Make the "Draw Bug" a Bannable Offense

0

I am. I was just replying to Morax that seemed to be under the impression that it is already forbidden. And it not being listed there convinces me of the opposite.

2

@rezy-noob said in Make the "Draw Bug" a Bannable Offense:

it won't be listed in the official rules cause it's not bannable,chill

Purposefully doing it definitely falls under our generic "do not abuse exploits/bugs/loopholes" rule.

E: For clarification, the bug/exploit being that in a draw situation the player who ctrl-k'd gets a "win" even though it should be a "draw".

0

I am soo confused at this point

0

To clarify, if you ctrl-k your com in a scenario where it makes sense (if the exploit didnt exist) is that considered abuse of an exploit?

Eg your com is about to die and/or lose range of enemy com (with you in a non-winning situation) but enemy com is low enough health that ctrl-k will kill it. Regardless of if the exploit existed ctrl-k is or could reasonably be thought to be the best move. Is that considered abuse of an exploit?

If yes I see that as a serious issue because you’re potentially banning someone for playing the game properly (ie they would play that way whether or not the exploit existed). If it isn’t then I’m confused what scenario would be considered abuse of an exploit.

Another example would be deliberately going for a draw by charging your acu into an acu fight but again that is a legitimate strategy if you don’t think you’ll come out on top long term.

M27AI developer; Devlog and more general AI development guide:
https://forum.faforever.com/topic/2373/ai-development-guide-and-m27ai-v12-devlog

0

Ofcourse ctrl-k'ing your acu in that situation is totally fine and part of the game mechanic. Doing so repeatedly on purpose, aka exploiting the current existing bug, is not ok.

0

@noonecares said in Make the "Draw Bug" a Bannable Offense:

I am soo confused at this point

The title of the thread and request is a bit misleading, yes. I will have to write something fresh and lock this thread as it has become quite convoluted.

For starters, a single offense - like any action - should result in a warning rather a ban.

0

As far as I can tell from this.
Your moaning about a feature in the game.

0

@noonecares said in Make the "Draw Bug" a Bannable Offense:

As far as I can tell from this.
Your moaning about a feature in the game.

It is not a "feature." If you need further explanation of that feel free to contact myself or Giebmasse.

0

The scenario is like this:

  • you ctrl-k your ACU, it explodes
  • the explosion damage kills your opponents ACU

Expected game outcome: DRAW
Actual game outcome due to bug: VICTORY for the player who ctrl-k’d

This has nothing to do with game mechanics. It is about a bug in the engine which causes this to be incorrectly reported as a win for you. If we could, we would just fix the bug so the scenario would be rated correctly, but we can’t, mostly because we don’t have the source code and it’s very difficult to debug without that. Thus the only option is to use moderation to prevent people from abusing this to intentionally inflate their rating.

So if you are in the situation where you are about to draw your opponent anyways, you should not be ctrl-k’ing your ACU in order to abuse the bug and get yourself a false win.

0

This can lead to a problem though, because sometimes you really are not sure if you are about to draw your opponent. Your opponent could be about to vet, or have a mobile shield on the way, or enough units coming to block your attack.

Is there perhaps some workaround that can be done by modifying the ACU's death weapon? Replace it in lua code with something that adds a flag to the replay file that the server can see when determining who won the game? I'm not very knowledgeable about FAF's inner workings but I know that as long as the server's able to see something that indicates a draw, then it can award the correct points instead of counting it as a win.

1

There have been years of time where there was no draw bug and ladder rating worked perfectly. Copy-paste the old code, put it in an if-statement to apply to ladder games, and then it works? Or rollback to two years ago when it worked and ban all the developers so it stays that way.

1

Am I gonna have to copy/paste a link to Brutus’ post on the subject for you Thomas?

0

Aging software has no relevance to the rating bug and is largely fake news as proven by the 14-year-old game that still works perfectly despite the fact you cannot modify its source code.

1

I think this topic has served its purpose and should be closed.

A work of art is never finished, merely abandoned

0

@giebmasse said in Make the "Draw Bug" a Bannable Offense:

Purposefully doing it definitely falls under our generic "do not abuse exploits/bugs/loopholes" rule.

E: For clarification, the bug/exploit being that in a draw situation the player who ctrl-k'd gets a "win" even though it should be a "draw".

Then it should be defined as an exploit on the exploit page because this is just an intuitive use of game mechanics, especially with vet existing.

I always ctrl+ for draw if I know I’m not leaving a push alive, why allow the risk of a vet. I’ll also play more aggressive with ACU against higher rated players that are disincentivized from trading with me. Does the sum of these two factors mean I’m intentionally bug abusing or is it incidental and regardless of the client report to the server I’d be doing it anyway?

Is this reportable now? Is it only reportable if I do it 5 games in a row? How about if I only do it during basically all my game draws even if I only draw like 1 game in 15? Problem is this falls into the same manipulation category as uneven team games being rated which didn't stop dudes that got 2.2k rating from 1v3's their friends in casual weekend games but it did stop all the high rated moaning about Sid and Suzuji.

This rule just seems like more shit added specifically with the intent of enforcing it at 1800+ without really caring about the exploit existing elsewhere simply because enforcing it is going to add hours of evidence searching by mods to see if it is a more pervasive problem or not in the replay vault of the accused. It only really results in a ban if it's a "known bug abuser" entity same as uneven teamgames, which I bet doesn't even exist right now and will instead be [insert ladder player] buttblasted that he lost 8 points in one game.

Also another thing, is this only considered an exploit based on the end result? Mainly, I can do the exact same behavior that would be considered bug abuse in ladder but it's OK in teamgames for me to intentionally draw Yudi in exactly an identical fashion because it doesn't impact the game results the client gives?

If so, then you're banning bad rating changes not "draw bug" behavior which makes no sense because it isn't the user's fault that bad rating changes exist.

1

Like really, if it’s a systemic rating problem that cannot be resolved your options are:

  1. Get rid of 5x5s from competitive games since they promote draws and therefore the exploit
  2. Petition balance team to change the game itself to lower draw chance
  3. Petition balance/game dev team to prohibit manual acu ctrl+k

Banning it is just incredibly flawed and lazy, doesn’t address any actual problem and instead puts the burden of faf’s own problems on the end user.

0

We can fix it - it is just a very time consuming bug to solve.

A work of art is never finished, merely abandoned

0

I mean I don’t mind if you take a year to fix it and the bug exists, but if people want a solution NOW then the only options are actual, systemic changes rather than banning behavior that may or may not lead to bad game results.

0

Could you parse the replay of every game with an automated replay parsing tool @Askaholic that looks for the ctrl-K command of acu followed by another acu getting killed and game ending. Can your tool do that?

if you could extract this information it would give you a fix provided you rewrote how games are reported...
This would at least give you an idea of how many games this occurs in even if that is not the case

0

I don’t think the game results are actually saved in the replay, so even if we had a bunch of replays where the bug occurred it wouldn’t help us. It’s also not that hard to reproduce the bugs with the game results. I have a pretty good understanding of what happens, but not why since that’s buried in the depths of the exe. I also wasn’t able to figure out a fool proof way of patching over it in the lua which is why there are certain edge cases like the draw bug that still happen.