In addition, I have been testing coh2 recently. I finally found the reason why the archive capacity of coh2 will grow larger and larger for no reason.The previous RTS game save system only saved the present situation and did not save the battle process. This is good. In this way, you can play a game as long as you want. AI will not become dull and stupid and bug due to the larger and larger archive. But when I play COH2, I find that the save mechanism of COH2 is actually the save the process of the battle. Why do I say so? Because when I play a game of battle and save, then quit the game, turn off the computer, and then play again next time, I find that AI becomes stupid and bugs appear earlier and earlier. So I did an experiment: when playing the same game of battle, I played for 20 minutes every time, then saved, then shut down the computer, let the computer rest for 1 hour, then read the archive, and then continue to play for 3 minutes, and deliberately let my troops be destroyed by the enemy, and a large number of them die together with the enemy, which reduced the number of enemy and our units on the battlefield, I did this to reduce the number of enemy and our units on the battlefield before saving, so as to eliminate the possibility of increasing the archive capacity caused by too many enemy and our units.And I was surprised to find that the archive capacity I played only for 3 minutes with a large reduce of unit number,was 2600 KB larger than the archive capacity I played for 20 minutes before, so I understand, The archiving of COH2 is not save the end of the battle, but the whole process of battle!This is very bad. First of all, the archive capacity is very large, usually more than 60000kb. In the later stage of the battle, it will cause AI to become stupid and bug. The game is more and more easy to jump out and other strange phenomena and Bugs will happen, which will destroy the game experience. I hope this bad save mechanism can be changed in COH3,the save mechanism shouldn't save the battle process, but save the battle present situation.
I think this stems from how the replay feature works. The game does not save any state of the units, but player inputs etc and based on this simulates the battle how you experienced it.
I assume since Relic wants to also have the replay for Bot games, they just used the same mechanism for save games. Otherwise you'd need to keep both the replay file as well as the save game in seperate files. Which is possible, but somehow double the work.
I don't fully understand yet though how this should influence the AI. Do you have any ideas? The game itself might become buggy, yes, but not fully sure how this influences the AI, especially after rather short games.
Could you also elaborate what you mean by "the AI becomes more and more stupid"? The AI is always stupid, even on the hardest difficulty. Apart from some rare and surprisingly coordinated pushes, it mostly forgets its squads on the front line and lets them die. |
Thread: FRP11 Feb 2022, 09:18 AM
If FRP were to be removed in a CoH game then you need to scrap all large maps, nobody wants to play a match where you need to spend 5 minutes for retreating to your base & walking back to the front every time. Maps like General Mud, Vielsalm, Vaux Farmlands and Steppes are 0 fun without FRP and every sane person would veto them
I like those maps especially for the reason that you need to think about how to approach and retreat WAY more than in most of the other maps where you just slam your squads in frontally because there is no room and no need to maneuver since every spot it crammed with enemy squads anyway.
Suddenly properly timing your troops becomes important, flanking is possible and teaming up on an enemy actually comes with the real tradeoff of abandoning your flank because you can't jump back within 20 seconds.
I think they are decent maps. Overall they should be slightly more rectangular (wider and less deep) to slightly adjust walk times where they are a bit too long. Also there should be specialized 3v3 maps. If you e.g. play 3v3 on Steppe, spawn middle-left and have to take the left flank, this adds an additional 15-20 seconds to traverse the empty 'base'.
Slight adjustments in that regard as well as some adjustments to make them fairer for both Allies and Axis and they're very good for me.
I'd rather have reinforcement points for soft retreats like OST bunkers cheaper and more available for all factions. This way you can fight an attrition war and overrunning an enemy does not mean he will be back in under a minute. These are real strategic victories, forcing your opponents to think more about counterstrategies. Way better than Hamburg where the main strategy is to spam arty no matter what happens. |
Thread: FRP10 Feb 2022, 17:48 PM
I agree with the general notion. 2v2 maps are usually to small so that it is not really worth the resources to go for an FRP. The only factions that sometimes do that are OKW if they go for T1 anyway and USF since they can establish a forward base by regular teching, although it is very susceptible to the Stuka. In 3v3, they can be super strong on some larger maps.
My solution to FRPs is to just remove them. They are hard to balance and devalue soft retreats. They punish proper micro and risk assessment. If you played too risky with your units, the retreat should force them back to base. If you micro properly and assess the risks correctly, a soft retreat to replenish troops should be enough.
OST has the best design on this in my opinion.
The only other option would be to make a real play style out of it, but then you need to commit with basically your whole base. I doubt this would really work in CoH though. In the current system of CoH2, you can have a full base in the back and still cheese FRP strategies with comparatively small cost and risk.
What should probably also be done is that units under fire cannot reinforce anymore, unless they are in the base sector. This could then also be used to set apart some elite squads like paratroopers that might still be able to reinforce in combat. |
Again, completely false. I have played games where I am 2v1 all game long and have my side and we still lose badly.
Just played a game on White Ball where all 3 of my team went middle at the start and were losing. How can anyone win this game?????? And our team ranks were much better then the enemies.
So me as soviets with only 3 squads on the field will beat an OST and an OKW who each have 4 squads and MGs??? In what universe?
But this affects ALL players and is not Soviet specific, which was your original point. What happened to you in your anecdotal game will happen to your Axis opponent too: They just end up with one buffoon throwing the game, leaving you the win. If this buffoon plays on the other side of the map, you probably won't notice. You're busy holding your own part, you will at some point get more support from team mates and win. However, you'll probably not have enough time to find out if the enemy player is indeed an idiot or just got outplayed. On the other hand, if the buffoon is in your team, you will notice because even if you're holding on, other enemies will flush you out while you get no support at which point you'll probably check what is actually going wrong with your team mates.
Anyway, if Axis are easier to play in 4v4 (which according to 4v4 data they slightly are and I am not even debating, although it would be interesting to get more info on the low rank side specifically), then all average skilled Allied players will lose to similarly skilled Axis players and downrank. You will end up with similarly skilled Allied players. The real problematic region in this case will be the low ranked portion of Allies, since player ELO will get compressed. This could be "real" ELO hell, but we have no real indication of how large this problem is. Especially since it is hard to tell apart from subjective and anecdotal experience.
Obviously, the larger the mode and the worse the matchmaker, the worse your match to match experience will be. But again, this will go both ways: Undeserved losses and undeserved wins.
The main issues in my opinion are:
1. No ELO compensation for arranged teams. As a random, you'll always oscillate between your actual skill level (when playing other randoms), and being outmatched (when playing against AT).
2. The duration of games in CoH2. If you get matched with an idiot, your game will still last 30 minutes. If by pure chance you get idiotic match ups, this can cost you 1-2 evenings of gaming. Obviously this is HUGELY frustrating, especially if all you have time for is maybe only 1-2 gaming evenings a week. |
One major issue right now is playing soviets in team games.
Currently in team games soviets are under performing badly.
Axis factions are easier to play. Your infantry works great and you have an early presence.
Any map you get with an OST and OKW player on the same side, you lose. By the time you get to the fuel they have already locked it down. So you lose ELO.
Now your ELO is worse, so you get stuck with noobs who dont even try to get the fuel. They run to the middle and camp out. Then get run over with blobs and early vehicles. So you lose ELO.
Now you are not only playing with noobs, but noobs that just bought the game. Meaning most of them do not have ALL of the allied factions. Which means you now get all soviet games against mixed axis teams.
GG ELO Hell.
Now add this all to premades, AFKs, bug splats, bad spawns, axis favored maps, etc. Playing soviets right now is a joke.
And that is where the MM sucks hard.
1) Getting premades 4-5 games in a row.
2) Games where the enemy is even remotely better than your team = loss because allies crumble after about 12 minutes.
3) Getting the same afk or bug splatter on your team 2 and 3 times in a row.
4) Playing the same map several times in a row. And getting the shit spawn every time.
5) Always getting the worst player on the team in your next match.
With the current state of the game and MM, I /L out of about 3 out of 5 games. Stuck in ELO hell just due to the MM.
This is quite an exaggeration. The game even in 4v4 is at least balanced enough that you don't have to sink THAT far below your actual skill level, because at some point you will just win by out-microing.
Some of these things are just BAD LUCK. And I have horrendous luck. But some things could easily be fixed by those cucks at Relic.
1) Calc an estimated ELO for premades.
2) Dont play the same map twice in a row.
3) Adjust ELO based on current win rates for the factions and mode.
4) Dont play with the person who just crashed the last game.
5) Adjust ELO on faction performance per map.
There are probably 3-5 more things that could easily be added by any competent programmer in about 20 minutes. But here we sit...
Sadly, the only point in time when Relic will gather a lot of data on winrates and maps without it being influenced by patches is when, well, there are no more patches. Catch 22... |
That's totally fine, I agree that the CoH2 match maker is far from perfect.
But I also don't think that the mechanism you describe really exists. Even if the game finds you completely fair games, you'll have a winning/losing streak of +/-3 every 4th session.
Plus on top no one is "fit" every day. If I am tired or just want a chill out session, I will more likely accumulate a losing streak and might even get fully outplayed. Gunther recently posted some study on sleep and car crashes, but there are many more on general reaction times and judgement. You might not even fully realize that you're already at a stage that impairs decision making. That's why most players will probably blame the match maker if they get beaten heavily, they don't even notice that they're having a bad day or at least the extend of it. On the other hand, winning streaks get deemed to be deserved because of good play, not because your opponents were tired or something else.
My point is basically: Most of this will be a subjective, personal bias, with some streaks extended by 1 or 2 due to actual shitty match making every now and then. |
Basically the whole problem is not with ELO ranking in and of itself, rather how the game uses it:
Lets say that in the last few games I suck hard right? Let's say -5 streak. Naturally, I will lose ELO points, my opponents will get them ofc and normally I would play with lower-ranking opponents so as to make the game fair.
The problem with COH2's implementation of that is the fact that they filter your teammates by elo, BUT not your opponents. Hence, ELO hell.
If you have -5 streak it is very likely that you will end up with a teammate that is also on a similar losing streak. HOWEVER, it is no guarantee that your opponents will also be at the same level ELO-wise, and if you work that out using induction, almost always you will be playing against better teams (or even worse, arranged teams) whilst at the same time the quality of your teammates will continually erode --i.e. ELO Hell.
I made the point bold that I don't get.
Do you have any backup for that? To my knowledge, as well as how any other skill based match making system works, it tries to match your ELO with an opposing player. I doubt Relic re-invented the wheel here, since it also does not make sense. Matching your ELO to an opponent is even more important for balancing than matching your ELO to your team mates' ELO.
What you explain does not make sense. A losing streak should have nothing to do with any of this, even if your assumptions were true.
There's these options:
1. The game does care about the ELO of both teams -> games should be fair at any level and previous games do not matter, you'll just get matched to higher/lower ELO. In my eyes, this is the current system with bad implementations and "hick ups" here and there.
2. The game does not care about any ELO -> completely random games. This is surely not the case, at least not by my experience.
3. The game cares only about similar ELO in your team, but not in the opponent's team -> Random opponent strength with random wins and losses. Any streak does not have an influence.
The only way your explanation would be true is if CoH2 somehow remembered your starting ELO at any given day. Then, when you lose, the game updates your ELO only when looking for team mates but not for opponents and will still continue to find opponents for your high ELO.
Why should it do that? This is creating hugely complicated extra steps. Also what happens the next day? Does the game fully update your ELO? Does it only update once you win a game? Do you then get continuous wins as well? What will the game take as a reference point for finding opponents?
None of this sounds logical, and therefore I assume the game does not do that. The match maker is just shoddy sometimes, let it be due to lack of players, due to some players searching too long or other badly tweaked parameters for the current player base. But overall in 2v2 and 3v3 I found it to work decently well.
The really bad thing is that it does not seem to adjust for arranged and random teams, leaving randoms less of a chance in such a matchup. |
Let's take for example a player which is level 10 on 2v2 as SOV. If he loses more than 3 matches straight (which is ridiculously easy to happen, given the absolute noobs/griefers/leavers that plague random matchmaking) not only will he lose 2 whole ranks (maybe 3), but also he will be continually matched with equal ELO teammates against higher ELO opponents. Rosbone has a pretty funny and depressing post which lists such matchmakings.
Could you elaborate why that is the case?
The game will try to match your ELO with similar ELO players on the other team. If you lose ELO, you also get matched with worse players on the enemy team. Why should your ELO go down but you still get upmatched?
The issue is that the match making algorithm seems to make odd decisions especially in team games and creates hugely unfair games every once in a while. |
Combat engineers can work as your mainline because they cost absurdly cheap and have a flamethrower upgrade, but even without it you can hold your ground
they have 2 models less, but you make 7 to 8 squads at the very least, and you stack multiple of them anywhere you are supposed to fight. Fun tactic
They don't
You're ignoring the fact that CE have to retreat after losing about two models. This means your army can only lose half their strength until you're incapable. If you retreat Cons the same way, you have to lose 67% of your army until you are pushed off.
CEs don't work as mainline. Not early game, and surely not in the late game.
Instead of spamming a ton of CE squads that you cannot afford flamers for anyway, just build two CE and 3 Conscripts and use merge. Same effect, better fire power, and cheaper on your MP while having the same amount of models on the field |
Agree with Rosbone. Some computer mice also allow you to unlock the wheel, which can result in some spinning just by rapid mouse movement.
The only camer issue I regularly have is that the camera is on rotation at the start of the game, as if I pressed the "alt" key. |