I was wondering if it’s possible for someone to help their score along by using a combination of surrender and delay tactics, or even helping opponents ..all in an effort to alter the order in which their scores are adjusted.

Consider your current score is 1000 and you are playing in two games against two different opponents (for a total of two opponents) whose scores are both 1000.  One game you are bound to win, and the other you are certain to lose. 

Winning first gives you +20 and -20.4 for a net -0.4, while  losing first gives you -20 and +20.4 for a net +0.4.

No big advantage here with range of just 0.8, but doing this more often than not should add up.

Now consider your current score is 1000 and you are playing in six games against five different opponents (for a total of 25 opponents), all of whose current scores are 1000.  Statistically speaking, you are expected to win one game and lose the others, and wouldn’t you know it, as the games progress you find that you are bound to win one game and you are certain to lose the other five.

Winning first gives you +100, -22, -21.6, -21.1, -20.7, and 20.3 for a net -5.7, while winning last gives you -20, -19.6, -19.2, -18.8, -18.4 and +110.6 for a net +14.6.

Depending on the timing of your win, there’s a 20-point swing in the balance.

There are tactical implications here.  In theory, you should take your time in games you are winning, even taking extra turns to win if there is no risk. Conversely, you should play games you are likely to lose more quickly ..and in those games when it appears another player is in a position to win, help them (especially if they are higher ranked, because you will lose even less points).  Regardless, if another player is certain to win, no matter who they are, it behooves you to speed up the win by helping them no matter what their score.

Some of us may find the above strategies are a little disturbing ethically speaking, but there they are. Or maybe I'm wrong and it doesn't make a difference  because wins and loses even out over time in the wash.

Thoughts? Comments?

Without doing the math as you have done, I would agree with your analysis.  Reason being is that the way the calculation is setup favors the lower ranking person (i.e. they have more to gain whereas the higher ranking player has more to lose).  So, you gain more points by winning while you are lower than when you are higher.

I believe this probably falls into the camp of in-game alliances; it's part of the system here on the site that can be exploited if one so desires.

I will admit that I have intentionally waited to win a game against Midnight Society in Zombies until I had lost a game I was sure to lose against him, to ensure i came out on the better side of the trade.

I agree with what you are saying M.  I guess that means that the calculation for the ranking change is based on your ranking when the game is over, not when it is started?  If the calculation is instead based on your rank when the game is started, and that will remain static, then I guess you wouldn't be able to 'game' the game.

SquintGnome wrote:

I agree with what you are saying M.  I guess that means that the calculation for the ranking change is based on your ranking when the game is over, not when it is started?  If the calculation is instead based on your rank when the game is started, and that will remain static, then I guess you wouldn't be able to 'game' the game.

I'm quite sure rankings are calculated based on the time the game is over. And though your idea is interesting, I would think that to base calculations on start times might be very unrealistic.

Basically it would require all games created before a specific time to be completed before an accurate or "authentic" score could be calculated.  In other words, suppose there were any games begun a year ago that were still ongoing (not completed). In order to be truly up-to-date, all scores (and I mean pretty much ALL scores site wide) could only include games started on or before before the date of the earliest start time, though I suppose Tom could make it such that scores are re-adjusted bi-weekly to reflect recently completed games, but the bottom line would be that scores would always be in flux.

I see what you are saying M, yes that would be very complicated.  How bout keeping it simple, lock in your ranking when the game is started.  Use that number when the game ends without any adjustment for what has been happening in other games.  It is not realistic in the sense that if a game takes a year, your score could have changed drastically, but I could take the approach that the ranking you started with was realistic at the time and it was the ranking your opponents saw when they joined.  Also, I would think that most scores would not change that drastically from beginning to end. 

SquintGnome wrote:

How bout keeping it simple, lock in your ranking when the game is started.  Use that number when the game ends without any adjustment for what has been happening in other games.

For some reason I kind of assumed the calculations were based on the player's scores going into the game, but now that I come to think of it I'm pretty sure this is not the case.  When you look at a game summary you'll notice that the amount of points you get added to your "old" score adds up perfectly to your "current" score (regardless of games that you may have won or lost in between the beginning and conclusion of the game being scored), implying that the "old" score used in the calculation was in fact your the most current score, and not the score you had going into the game.  To calculate otherwise would be VERY confusing for players trying to figure out where the numbers came from.

Under the system you are describing, it still sounds complicated to me. For one, the funny looking math in the summary window described above would apply.  Moreover, if I started a game right now, you're suggesting that my current board score is the one that will be used in the calculations at the conclusion of the game. However, that "current" score is not really current when you consider those unfinished games I've started on the same board weeks, or even months ago.

I think for simplicity's sake the system in use now makes the most sense.  At any given time, the scores will be up-to-date because all calculations are based on a "current" state.  Besides, I'll bet any system that favors the lower ranked player can be analyzed and gamed to some degree.

SquintGnome wrote:

I see what you are saying M, yes that would be very complicated.  How bout keeping it simple, lock in your ranking when the game is started.  Use that number when the game ends without any adjustment for what has been happening in other games.  It is not realistic in the sense that if a game takes a year, your score could have changed drastically, but I could take the approach that the ranking you started with was realistic at the time and it was the ranking your opponents saw when they joined.  Also, I would think that most scores would not change that drastically from beginning to end. 

Would this calculation still be zero-sum?

AttilaTheHun wrote:

SquintGnome wrote:

I see what you are saying M, yes that would be very complicated.  How bout keeping it simple, lock in your ranking when the game is started.  Use that number when the game ends without any adjustment for what has been happening in other games.  It is not realistic in the sense that if a game takes a year, your score could have changed drastically, but I could take the approach that the ranking you started with was realistic at the time and it was the ranking your opponents saw when they joined.  Also, I would think that most scores would not change that drastically from beginning to end. 

Would this calculation still be zero-sum?

The change in score still adds to zero. I like the logic of updating as results come in better than using the scores from the beginning. The skill estimate being used in the calculation is "more informed" than if you insist on using the rating at the beginning.

That said, either way of doing it should result in the player reaching their equilibrium rating in the long term. (If you game the system, you bloat your rating quicker, either making the next step towards equilibrium slower, or surpassing equilibrium to set up a later fall.)

Using the beginning rating has this problem: Suppose you know (or believe) you'll dominate on a new map. You could then start a ton of new games on that map and get the benefit of the 1000 rating for each of those games. You could use the previously mentioned delay tactics right before starting the next large set of games as well.

Side question - Maybe this is obvious, but I never thought of it before (I've only played a handful of public games).  If you are in a game with two people, one with a score of 800 and one with a score of 1500, and you are 90% sure you are going to lose, it is better to help the better player win, because you will lose less points off your global ranking.  Is that right, and do you guys play like that?

Ozyman wrote:

Side question - Maybe this is obvious, but I never thought of it before (I've only played a handful of public games).  If you are in a game with two people, one with a score of 800 and one with a score of 1500, and you are 90% sure you are going to lose, it is better to help the better player win, because you will lose less points off your global ranking.  Is that right, and do you guys play like that?

I do, but the opportunity doesn't present itself as often as you might think.  FWIW, this might be tempered by the fact that in the earlier stages of a game I'll go after top ranked players a bit more aggressively.

Ozyman wrote:

Side question - Maybe this is obvious, but I never thought of it before (I've only played a handful of public games).  If you are in a game with two people, one with a score of 800 and one with a score of 1500, and you are 90% sure you are going to lose, it is better to help the better player win, because you will lose less points off your global ranking.  Is that right, and do you guys play like that?

Yes, that strategy would work if you are at the point of mitigating damages. 

Havent read the whole topic so i might repeat something, but these strategy's only work if you stop playing after a set of games. Most of us will keep on playing more games, so the advantage you had will result in a bigger disadvantage next game. In the end when you keep playing every player will reach his balanced score as a result of his skill/luck.

Yeah, and I was also thinking that the initial advantage that you may gain (and manage to keep at equilibrium) may very well be off-set by the fact that you have lost more games "on schedule" than you have won (because you have been delaying your wins), thus keeping your score artificially low.