definitely get leaked, no way to lock it down, but it would still be a reasonable excuse for people to support you and have you feel like you’re giving back. I won’t complain about interesting free stuff though!
could you do all way down to 3.0 or 2.7 since its the actual start of energy being meaning full in a dog fight and not just biplanes 100% pulling at their stall speed
maybe, but i think id still prefer everyone to have access to the stuff i make and if people want to support me they can choose whether or not they want to send in donation money. People aren't entitled to the stuff I make but reason i make videos is to share stuff, that's what i enjoy the most, so i'll keep doing it the way i am for now and ill make a channel for donation if someone thinks they'd want to support me that way.
@@CatWerfer Btw did you see statshark? On that website you can generate EM diagrams for every plane in wt. But I don't know how accurate those are in relation to yours
This is so cool man. I didn't even realize the f4 was better than the g2, or that the spit Vb was so much better than the mk. IX. Thanks for all your hard work
Early Bf 109 G models were hamstrung by a performance restriction on the DB605. The intake manifold pressure was limited to 1.3 ata due to reliability issues early in the DB605's life. As a result, the DB605 in early G models produced around 1300 hp, around the same of the DB601E powering the F-4. The G was heavier than the F, so flight performance of the early Gs was worse than the Fs. In 1943 this restriction was lifted, so maximum power was increased to 1500 hp at 1.42 ata. Unfortunately War Thunder does not have the flight models with these specs.
The Spitfire Mk.9 is far superior to the Mk.5. However this reality is not modelled in War Thunder because the Mk.9 in the game uses the wrong specs. The Merlin 61 should be producing around 1560 hp in its most powerful setting, but in the game it's modelled as 1400 hp. Problem is, that 1400 hp is its rated power at 7000 m altitude whilst in its first supercharger speed (aka the speed for low altitude performance).
@@CatWerfer It's confusing. The F Mk.9 started out with the Merlin 61, but late-model F Mk.9s got the Merlin 63 which could produce 1710 hp. The LF Mk.9 got the Merlin 66, which was basically a low-alt version of the 63. Produced about the same power, but at a lower altitude.
I don't know what background defyn has in aerodynamics or physics, he always kinda struck me as a "vibes-based" player where he sorta just played planes and gave advice without evidence beyond his own experience (which isn't a bad thing since there really isn't an easy way to gather airplane data outside of just playing the game), but if he has that kind of a background and has interesting things to say about em diagrams then sure
studying mechanical engineering (undergrad). I don't have an aero background myself but I've done a fair bit of reading about this in my own time. People more well versed in aero come around and correct me when im wrong so it's nice there are people who can help me convey accurate information in that regard
Bro this is awesome! You're an absolute legend for doing this. Would love to see some charts for La-5 variants in this BR range. I know you're burned out but I think the next data points to consider would be like 3000m as that 'mid-range engagement' that happens at the start of a match. Again, thank you so much for what you've already done. This is such awesome info. Subbed and shared on my channel.
I certainly would like to do altitude performance tests as well but frankly speaking I don't think I want to spend time comparing planes at multiple altitudes. I know that it's important and some planes gain a significant advantage at altitude over others, but its pretty time consuming so when I do make new graphs, it'll probably be on different planes. My code and methodology are all shared in the drive and previous videos so if you want you can test for yourself too, though it can be pretty painful. La5/7 and other planes at other brs coming not-so-soon(tm)
lol thanks for the complement but I don't think the average war thunder player would sit through a 30min video about data analysis like i did in pt 2. (though i guess the average wt player would also probably not be that interested in em diagrams either)
@@CatWerfer hey now! Don't just destroy my limited faith in humanity like that! You mean to tell me the average WT pilot is an unpaid actor actively trying to troll their own team?!? Inconceivable! Looking forward to future videos. Maybe you can figure out what the rocket crosshair is supposed to represent :D
im pretty sure the rocket crosshair is just tied to nose direction and maybe the angle the rocket is mounted to the wing at. maybe a more interesting to look into is to try and see how radar lead indicator and arcade lead indicator works
12:39 - Yak-3 has a smaller, more aerodynamic airframe than Yak-9s. A smooth mini Yak-9. Still, a great video. 30 planes, damn, that's lot's of effort! I was suprised by energy advantage of F4U-4 vs Bf 109F4 at most speeds. Tho it makes sense cause you tested F4U-4 at low altitude where it has 2470hp, while it already drops to 2100 hp at 1.7km. 109F4 gains big power advantage in at 2km compared to sea level.
huh, the airframe limits in air rb are basically identical for the em plot, i guess I shouldve done a bit more research instead of assuming the airframe is the same
F4U-4 is carried heavily by the methanol-water injection (limited WEP). For the altitude performance drop though, is that with manual supercharger gear? I found F8Fs and F4Us on AEC have delayed supercharger gear switch compared to what the manuals suggest for fast-scrambles.
I wonder why gaijin is so cagey with their in-game aircraft statistics and how things in game actually work? Why would they want to keep players in the dark on these matters?
likely they're wary that people might try to develop a competitor to war thunder using similar aero methods. they have patents on their instructor mouse aim after all...
@@CatWerfer That patent also doesn’t go into detail about how the mouse aim works. It does not show any control loops, so the exact method is not protected unlike with other engineering patents that show and claim rights on the exact method used. For example, many engineering patents claim rights on a specific circuit topology used.
@@fablearchitect7645 as far as I know, software patents don't need to go into the actual details of how it works, just what it's supposed to do. It's the reason why Sega successfully sued Fox and EA in the early 2000's for patent infringement due to the game "The Simpsons: Road Rage" using an arrow to point towards objectives, which functioned the same as the arrow in "Crazy Taxi" despite the code likely being different.
yeah, though keep in mind that that is the ideal climb speed at sea level and the ideal climb ias drops as altitude increases. Also theres a pretty wide range for the plateau of max SEP so it may not be exactly that number because of the curve fit (check out my "optimizing climb" video if you haven't)
Very impressive and thanks for the effort. Now its more clear to me why most of the German planes tactic to boom and zoom is so important but also wont work in many situation.
Great work! Rare to see some Em diagrams comparison for war thunder. Now we need some 5.0-6.0 BR palne comparion such as J2M2, Ki84, Spitfire LF Mk IX,Yak3U, Ta152H and Bf109K4.
I like the ITP. It can pull something like 11g or 12g turns. The guns make it quite deadly, but outside of a dive it is to be used carefully, since it is not the best climber or accelerator.
I like transparency with some flaws of these diagrams. Still a great resource for anyone who wants to use a few neurons to learn from these diagrams, there's noting like this that's available.
yeah i think i make sure to make disclaimers pretty adequately for my videos. data isn't useful if its mixed in with garbage so i gotta make sure people know what is and isnt high confidence in my data
i'm probably only gonna do subsonics if i even do jets, i have a feeling that the hexic curve fit will start to become not good enough once i have to account for wave drag...
@@CatWerfer the wave drag is the largest drag component at supersonic speeds, but typically the drag force will be approximately proportional to the sears haack body drag which is proportional to dynamic pressure and max crossectional area to length ratio squared. The way it shakes out though that’s basically just a large drag term proportional to mach squared coming online though the transonic region. I think a 6th order fit should be able to handle that alright, but if not there are some fairly simple things you could do that don’t assume an underlying form like a smoothed spline or k nearest neighbor model
30 aircraft?!? get some sleep, damn. Excellent work, really. These graphs make it apparent the Pyorremyrsky, XP-50, Wyvern, Yak-3 and maybe P-51 D10 are in their own league, which anyone playing this bracket already knew (tldr: p2w confirmed) The P-63 seems excellent too, but this is only at sea level Maybe the color gradient could show the relative difference in MER instead. This would make advantages in stall-fights and climb-rate more apparent, which are quite relevant to the meta
Interesting to see the A6M3's current EM diagram. Per US TIAU no A6M they tested could turn inside of a P-47 at 300mph IAS. Also per Japanese reports the Ki-43 could always pull more Gs than any A6M at high speeds. This was a deliberate decision by Mitsubishi for the very reason that came to plague early Ki-43s giving them an early bad reputation in the IJA. Pilots were ripping the wings of early Ki-43s trying to pull out of high speed dives. Mitsubishi was concerned about this during development of the A6M (because of the insane performance demands of the IJN) so restricted how much elevator could be pulled via balancing the elevator for low speed. This resulted in elevator compression earlier and heavier than the Ki-43. If the A6Ms ever get this elevator nerf in game they are going to handle like trash above 300mph IAS. Also this elevator nerf would make the historical get away tactic for US pilots against the A6M actually viable to use in game. Dive to 400mph IAS, roll to the right and then pull out of the dive.
huh that's pretty interesting, thanks for sharing. I don't know very much about aircraft flight performance IRL and am really just modelling things ingame, but its cool to think about how they could implement that. Do you think it's likely they'll change that? there are a lot of inaccuracies that war thunder has like not using the G limiters on f16s that are there since the pilots in war thunder are kind of superhuman with their g tolerance.
@@CatWerfer They have been implementing historical elevator compression randomly to a number of airplanes since the F-16 was added. The F4U-4s and 7 were the first. The MiG-23s and 27s were maybe the most famous recent example. So will they? No idea. I think it would be nice because elevator compression has been the one major performance aspect that Gaijin has used for balance rather than sticking to historical documentation.
@@CatWerfer If you do throw in a bug report on the top speed of the J2M2, 3 and 5 while you are at it. Their in game top speed is based on Japanese top speed data when nearly every other major Japanese aircraft of ww2 in game uses US top speed data. Why does this matter? Japan didn't use WEP to collect top speed data. The US did. The J2M2 and 3 topped out around 405mph and the J2M5 topped out around 425mph. I didn't include the J2M4 because the US never tested it. It's actual top speed is a historical unknown.
If someone's gonna bugreport it, it probably won't be me cuz idk much about irl flight performance and i also don't know what sources to look up for that kind of testing info. maybe you could make it since you sound very knowledgeable on this topic?
@toki_tsu_kaze5511 gaijin is notoriously cagey with their in-game aircraft statistics and how things in game actually work so it's not likely but maybe
8:24 - I point out that at speeds above the sustained turn, the SEP gradient (color of graph) is deep red along the top edge (max turn) meaning that while the f6f does turn better (blue outline is farther up than the red), the 109 has a better maneuvering energy retention (loses less speed in a turn). 14:08 - The opposite is true with the typhoon, since the top edge has a blue gradient, meaning the typhoon has a higher SEP, loses less speed in a turn, and so has a better maneuvering energy retention
@catwerfer Nice work! Wish Gajin would put EM diagrams in the game, definitely makes mastering a specific aircraft a science, and gives you an idea of where you need to be when picking a fight. If you did put these behind a paywall, I would gladly subscribe to your patreon. Also consider videos where you discuss the EM chart for a specific aircraft while in a match and demonstrate where those advantages and gradients are. If you could animate one to show a dot where you and an enemy aircraft are on the corresponding charts, you could pinpoint where a fight turns in favor of the victor. Would love to fly with you sometime maybe a stream? I’m an EE in the aerospace industry, but I also fly my own yak 52 :)
lol im too much of an introvert to have considered streaming, I dont know what i'd do to make war thunder gameplay entertaining to watch. Thanks for the complements! I am actually working on a video on how to compare EM diagrams a bit more indepth cuz this one i kinda glossed over it quickly. I laid out the foundations in my part 1 and 1.5 videos but i need to do a better job of explaining how the 2-plane plot works.
Did you tested Bf109G2/trop or normal G-2(unavailable in germany, only available in italy and sweden)? the trop has more drag due to the sand filter on the air intake, so it performs sligthly worse Also worth nothing that the DB605 has better cooling compared to the DB601, so the G-2 can use WEP more often than the F-4, something that can make the difference in a fight Will be cool to see the 3.7 graph and how the C205 serie 1 murders the E/M graph of everyother plane or the 5.3-6.3 superprops graph will be very interesting to see
the club penguin mafia really thought they could get me but now i have relocated to a remote island with a bunch of gigarich people where the government cannot monitor my activities
Maybe you could create a detailed manual on how you tested and gathered your data. Then create a public google document or something and everyone can test his favourite planes and share the data. With a few hundred people doing that you should be getting good averaged data.
sim mode basically removes the instructor the instructor does a few things: - translates mouse movement into control surface inputs to make your aircraft's nose point to the mouse cursor - limits aircraft turn by AoA so the plane doesn't stall The base physics of the aircraft stay the same, but without the instructor, you're able to turn harder but also need to make sure you don't pull too hard and stall your plane. you also have to manage control surface inputs and trim yourself (until you get to fly by wire and jets with computers helping you fly). That's why a skilled sim pilot can outturn someone using mouse aim in RB.
@@lonemarkkingoftypos3722 It is a 4.3 maybe 4.7 plane, not a 5.0 or 5.3. It is close overall to the Pyro. A bit slower, better climb sub 2km, worse climb the higher above 3km you go, better roll and better turn. Absolutely cracked sub 2km for 4.3 even if a bit slow. Good 2km-5km. Don't take it 6km and up. That said it could end up 5.0 because bad player base + turn fighter syndrome.
@@CatWerfer Yes, sadly. Thanks for all the effort you put into these videos. A friend recommended me this video today and now I'm subscribed to the channel. Good work and I'm looking forward to seing more of these videos
if there are graphs for 2 planes on a single plot you'd like me to do for planes I've shown in this video i can make that but for new planes I would have to go in and manually test them and i am not going to do that for now
I understand, that's really hard and time-consuming work. It is just a suggestion of two of my favourite aicraft) Thank you for this amazing info, nonetheless!
Dude, instead of doing them all in sheets, see if you can make a site or comparison tool or something (I can't rlly help with something like that sadly)
it can be kinda fun sometimes since very often the enemy will try to take you on in a 2v1, but the a6m2/3 is also pretty damn good at juggling multiple people if they try to turn with u
the yak 3 got some kind of upgraded fuel mixture iirc, something like 100 octane fuel that the yak9t/k didnt get tho so i assumed it was some kind of engine related upgrade
The Yak-9K and 9T by no means have the same airframe as the yak-3. The yak-9 airframe has quite a lot of differences, most notably larger wings and more weight
My only nitpick is that you picked the worst A6M3 instead of the more maneuverable Model 22 at 4.3. Those clipped wings just weren't it. Overall this is mostly unsurprising, though just from looking at these EM graphs I have a hard time keeping up with your conclusions. Btw aren't you glad that there's now Eminem's Yak3 at 4.0 with that kinda performance? Totally balanced! Also, A7M2 needs to be on this. It's an incredibly maneuverable plane that will laugh at any Spitfire that decides to turnfight it. Ki-27 would be funny too because IIRC it has the single highest sustained turn rate of any actual fighter, only beaten by a min fuel BI - coupled with a low rip speed I assume it's EM graph is gonna look more like a fence post than a doghouse.
Really? I like the base a6m3 much more than the others, it already outturns everything like crazy so the extra turn you get from having unclipped wings really isnt worth it imo. If its difficult to read the em diagrams, i might make a video explaining how to read a 2plane plot separately. The conclusions weren't really supposed to be surprising, just a more concrete way of describing what we already know from playing the game. A lot of wt commentators will list off a bunch of characteristics about aircraft without any hard facts to point to. They aren't wrong, but its hard to verify without just playing the game a lot
@@CatWerferbtw, do you have these graphs in a google doc or spreadsheet? I think they could be very useful for helping my squadron learn flight characteristics at this br.
@@CatWerfer obviously sarcasm, bad tone which i do not appriciate, but yes it does, just kinda spam the flaps and force a stall fight ASAP and the f4 is the superior plane. if the fight is prolonged the xp50 will win, but f4 is the better flight module
@@ProudOne the 109 only wins in the lowspeed stallfight if you look at the graphs. If you can force it into a lowspeed stallfight then you have a solid chance of winning, but in basically every other scenario (airspawn, headon, horizontal ratefight, scissors, chase), the xp50 is better. If you can consistently win against the average XP50 player in air rb thats all well and good, but part of that is because most xp50 players aren't that good and dont use throttle control+flaps themselves. Put two competent players who know their planes and their advatanges, one in an xp50 and one in a 109 and the xp50 can see you're forcing a stallfight and use its better linear energy retention to disengage and force a fight that's more favorable to it. Do you see people complaining about the xp50? yes Do you see people complaining about the 109? no I think your argument is pretty bad faith so i responded in kind
@@CatWerfer no it was not in bad faith, and you’re still wrong. The f4 is the better flight model, although it is a good match up. I don’t base this off of beating them consistently, but rather one of the best prop pilots in the game if not the best says this and has explained why; with actual fights against another very competent pilot.
mmm yessss the very gigachad and definitely not whiny baby move of crashing to deny the kill so that both you and your opponent who is playing the same painfully grindy game lose together and the snail wins. it would've been funny if i made a video talking about technical stuff and then i did a psych-out at the end where i just tell everyone to crash lol
@@antonisashitteragain6993 Or you could play smarter, like Leonidas meant, rather than harder and force them to use as many of their resources to kill you as possible a la just like what Leonidas did at Thermopylae. Most props around this BR have less than 20 seconds of total firing time. Getting them to expend as much ammo as possible can be a difference maker in a match. Likewise getting the opponent to burn as much energy as possible by dragging the dogfight out as long as possible makes them vulnerable for longer to any teammate that might come along. Nothing is better for me than an opposing player crashing if I so much as get near them. I don't have to expend ammo, fuel, effort or energy to get you out of the match beyond just flying in your general direction. I get to move on to taking out your teammates with the least amount of effort and resources expended possible on you. Quoting Leonidas but then not following his example. Nothing could scream fake poser any louder than that.
By popular demand, I will continue this series and keep doing these tests for different aircraft at different brs (just not for the next week and the next video).
I'll probably start with 3.7, then skip to 4.7 and work my way up to maybe 6.3
However, due to the labor intensive nature of the process, I will be locking the graphs for the vehicles behind a paywall.
To unlock each new EM diagrams you must pay into a gacha lootbox system and each pull will cost you 3 doubloons and a sheep liver with a 0.0000% of getting an EM diagram that you haven't gotten yet.
In all seriousness though, this does take a lot of time and I do this as a hobby and not a job so I'll get to it when I feel like it but it may be some time until I start making more graphs.
It's pretty annoying since my laptop is also limited in performance meaning stutters and crashes and having to go in and prune the data.
If you want to try and make graphs yourself, my testing method and code are all shared in the previous video and google drive, though it isnt annotated so you'll have to figure that out on your own.
@@CatWerfer honestly Id be happy to pay for these in future (not a lot Im poor lmao), but would also consider contributing if there are things I could do without risk of backseat driving or making you wait on a rando for something haha
@youngbloodbear9662 lol nah even if i did want to pay wall this stuff it would be like making nfts that can just be screenshot and shared
@@CatWerfer the simplest thing I can say is it wouldn’t be unreasonable to release more stuff on patreon than you do publicly, people aren’t entitled to your work. It will definitely get leaked, no way to lock it down, but it would still be a reasonable excuse for people to support you and have you feel like you’re giving back.
I won’t complain about interesting free stuff though!
could you do all way down to 3.0 or 2.7 since its the actual start of energy being meaning full in a dog fight and not just biplanes 100% pulling at their stall speed
maybe, but i think id still prefer everyone to have access to the stuff i make and if people want to support me they can choose whether or not they want to send in donation money. People aren't entitled to the stuff I make but reason i make videos is to share stuff, that's what i enjoy the most, so i'll keep doing it the way i am for now and ill make a channel for donation if someone thinks they'd want to support me that way.
Now its time for every 1.0-13.7 plane em diagram
💀
You're gonna kill the guy faster than the snail
maybe with crowd-sourced data it could be sped up somewhat
@Jay-kc2pm maybe but idk how quality control would work then
@@CatWerfer Btw did you see statshark? On that website you can generate EM diagrams for every plane in wt. But I don't know how accurate those are in relation to yours
This is so cool man. I didn't even realize the f4 was better than the g2, or that the spit Vb was so much better than the mk. IX. Thanks for all your hard work
yeah i was surprised too, apparently some people were aware but i guess i just dont interact with them a lot
Early Bf 109 G models were hamstrung by a performance restriction on the DB605. The intake manifold pressure was limited to 1.3 ata due to reliability issues early in the DB605's life. As a result, the DB605 in early G models produced around 1300 hp, around the same of the DB601E powering the F-4. The G was heavier than the F, so flight performance of the early Gs was worse than the Fs. In 1943 this restriction was lifted, so maximum power was increased to 1500 hp at 1.42 ata. Unfortunately War Thunder does not have the flight models with these specs.
The Spitfire Mk.9 is far superior to the Mk.5. However this reality is not modelled in War Thunder because the Mk.9 in the game uses the wrong specs. The Merlin 61 should be producing around 1560 hp in its most powerful setting, but in the game it's modelled as 1400 hp. Problem is, that 1400 hp is its rated power at 7000 m altitude whilst in its first supercharger speed (aka the speed for low altitude performance).
hmmmm is this not just a difference between the f and the lf mk9?
@@CatWerfer It's confusing. The F Mk.9 started out with the Merlin 61, but late-model F Mk.9s got the Merlin 63 which could produce 1710 hp. The LF Mk.9 got the Merlin 66, which was basically a low-alt version of the 63. Produced about the same power, but at a lower altitude.
We gotta get Defyn to weigh in on this. I think he could give some valuable insights into what exactly you can deduce from these EM graphs
I don't know what background defyn has in aerodynamics or physics, he always kinda struck me as a "vibes-based" player where he sorta just played planes and gave advice without evidence beyond his own experience (which isn't a bad thing since there really isn't an easy way to gather airplane data outside of just playing the game),
but if he has that kind of a background and has interesting things to say about em diagrams then sure
Out of curiosity, what's your background? Military or aeronautical engineering?
studying mechanical engineering (undergrad).
I don't have an aero background myself but I've done a fair bit of reading about this in my own time.
People more well versed in aero come around and correct me when im wrong so it's nice there are people who can help me convey accurate information in that regard
@@CatWerfer nice, I did too but dropped out. Takes a fucking work ethic
yeah :( meche isn't quite what I thought it would be
Bro this is awesome! You're an absolute legend for doing this. Would love to see some charts for La-5 variants in this BR range. I know you're burned out but I think the next data points to consider would be like 3000m as that 'mid-range engagement' that happens at the start of a match. Again, thank you so much for what you've already done. This is such awesome info. Subbed and shared on my channel.
I certainly would like to do altitude performance tests as well but frankly speaking I don't think I want to spend time comparing planes at multiple altitudes. I know that it's important and some planes gain a significant advantage at altitude over others, but its pretty time consuming so when I do make new graphs, it'll probably be on different planes.
My code and methodology are all shared in the drive and previous videos so if you want you can test for yourself too, though it can be pretty painful.
La5/7 and other planes at other brs coming not-so-soon(tm)
Holy moly, this video came out right when I bought the A6M3 to train my dive bomber pilot
Thanks a lot!
wdym finally a useful video?! They were all useful and highly informative and entertaining!
lol thanks for the complement but I don't think the average war thunder player would sit through a 30min video about data analysis like i did in pt 2.
(though i guess the average wt player would also probably not be that interested in em diagrams either)
@@CatWerfer hey now! Don't just destroy my limited faith in humanity like that! You mean to tell me the average WT pilot is an unpaid actor actively trying to troll their own team?!? Inconceivable!
Looking forward to future videos. Maybe you can figure out what the rocket crosshair is supposed to represent :D
im pretty sure the rocket crosshair is just tied to nose direction and maybe the angle the rocket is mounted to the wing at.
maybe a more interesting to look into is to try and see how radar lead indicator and arcade lead indicator works
12:39 - Yak-3 has a smaller, more aerodynamic airframe than Yak-9s. A smooth mini Yak-9. Still, a great video. 30 planes, damn, that's lot's of effort!
I was suprised by energy advantage of F4U-4 vs Bf 109F4 at most speeds. Tho it makes sense cause you tested F4U-4 at low altitude where it has 2470hp, while it already drops to 2100 hp at 1.7km. 109F4 gains big power advantage in at 2km compared to sea level.
huh, the airframe limits in air rb are basically identical for the em plot, i guess I shouldve done a bit more research instead of assuming the airframe is the same
F4U-4 is carried heavily by the methanol-water injection (limited WEP). For the altitude performance drop though, is that with manual supercharger gear? I found F8Fs and F4Us on AEC have delayed supercharger gear switch compared to what the manuals suggest for fast-scrambles.
@@CatWerferthey rip the same but the 3 os noticebly lighter
I wonder why gaijin is so cagey with their in-game aircraft statistics and how things in game actually work? Why would they want to keep players in the dark on these matters?
likely they're wary that people might try to develop a competitor to war thunder using similar aero methods. they have patents on their instructor mouse aim after all...
@@CatWerfer gajins godawful spaghetti coding is an acceptable price to pay for the code's true purpose: anti-theft
@@fablearchitect7645 lol its so bad noone would steal it
@@CatWerfer That patent also doesn’t go into detail about how the mouse aim works. It does not show any control loops, so the exact method is not protected unlike with other engineering patents that show and claim rights on the exact method used. For example, many engineering patents claim rights on a specific circuit topology used.
@@fablearchitect7645 as far as I know, software patents don't need to go into the actual details of how it works, just what it's supposed to do. It's the reason why Sega successfully sued Fox and EA in the early 2000's for patent infringement due to the game "The Simpsons: Road Rage" using an arrow to point towards objectives, which functioned the same as the arrow in "Crazy Taxi" despite the code likely being different.
Also if for nothing else, optimal climb speed is super helpful, i didn't realize so many planes like to climb above 300 IAS
yeah, though keep in mind that that is the ideal climb speed at sea level and the ideal climb ias drops as altitude increases.
Also theres a pretty wide range for the plateau of max SEP so it may not be exactly that number because of the curve fit (check out my "optimizing climb" video if you haven't)
Very impressive and thanks for the effort. Now its more clear to me why most of the German planes tactic to boom and zoom is so important but also wont work in many situation.
This is video I was waiting for!!! Thank you!
I got lost like 2 mins into the video but I stayed for the big brain presentation
@@roywenton1605 it might be easier if you watch the whole series, i kind of build up to this point in terms of explanation, starting simple
Great work! Rare to see some Em diagrams comparison for war thunder. Now we need some 5.0-6.0 BR palne comparion such as J2M2, Ki84, Spitfire LF Mk IX,Yak3U, Ta152H and Bf109K4.
maybe eventually but not now
@@CatWerferall right. Youre a legend m8
very excellent video
This was great, thanks man
I like the ITP. It can pull something like 11g or 12g turns. The guns make it quite deadly, but outside of a dive it is to be used carefully, since it is not the best climber or accelerator.
@@Orodreth888 ok buddy
I like transparency with some flaws of these diagrams. Still a great resource for anyone who wants to use a few neurons to learn from these diagrams, there's noting like this that's available.
yeah i think i make sure to make disclaimers pretty adequately for my videos.
data isn't useful if its mixed in with garbage so i gotta make sure people know what is and isnt high confidence in my data
Thanks. You did great
The Spit IX is truly the king of Rank III. There is almost nothing it can't do at least reasonably well at almost all altitudes :D
@darthcalanil5333 yeah, though I'd play the 5b more if it was rank 3
it cannot dodge these hands when it goes headon, the sheer size of its wing is the major success and downfall of the spitfire.
@@Akhaaelikian hispanos are really good sniping guns tho, you rarely lose headons if you know what you're doing
109 F-4 is king imo
Awesome work! Really great visualizations. Now you just need to extend it to supersonics… :)
i'm probably only gonna do subsonics if i even do jets, i have a feeling that the hexic curve fit will start to become not good enough once i have to account for wave drag...
@@CatWerfer the wave drag is the largest drag component at supersonic speeds, but typically the drag force will be approximately proportional to the sears haack body drag which is proportional to dynamic pressure and max crossectional area to length ratio squared. The way it shakes out though that’s basically just a large drag term proportional to mach squared coming online though the transonic region. I think a 6th order fit should be able to handle that alright, but if not there are some fairly simple things you could do that don’t assume an underlying form like a smoothed spline or k nearest neighbor model
@youngbloodbear9662 hmm ill think about it
30 aircraft?!? get some sleep, damn. Excellent work, really.
These graphs make it apparent the Pyorremyrsky, XP-50, Wyvern, Yak-3 and maybe P-51 D10 are in their own league, which anyone playing this bracket already knew (tldr: p2w confirmed)
The P-63 seems excellent too, but this is only at sea level
Maybe the color gradient could show the relative difference in MER instead. This would make advantages in stall-fights and climb-rate more apparent, which are quite relevant to the meta
even more p2w when you consider the eremin yak3 with one less .50 at 4.0 💀
Very interesting vidéo
Interesting to see the A6M3's current EM diagram. Per US TIAU no A6M they tested could turn inside of a P-47 at 300mph IAS. Also per Japanese reports the Ki-43 could always pull more Gs than any A6M at high speeds. This was a deliberate decision by Mitsubishi for the very reason that came to plague early Ki-43s giving them an early bad reputation in the IJA. Pilots were ripping the wings of early Ki-43s trying to pull out of high speed dives. Mitsubishi was concerned about this during development of the A6M (because of the insane performance demands of the IJN) so restricted how much elevator could be pulled via balancing the elevator for low speed. This resulted in elevator compression earlier and heavier than the Ki-43.
If the A6Ms ever get this elevator nerf in game they are going to handle like trash above 300mph IAS. Also this elevator nerf would make the historical get away tactic for US pilots against the A6M actually viable to use in game. Dive to 400mph IAS, roll to the right and then pull out of the dive.
huh that's pretty interesting, thanks for sharing. I don't know very much about aircraft flight performance IRL and am really just modelling things ingame, but its cool to think about how they could implement that.
Do you think it's likely they'll change that? there are a lot of inaccuracies that war thunder has like not using the G limiters on f16s that are there since the pilots in war thunder are kind of superhuman with their g tolerance.
@@CatWerfer They have been implementing historical elevator compression randomly to a number of airplanes since the F-16 was added. The F4U-4s and 7 were the first. The MiG-23s and 27s were maybe the most famous recent example. So will they? No idea. I think it would be nice because elevator compression has been the one major performance aspect that Gaijin has used for balance rather than sticking to historical documentation.
@@rainsilent huh i guess someone needs to send a bug report in
@@CatWerfer If you do throw in a bug report on the top speed of the J2M2, 3 and 5 while you are at it. Their in game top speed is based on Japanese top speed data when nearly every other major Japanese aircraft of ww2 in game uses US top speed data. Why does this matter? Japan didn't use WEP to collect top speed data. The US did. The J2M2 and 3 topped out around 405mph and the J2M5 topped out around 425mph.
I didn't include the J2M4 because the US never tested it. It's actual top speed is a historical unknown.
If someone's gonna bugreport it, it probably won't be me cuz idk much about irl flight performance and i also don't know what sources to look up for that kind of testing info. maybe you could make it since you sound very knowledgeable on this topic?
You should definitely make this into a playlist. A bit problematic to go from one to the other without it being organized.
ok i will make it right now
@@CatWerfer Thank you, my good sir!
we need this for top tier fighters, i want to see if there is any regime where my Tornado can do something against anything
lol i dont think you need a graph to show that the tornado is a shitbrick
god YES I wish I could have done the VO for this video
VO? voiceover? why so?
It should be in stat cards
You want to shove an entire EM diagram into the stat cards?
@@MrSeon123 It would be cool if you could compare aircraft in the game
that would make the statcard actually useful lol.
maybe they could add a comparison mode like they did with shell penetration for tanks
Someone should make a suggestion about this in the forums
@toki_tsu_kaze5511 gaijin is notoriously cagey with their in-game aircraft statistics and how things in game actually work so it's not likely but maybe
hey, some air frames lose speed far more quickly when turning. How to interpret that in these graphs?
8:24 - I point out that at speeds above the sustained turn, the SEP gradient (color of graph) is deep red along the top edge (max turn) meaning that while the f6f does turn better (blue outline is farther up than the red), the 109 has a better maneuvering energy retention (loses less speed in a turn).
14:08 - The opposite is true with the typhoon, since the top edge has a blue gradient, meaning the typhoon has a higher SEP, loses less speed in a turn, and so has a better maneuvering energy retention
@@CatWerfer went through your older vids. now i understand whats happening for somewhat... hehe... need to think about this..... thanks man!!!
@catwerfer Nice work! Wish Gajin would put EM diagrams in the game, definitely makes mastering a specific aircraft a science, and gives you an idea of where you need to be when picking a fight. If you did put these behind a paywall, I would gladly subscribe to your patreon.
Also consider videos where you discuss the EM chart for a specific aircraft while in a match and demonstrate where those advantages and gradients are. If you could animate one to show a dot where you and an enemy aircraft are on the corresponding charts, you could pinpoint where a fight turns in favor of the victor.
Would love to fly with you sometime maybe a stream? I’m an EE in the aerospace industry, but I also fly my own yak 52 :)
lol im too much of an introvert to have considered streaming, I dont know what i'd do to make war thunder gameplay entertaining to watch.
Thanks for the complements! I am actually working on a video on how to compare EM diagrams a bit more indepth cuz this one i kinda glossed over it quickly. I laid out the foundations in my part 1 and 1.5 videos but i need to do a better job of explaining how the 2-plane plot works.
Did you tested Bf109G2/trop or normal G-2(unavailable in germany, only available in italy and sweden)?
the trop has more drag due to the sand filter on the air intake, so it performs sligthly worse
Also worth nothing that the DB605 has better cooling compared to the DB601, so the G-2 can use WEP more often than the F-4, something that can make the difference in a fight
Will be cool to see the 3.7 graph and how the C205 serie 1 murders the E/M graph of everyother plane
or the 5.3-6.3 superprops graph will be very interesting to see
@@toki_tsu_kaze5511 i wanted to minimize redundant testing due to how long it takes so no I didn't test the trop vs non trop
Wow this is super useful, i already fly the 109 F-4 a lot. Saving this to study again soon!
holy shit he is back. i tought we lost him to cp
Lost him to what??? 😧😧😧😳😳😳
the club penguin mafia really thought they could get me but now i have relocated to a remote island with a bunch of gigarich people where the government cannot monitor my activities
@@CatWerfer not malta!
I hope you do it for 4.7-5.0 or 3.3-3.7 next time, very fun and useful diagrams!
eventually, when i get the stomach for testflying again
Maybe you could create a detailed manual on how you tested and gathered your data. Then create a public google document or something and everyone can test his favourite planes and share the data. With a few hundred people doing that you should be getting good averaged data.
@@robzinger did u watch part2
Super cool! I've been working with Alpakinator to get calculations for top speed charts for all planes. Do you have a discord?
i dont have my own channel but i lurk around in adam's server
Do you have any experience with sim mode? Was curious if there is any performance/physics differences compared to RB.
sim mode basically removes the instructor
the instructor does a few things:
- translates mouse movement into control surface inputs to make your aircraft's nose point to the mouse cursor
- limits aircraft turn by AoA so the plane doesn't stall
The base physics of the aircraft stay the same, but without the instructor, you're able to turn harder but also need to make sure you don't pull too hard and stall your plane.
you also have to manage control surface inputs and trim yourself (until you get to fly by wire and jets with computers helping you fly).
That's why a skilled sim pilot can outturn someone using mouse aim in RB.
Should have included xp-55, it's one of the best 4.3 fighters.
@@LaminarTurbulence i forgor i will do it next time
Why not do a Ki-44-II comparison with the 4.0 and 4.3s too? Doesnt ut have the performance of a 5.0 ir 5.3 down in 3.7?
Its performance no reach 4.7
i compared 4.0/4.3 planes, not 4.0/4.3 flight performance planes
@@CatWerfer fair, still it would be interesting to see just how ridiculous it really is.
@@lonemarkkingoftypos3722 It is a 4.3 maybe 4.7 plane, not a 5.0 or 5.3. It is close overall to the Pyro. A bit slower, better climb sub 2km, worse climb the higher above 3km you go, better roll and better turn. Absolutely cracked sub 2km for 4.3 even if a bit slow. Good 2km-5km. Don't take it 6km and up. That said it could end up 5.0 because bad player base + turn fighter syndrome.
Could you compare the early war plane like the bf-109 e3, spirfire mk1 and d520 ?
I feel like the lowest I'll go is 3.7 maybe because below that br, the dogfights aren't that difficult to win since most players are newer
Can you add Lavochkins to the graph?
eventually but not in the immediate future
Can you do a EM for the IL-10 (1946) or the IL-28?
@MaxerGamex ill do thr il10 eventually but the 28 might be hard
@@CatWerfer Thanks man, don't worry about the IL 28, im fine with only the IL10, take your time and take care
I am studying physics and I love this. I just wish War thunder would give us more stats
yeah, but they are notoriously cagey with how their physics engine works so if we do figure out stuff its gotta be on our own
@@CatWerfer Yes, sadly. Thanks for all the effort you put into these videos. A friend recommended me this video today and now I'm subscribed to the channel. Good work and I'm looking forward to seing more of these videos
Ki83 & Hornetmk3 comparison plz?
someday but not so soon(tm)
You can insert a comparison of p51 vs Russian and German planes
are you making a request? and for the p51d5?
@@CatWerfer yes, compere to yak3 and yak9
New folder added called "P-51D-5 VS 2 Plane Plot", ive added the graphs there if you want to see the comparisons
13.7 when
I won't be doing high thrust jets any time soon because theyre a lot harder to test (you can watch the part 2 vid to see why)
YEEEEEEEEEEEEEEES, NO MATH!!!!!
YEAAHHH BOIIIIIIIIII GRAPHHHHH THUNDERRRRRRRR
Can you please do Yak-9U vs Bf 109 G-6 at 4.7?
if there are graphs for 2 planes on a single plot you'd like me to do for planes I've shown in this video i can make that but for new planes I would have to go in and manually test them and i am not going to do that for now
I understand, that's really hard and time-consuming work. It is just a suggestion of two of my favourite aicraft)
Thank you for this amazing info, nonetheless!
@@DimaMuskind yeah, itll be happening, just not in the immediate future
5:34 20.5m/s to 21.9m/s max SEP after breaking the flaps on the P-51D-5
Got it, I will make sure to rip my flaps off every match before climbing
as he said, that's measurement error. I tested this in the past and theres no differnce in climb, or turn.
lol i think he's joking (please do not rip off your flaps before climbing)
Dude, instead of doing them all in sheets, see if you can make a site or comparison tool or something (I can't rlly help with something like that sadly)
i don't know how to do that, maybe eventually but this is what i got for now
Using the Zero feels like cheating sometimes lol, unless you're tryina chase someone
it can be kinda fun sometimes since very often the enemy will try to take you on in a 2v1, but the a6m2/3 is also pretty damn good at juggling multiple people if they try to turn with u
Yak9t/k are diffrent to the 3 airframe wise yak3 is stubbier if i remember
yeah i saw that the profile was the exact same and so thought it was identical but i guess thats just the instructor limiter
@@CatWerfer yak 3 is lighter thats the only diffrence basically
the yak 3 got some kind of upgraded fuel mixture iirc, something like 100 octane fuel that the yak9t/k didnt get tho so i assumed it was some kind of engine related upgrade
@@CatWerfer The engine is better but thats because the t/k are the 3.0 yak9
The Yak-9K and 9T by no means have the same airframe as the yak-3. The yak-9 airframe has quite a lot of differences, most notably larger wings and more weight
in air rb they behave the same and have the same airframe limits
Please guys subscribe because that's really good and hard work here, you Sir deserve more followers 🙂
No math = Me sad
@rimantasginiunas1848 math in next video :]
This should help in 4.0-4.3
maybe eventually but not for another week or so, testing is annoying
lmao use the aria math and its consequences in the next video
@@BONGUS1-g7t lol maybe
My only nitpick is that you picked the worst A6M3 instead of the more maneuverable Model 22 at 4.3. Those clipped wings just weren't it.
Overall this is mostly unsurprising, though just from looking at these EM graphs I have a hard time keeping up with your conclusions.
Btw aren't you glad that there's now Eminem's Yak3 at 4.0 with that kinda performance? Totally balanced!
Also, A7M2 needs to be on this. It's an incredibly maneuverable plane that will laugh at any Spitfire that decides to turnfight it. Ki-27 would be funny too because IIRC it has the single highest sustained turn rate of any actual fighter, only beaten by a min fuel BI - coupled with a low rip speed I assume it's EM graph is gonna look more like a fence post than a doghouse.
Really? I like the base a6m3 much more than the others, it already outturns everything like crazy so the extra turn you get from having unclipped wings really isnt worth it imo.
If its difficult to read the em diagrams, i might make a video explaining how to read a 2plane plot separately.
The conclusions weren't really supposed to be surprising, just a more concrete way of describing what we already know from playing the game.
A lot of wt commentators will list off a bunch of characteristics about aircraft without any hard facts to point to. They aren't wrong, but its hard to verify without just playing the game a lot
Math 🤪🤪
Friendship ended with math thunder, now graph thunder is my best friend
@@CatWerferbtw, do you have these graphs in a google doc or spreadsheet? I think they could be very useful for helping my squadron learn flight characteristics at this br.
as I've said, theyre in the shared google drive folder in the description. lmk if you have trouble accessing them
I am an IDIOT nevermind that
lol we can delete these comments and pretend this discussion never happened
109f4 actually beats xp50 if the f4 plays aggressively.
@@ProudOne ok buddy
@@CatWerfer obviously sarcasm, bad tone which i do not appriciate, but yes it does, just kinda spam the flaps and force a stall fight ASAP and the f4 is the superior plane. if the fight is prolonged the xp50 will win, but f4 is the better flight module
@@ProudOne the 109 only wins in the lowspeed stallfight if you look at the graphs.
If you can force it into a lowspeed stallfight then you have a solid chance of winning, but in basically every other scenario (airspawn, headon, horizontal ratefight, scissors, chase), the xp50 is better.
If you can consistently win against the average XP50 player in air rb thats all well and good, but part of that is because most xp50 players aren't that good and dont use throttle control+flaps themselves.
Put two competent players who know their planes and their advatanges, one in an xp50 and one in a 109 and the xp50 can see you're forcing a stallfight and use its better linear energy retention to disengage and force a fight that's more favorable to it.
Do you see people complaining about the xp50? yes
Do you see people complaining about the 109? no
I think your argument is pretty bad faith so i responded in kind
@@CatWerfer no it was not in bad faith, and you’re still wrong. The f4 is the better flight model, although it is a good match up. I don’t base this off of beating them consistently, but rather one of the best prop pilots in the game if not the best says this and has explained why; with actual fights against another very competent pilot.
what a nerd, just hold turn key and crash if neccesary to deny the kill to the enemy
mmm yessss the very gigachad and definitely not whiny baby move of crashing to deny the kill so that both you and your opponent who is playing the same painfully grindy game lose together and the snail wins.
it would've been funny if i made a video talking about technical stuff and then i did a psych-out at the end where i just tell everyone to crash lol
@@CatWerfer nah like my boi leonidas said “Give them nothing, but take from them everything.”
@@antonisashitteragain6993 "an eye for an eye means everyone gets to eat eyeballs, mmmm yummmm"
@@antonisashitteragain6993 Or you could play smarter, like Leonidas meant, rather than harder and force them to use as many of their resources to kill you as possible a la just like what Leonidas did at Thermopylae. Most props around this BR have less than 20 seconds of total firing time. Getting them to expend as much ammo as possible can be a difference maker in a match. Likewise getting the opponent to burn as much energy as possible by dragging the dogfight out as long as possible makes them vulnerable for longer to any teammate that might come along.
Nothing is better for me than an opposing player crashing if I so much as get near them. I don't have to expend ammo, fuel, effort or energy to get you out of the match beyond just flying in your general direction. I get to move on to taking out your teammates with the least amount of effort and resources expended possible on you.
Quoting Leonidas but then not following his example. Nothing could scream fake poser any louder than that.
@@rainsilent it's not about getting killed or one member of the team less is about if they have not touch me negate them any SL or RP :)