PDA

View Full Version : If the AI wins it should be a draw....


anarres
15-09-2003, 09:53
I think that if the AI wins (by Diplo, for example), the game should be a draw (i.e. no ladder changes).

Having a better score is not a good measure of how well you are doing, and it means the player in the lead in score as you reach the modern age can aim for an AI diplo win, or can slow the other human down with WW to force an AI space win.

I think this rule is unfair, I would like the ladder admins to consider changing it. I am not asking this be changed for games in progress, but for new games.

Aggie
15-09-2003, 10:34
I do agree that we could change it for new games. But I'm afraid that also could be misused. I don't have good examples yet, but something tells me that also isn't waterproof.

DrA yesterday had a great suggestion for games were the AI wins: let the score count for half. So someone gets/loses about 25 points in these cases.

Lt. Killer M
15-09-2003, 11:19
I think it depends on the game: say, both players are at each others throat and don't amount to much all game long. But on player manages to stay ahead in score until an AI gets a culture or UN win - why shouldn't that count as a win? He/She did after all fight for a better score?


Admittedly, that is not the real aim of the game, so I think DrA has a good point there!

digger760
15-09-2003, 11:39
i'm not sure that a draw is such a good idea, what if you made a few bad moves at the start or had a bit of bad luck, you could throw the game. I think if you loose to the AI it is tough titties and highest score wins.

anarres
15-09-2003, 14:29
digger, how could you 'throw' the game? The only realistic way for the AI to win is by Diplo or Spacerace. How can you get to that part of the game unless you are competitive. If you make sucha bad mistake the other player would squash you.

My point is that the person in the lead can deliberately go for an AI win if they know they can't win themselves. This is in my mind unfair, and if it is possible for anything to be "misused" then this is it...

I have a good example too, for why score is a shit measure of performance: Player 1 has 1 or 2 lux's on their half of the world. Player 2 has 7 lux's on their side. Player 2 will always have a higher score, but player 1 may be amazing in even getting to the modern age with only 2 lux'sin their territory. Now player 2 can push for an AI win (defensive stalemates are easy if you go for it - especially since you can drop research to 0 and mobilise at the cost of science buildings).

Can anyone tell me how counting them as a draw could be abused?

Shabbaman
15-09-2003, 14:49
If the AI wins, both players should be whipped.

anarres
15-09-2003, 14:51
Taking my example of why score is not a good representation of how the game is going - you don't even need a disparity as bad as 2 lux's vs 7: if one player has 5 and the other 7, that is almost as bad....

digger760
15-09-2003, 15:14
quote:Originally posted by anarres

digger, how could you 'throw' the game?

opps was'nt thinking straight. I suppose in the situation i suggested all the opponent has to do is play it out till he's past your score [blush]

Nightfa11
15-09-2003, 15:42
I agree with shabbaman, how about if the AI wins, both players get a loss? I'm not sure how fair that is, but I can see someone playing for a draw to avoid a loss, and this eliminates that possibility

Skyfish
15-09-2003, 16:38
If the AI wins it should be a loss for both players...
Thats the principle, now how to make it work I have no idea yet....

Even anarres solution could be used as a cheap escape, as aggie says...

DrAlimentado
15-09-2003, 17:01
Well I think getting a default win thru histo is flawed for all the reasons why score is a crappy indicator of performance. However... making it a draw has it's own problems:

1) A player who is losing the game in the modern age can play for an ai win knowing it will be a no-score draw, this is just as fucked up as the player leading on points going for an ai win imo. In fact... it is even more fucked up!

2) If a high ranking player is fought to a draw by a low-ranking player why should there be no points awarded? In fairnes even if we did say an AI win was a draw then there ought to be some points given according to rank.


So... unless we can think of a better indicator than points then I think we are left with little alternative but to use histo as an indicator of who is winning. I think the half points award (for a default win going on points) is a fair solution.

Either that or we work out a scoring formula for draws based solely on rank. But the half-points (or 1/3 if 1/2 is too much) solution is much easier.

anarres
15-09-2003, 17:24
quote:Originally posted by DrAlimentado

1) A player who is losing the game in the modern age can play for an ai win knowing it will be a no-score draw, this is just as fucked up as the player leading on points going for an ai win imo. In fact... it is even more fucked up!Hmm, not sure I agree.

Current Rules:

* If the player with more points is losing they can go from loss to win under current rules - a huge score difference.

* If the player with less points is losing they can't salvage anything atm.

'No Score Draw' Rules:

* If the player with more points is losing they can go from loss to a draw - a medium gain for them.

* If the player with less points is losing they can go from loss to a draw - a medium gain for them.

It seems to me the current rules favour the one with most score heavily - going from loss to win is huge.

The 'No Score Draw' rules would mean both players could try to engineer a draw if they were losing, going from losing points to nothing. Note that in this instance the difference between winning and drawing is not great, whereas the difference as it stands is *huge* - gaining the points where you would have lost them.
quote:2) If a high ranking player is fought to a draw by a low-ranking player why should there be no points awarded? In fairnes even if we did say an AI win was a draw then there ought to be some points given according to rank.IMO we already take difficulty in to account for wins. An AI win is more of a 'loss' for both in my eyes - neither should be rewarded.

Aggie
15-09-2003, 17:57
OK, what about this: when the AI wins, both humans will lose the points, so I agree with Skyfish but take it a step further. Sounds fair to me :).

EDIT: I didn't read Nightfa11's and Shabbamans's comments... They were the first to suggest that both lose the points.

DrAlimentado
15-09-2003, 18:11
It seems the main objection is that points are no indicator of winning or losing, but points are some kind of indicator. A crap indicator yes... but by the late stages of the game they mean something. Be it more land and pop, or just more happiness, that shows something imo. And whilst it's true that it could be close with the only difference being a lux say, that is no different to the usual vagaries of the map that we all accept as being part of the game.

You are right that for a winning player to go to a loss thru an AI win is a big difference, which is why getting 1/2 or 1/3 points would be fairer that that.


And if a high ranked player draws to a low ranked player why isn't it logical that the low ranking player should take some points? That's what happens in every other ranking system I know of. Even if we say both players lose I would still argue for the higher ranking player losing more on this basis.

anarres
15-09-2003, 18:21
There is the added problem that atm this is a 'zero sum' game. Allowing AI wins to mean losses for both players would make this a 'negative sum' game. That means that peoiple who come in new at 2500 points aren't going to be the median value. They will be in the upper half, and that is not a good idea. I can promise that this needs to be zero-sum to work...

DrAlimentado
15-09-2003, 18:29
hmm... yes, I was wondering about that. I agree on the zero-sum aspect.

I still think a 1/2 points win is a better solution than a no-score draw though ;)

digger760
15-09-2003, 21:58
maybe we should add some new players to the ladder called Xeres, Shaka, Abe..ect

[EDIT] another though (or passibly lack of thought) if you recorded losses for the AI then there score, after time gets real low, cos generally they loose all the time...if you were to account for AI scores and you lost to the AI then you would be up for a huge drop in points [/end thought of madness]

Kemal
15-09-2003, 23:06
Though I agree an AI win actually resulting in a ladder victory for one of the losing (to the AI) players seems a bit strange, it does fit perfectly in the point of view that in PBEM, the AI can be seen and used as tools by the human players. A player actually letting an AI win if that would mean a ladder victory for him in my opinion just seems to be an extreme form of this "AI-are-tools-philosophy", but in essence not very different from letting them help you win the game by forming alliances against the other human player, or letting them fund your research projects etc.

Though I doubt this "trick" can and will be used in many games, why view this any different than just another one of the numerous factors which players will have to take into account when plotting their strategies for either winning the game themselves or disrupting the efforts of the other human or AI players, as it seems to me that if you're unable to prevent these situations as the would-be-losing player, it is already very questionable in itself whether you shouldn't be the losing party, since you can either stop the AI from winning this way by (amongst others of course):

1. Economical domination, ensuring you reach the UN or Spaceship technologies first and thus preventing the AI from winning by not holding the vote in the first case, or launching the ship sooner and winning by space race yourself in the latter.

or

2. making sure your empire is powerful enough to get on the voting ballot yourself, so that you yourself, alongside the other human player, will be able to manipulate the AI's voting behaviour, thus creating equal possibilities for both players to get a win in the final UN-vote, if the AI chooses to hold them.

Besides the fact that if a player's empire is unable to fullfil either of the above criteria, while the other player is able to set up such an AI win, I personally would feel that a victory for the latter player is a just result, it should also be noted that in itself the situation of the AI building the spaceship/UN wouldn't be much different to the situation in which the human player who would recieve the ladder win would be building them. The only difference would be that the city building it would be of a different colour/side (and probably be very poorly managed and defended), since it will change nothing to the fact that the winning player will go all-out to try and get them completed, while the losing player will do anyhting to prevent it, same situation as if the winning player was completing it in one of his own cities.

Taking all this into account, I personally do not see the reason to adjust the current rulings regarding AI wins in PBEM resulting in ladder victories for one of the two human players. After all, whether you win or lose is all in your own hand. Thanks to the mapmakers, starts will be about equal for both players, and if it should somehow happen to be that clearly it is not, just restart the game on a new map. :)

ProPain
01-10-2003, 23:37
I see the problem with the histoscore but good alternatives are hard to find. Zero sum must be kept intact imo.

Possible solution, I add a player called AI (or SidM :) )to the ladder who is awarded the win when an AI wins and gets points from both humans. This ensures zero sum.

This has several drawbacks
- Theoretically this will also result in the AI taking no1 spot over time because it will never lose. When the ai loses, also a human player loses and the human gets the score deducted.

- 2 players who are losing to the AI will be tempted to abandon the game. No incentive for either player to continue.

- It's complicated to add to the database algorithm.

For me it's plain this is not gonna work. So I say, stick to the present system.

I'm gonna agree with Kemal:you know the rule, so you can act upon it.Besides that I think a winning AI is not gonna happen a lot.

anarres
02-10-2003, 00:59
FWIW Kemal did a pretty good job of convincing me.