Asgards vs. Stations: X4 7.0 AI Test
ฝัง
- เผยแพร่เมื่อ 28 มิ.ย. 2024
- The mighty Asgard is the uncontested king of the cosmos...as long as you don't look too closely. We will be looking very closely.
00:00 Background
02:11 Test Design
04:22 Test vs. Wharf
08:18 Test vs. Defense Station
12:32 Results & Discussion
15:15 Conclusion
![X4 Foundations: Fleet AI vs. the Endgame Crisis [Beta 7.0]](http://i.ytimg.com/vi/Qp4Xb2cPP8s/mqdefault.jpg)
![X4 Foundations: Hijacking a Xenon H! [Beta 7.0]](http://i.ytimg.com/vi/NuMwUAY_eiA/mqdefault.jpg)
For those interested, here’s how I arrived at my conclusion about the nature of the Asgard:
Egosoft has created a universe in which battles take place at certain ranges and velocities, all ships obey certain rules of physics, and the *core* AI handles those conditions in a certain way. Under those constraints, the Asgard does not come anywhere near its on-paper potential in high-attention combat. There are two possible reasons for that:
A. Egosoft understands their universe and the constraints it imposes on XL ships, and they designed the Asgard knowing its capabilities in high-attention combat would be severely limited.
B. Egosoft does not understand how XL ships operate in their universe and believed the Asgard would be able to apply its full theoretical power in high-attention combat.
Based on the very practical designs of the other XL direct combat ships (K, I, Erlking), I think option A is much more likely here. Which then begs the question, why did Egosoft knowingly implement a ship with such a huge gap between its theoretical power and its practical power?
A. For metagame reasons.
B. For lore reasons.
The argument for option A goes something like: “mimicking the M8 Bombers of X3, Egosoft implemented the Asgard in X4 as an ‘I win button’, but they wanted to restrict its power mostly to low-attention combat”. You’ve heard the argument for option B, but I will add that it also has precedent: the Phoenix in X3, among others.
After thinking through that line of reasoning and considering both of the last two options, I felt option B was both cooler and better supported by the evidence. Do you agree? Do you have a different theory? Let me know!
The lore reason falls apart because they actually build and use these ships.
A ineffective combat ship that is only scary on paper is only scary as long as it remains a theoretical threat, the moment it is put into actual USE the theoretical danger of a laser THIS BIIIIIIIIIG! is lost and even worse, the RnD time and cost, build time and cost is also wasted because now the Terrans are sitting on big expensive paper weights that no one is scared of.
There is also the whole thing of them being able to actually build these functional ships but not deciding to build a functional design... why would they not?
Even if the laser had to be scaled down it would be infinitely better to do that if they are going to be building and using them.
As it is now you are basically saying the Terrans are knowingly building and using ships that are terrible for the reason that on paper they are scary but it defeats the purpose to do so, they are not scary once everyone else knows how ineffective they are.
Thank you for your sacrificed time testing things for the video though, I liked it! 🙃
Well, since you asked: all my fleets have 1 Asgard commander with 4 Asgards on Follow. Each one of the latter has 4 destroyer subordinates on Attack. What I do is I pilot the commander Asgard up to the station to a safety distance, so that then I can use my position as a point of reference in the map. I then manually tell the other Asgards to park around the station at a similar distance to mine on the map. Then just order attack! I've never had a loss and they're pretty quick in getting down to business. I've always used this strategy in X games and I never really lose ships.
Enjoying these research videos by the way!
behold! The Fleet AI result after 6 years of release!
Must admit, I'm impressed with the improved AI. A decent destroyer captain will very rarely blunder weapons range, at least with Xenon stations. And the destroyers that do, well, their sacrifice is noted and the names of the crews placed on the Wall of Remembrance at Storm Industries HQ
Since updating the ai is night and day, it's significantly better to the point I've really noticed
I see Asgard, I upvote.
The context you add to the Lore of the ship is really insightful. Great analysis and completely agree. OOS you can use an Asgard to clear a sector of stations, as soon as you're in sector the DPS and IQ of the ship drop. Maybe they're just shy.
The Asgard is the ultimate Hero Ship.
The only way to use it effectively is fly one by yourself and let one or two follow you.
Shoot the main battery, teleport, shoot the main battery of the second ship, teleport, repeat till everything is done.
agreed, I see the asgard as the flagship of a fleet and intended to be piloted by player to be used properly. with appropriate upgrades on the main gun, you can get much faster reload. also helps clean up K and I from a sector easy (when used by player) before eliminating all stations. By the way could be interesting to investigate what are the best upgrades for optimal asgard main gun use.
Why only ~400 subs dude???? Your content is awesome and duper informative! Thank you! Please keep it up 😊
Thank you for the informative video! This must have took some editing, and the result shows.
Thank you for the video and taking feedback into consideration! I have no complaints about this one :)
Good enough for me. Thanks for the Asgard science !
In all my fleets I always have an Asgard, he is always the captain ship of the fleet, the Asgard is very frustrating but I still love seeing his laser fire and for me it is one of the most beautiful ships in the game.
I like the lore you have built around the reasoning, but I don't agree on the technical aspect. A fixed gun on a spaceship is actually a very practical idea. You can't compare it to a battleship on the sea; space operates under different principles. At the scale of the Asgard's weapon, it makes even more sense.
Consider momentum: rotating a massive gun on a floating object in space would require incredibly strong gyros and RCS or an even more massive ship to carry it. Otherwise, the gun would rotate the ship instead of itself. It's far more efficient to build the spaceship around the gun and rotate the entire craft.
Take the Hubble Telescope as an example. It's essentially a bus-sized primary mirror with solar panels and thrusters strapped on. The Asgard is built on the same principle but designed to spit light instead of collecting it.
I agree with the physics there, but I think there's a big difference between "giant space lasers are best deployed as fixed weapons on XL hulls" and "XL hulls built around giant space lasers are practical weapons in the X Universe". Egosoft has made some clear design decisions around engagement ranges, transversal velocities, and the ability of even 5-star pilots to track targets at those ranges and velocities, and the Asgard just isn't a good fit for that combat environment. My comment was meant to convey that "fixed weapons on battleships in the X Universe are impractical for X Universe reasons, in a similar way to fixed weapons on real world battleships being impractical for real world reasons." But I appreciate you taking the time to put a real-world lens on it!
@@CptSnuggles07 Well, from a game design perspective, the Asgard isn't made to track fast moving targets with its main guns. I would rather argue that the AI code simply still is incapable handling the ship.
If you use CUDA AI tweaks, the Asgard actually becomes an effective tool and can kill K's with one hit, occasionally even directly out of travel drive. There have been issues with all capital ships all the time. Common destroyers have been improved but the Asgard is a unique ship and simply does not have combat routines and priorities to make it effective in vanilla. For instance, there is no priority to target capital ships, so the Asgard might be sitting there trying to target a light fighter while a K is approaching and it will not change targets.
Another issue is that the code for capital ships prevents them actually moving in 3D space properly, they always try to stick to the ecliptic and that results them being incredible dumb when attacking stations. This is also a thing improved in CUDA allowing them to take positions with a steep angle above or below the station.
I think Egosoft has teaked their settings in that regards a bit, but its still not there yet.
@@madrooky1398 I don't like "just make the core AI better" arguments because if the offensive scripts get improved, the defensive/evasive scripts should be too, and then we're back to square one. But we can agree to disagree. Thanks for the thoughtful comments!
Also, the "max pitch" behavior has been greatly altered in v7.0. Ships still prefer the ecliptic but will maneuver and shoot at very steep angles much more readily. Documentation in my destroyer video, and in the thumbnail of this video. A welcome change!
So at least for me, I will always use the Asguard for In sector combat, mainly because I will be the one flying it. I have an AI captain, but he's just there to chauffeur the ship around while I attend to other business. As for how I use it, that is simple, I park between 11-14 km out from the station, preferably above it, and I use the main cannon to destory defense segments. I do this in conjunction with a destroyer assault to basically minimize the damage to the destroyers if one misrolls or move out of line.
Nice😊
Suspect operative difference is Asgard main gun being completely fixed. Suggest disabling main gun to confirm. AI should adjust distance and positioning accordingly.
That is very possible! I'll add it to the "misc testing" list.
The Asgard parking below station is probably because you forgot to untick the option "attack weak targets first" in the coordinated attack options. This option is enabled by default and makes your ships chase subordinates of the station, drones for example
I deliberately left the C Atk settings on default for testing purposes, and the defense stations didn't deploy drones, but yes, it is possible that that setting contributed to the strange (but consistent) ship movement. Check out my Destroyer Behavior video for a more in-depth look at the effects of different C Atk settings!
@@CptSnuggles07 It's just weird because my ships would never fly into a station like this, if I used a coordinated attack order. But also, they seem to have changed the way it works in 7.0.
are u also setting the L XL ships to bombard if u are using a coordinated attack ?
No, they're always kept in the same configuration I show early in the video. Ships set to Bombard will only attack capital ships. They ignore stations. If they are included in a Coordinated Attack, and there are no valid Bombard targets within range, the fleet commander usually assigns them a default Attack role instead.
I've watched a few, but not all of the 7.00 videos and I didn't catch you testing bombard orders for station attacks. Aren't those meant for this use case?
Bombard is to capitals what Intercept is to fighters. Ships assigned to Bombard only attack capital ships and will completely ignore stations. Very poor word choice on Egosoft's part.
Asgard's only drawback is that it's too strong. You don't need L-plasma, you need Terran L-pulse and M-Flak because you want to effectively destroy M and S ships. The Asgardian AI is great at destroying stations while you're away from the scene.
I seriously think there should only ever be 1 of these in your entire fleet... as the players ship. And should only ever be given to the AI when you can't be bothered flying from a to b. They fly it so bad. A destroyer/fighter combo does everything better than an Asgard. In the players hands though, it's an unstoppable beast.
Agree, if the AI also has limit of 1 TER Intervention Asgard which it supposedly had at least at some point.
I haven't touched X4 since 6.0, apart from buying Timelines and playing through it's "story." It's good to see that AI Asgards don't ram Xenon stations anymore, dying in ten seconds, making me delete the game.
Ok so ive been playing the same save since i bought it at cradle so while i played kingdom end, not much large scale combat due to being on good terms with the surviving factions(xenon are locked in a corner, and patriarchs decided not to hold a grudge for almost exterminating them(ok so i shot criminals)) my question is when did the ai stop consistently tring to lick the windows of whatever station you told them to shoot. Also this is the first time ive heard asgard criticism not aimed at it being an overpowered i win button. I mean if the ai can just about use that gun now i dont see how its a bad design(in universe) it rides at a crawl but unless the enemy gets multiple heavy warships to it at the same time they all die the second they get in range. Still nice to see a different opinion then the norm. I mean i love it but i know aomthing like would be aweful in any kinda realistic setting
The window licking has gotten a little better with every major patch, especially 6.0 (new physics engine) and now 7.0 (big focus on AI improvements).
When I first started designing this test, I was actually operating on the standard assumption that "Egosoft gave us the Asgard as an I win button because that's what they always do". But the more I thought critically about that, the more I realized that it doesn't tell the whole story (although it may be partially true). So, I went back to square one and tried to think objectively about "what is this ship and why does it exist", which then turned into "in a universe where most pilots struggle to hit moving targets with gimballed weapons on relatively agile destroyers, why would anyone put a fixed weapon on a battleship?" After going deep down the rabbit hole of in-universe lore and Egosoft's game design philosophy, this is what I came up with.
The ships need to be at almost the same level as their target. Otherwhise they cant hit it. Because the ships are limited on how much up and down they can aim for some reason. Its weird.
They've actually corrected the "max pitch" behavior in 7.0. That's shown several times in my destroyer test video and in the thumbnail for this video, which shows an Asgard shooting straight up (I flipped the image upside down for formatting reasons).
The problem with the asgard is more to do with the AI. capital ships are not hard to hit.
That is really disappointing. I still think that this odd behavior is related to the way you configure your L and M weapons. You're using new options for the weapons, do you know exactly how they work?
Not exactly. Based on what I've seen, both Capitals First and Missiles First choose targets exactly how you'd expect based on the names, but I haven't seen a good guide or done any really thorough testing of my own.
Because the destroyer tests showed identical ships displaying different behavior in/out of fleet and under different fleet orders, I think it's most likely that the odd behaviors are due to the fleets and orders, not to the individual ship setups. It's possible that there's some odd interaction between fleet orders and turret commands, but that doesn't seem like the most likely cause.
I did take note of your turret command notes on my destroyer video and have the comment marked down on my "miscellaneous tests" list, but I have no idea if/when I'll get to everything on that list. If you do any testing of your own, let me know what you find out!
The Average skill level of combat ships influences the maximum range the ship will use for the combined average of all its weapons
5 Star is 100% or maximum weapon range
4 Star is 90%
3 Star is 80%
2 Star is 70%
1 Star is 60%
0 Star is 50% of Weapon range
Average skill level also determines how mobile the ship is meaning they will get in position faster in the attack commands and between modules (It also affects miners to a massive degree on Silicon Specifically)
This is worth keeping in mind
Do you have a source for this? It doesn't match the datamined scripts we have from v6.2.
Awesome testing. Disappointing IA 🙂