Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Showdown35]

Hello All! I wasn't sure where to put this, so I figured General Discussion works as it's not really a Tournament or Event post.

I have created a ranking system to use with my playgroup, just as a fun way to keep track of stats and have a leaderboard, without having the constraints of a play schedule or tournament format.

Pretty much the idea is that whenever we get together to play, I just write down the final scores of each game, and enter them into a spreadsheet that I created.  It doesn't matter which players are present, or if some players play way more than others, the scoring is a ranking system based on average stats, and as soon as a player plays their 4th game, they are added to the rankings.

So take a look, feel free to steal the idea and even use my excel scoresheet (which you can download from my league site).

The statistics and rankings are explained on my site as well.
 Here's the site: http://scon.comze.com/dominion/

The "Player Cards" are just some fun I had with photoshop and I think are a funny way to keep people interested in checking out the site.

Let me know what you think, or if there is a better way to do the scoring.  It took me quite a while to get all the formulas right for the spreadsheet, so any changes (other than simple ones like changing the amount of points awarded per game) may take a while to implement properly.

Thanks!

[eHalcyon]

Looks good!  For a casual group, these stats can be really fun to track.  Take it with a grain of salt though.  If you're more serious about it, here are some criticisms:

You're only counting games with 3-5 players.  Why not 2?  Even with 3 players, it's a little rough, but with 4-5 it's kind of an enormous luck-fest.

It's not entirely clear when you say "points" whether you mean points awarded based on what place you finish (which seems to be what you are using for PPG) and actual VP scored in the game (which seems to be for the Victory Differential). 

Using points from the game is not really a useful measure.  Suppose it's a 3p game and there is a really great mega-turn engine on the board.  Players 1 and 3 recognize this and both pursue that strategy.  Player 2 is not really paying attention and decides to just buy as many Estates as he can.  Eventually, both p1 and p3 are just about ready to trigger their mega-turns.  Thanks to first player advantage, p1 does it first and is able to pile drive all 12 Provinces.  Meanwhile, p2 ended up collecting all 12 Estates.  Final scores are 75, 15 and 3.  The average score is thus 31.  The differentials are +44, -16 and -28.  This makes it seem like p1 was far better and p3 far worse than the other players, but actually p1 and p3 were really close and p2 was terrible.

For a simpler example, consider the point spread that can happen in Goons games.

Edit: fixing a math fail.

[Showdown35]

This is definitely meant for casual play.  Just a way to make a casual playgroup a little more interesting.  Like you're playing towards something overall and not just on a game to game basis.

You're points about Victory Point scoring are definitely valid, and I considered how much VP can fluctuate from game to game when developing the scoring system.  That is why I wanted to create the Differential scoring, so it at least compares each score to the average score of that particular game.  Even if you win a game with -10 VP, you're ranking won't be worse than if you lost a game with 15 VP.

You're points are also the reason I made Victory Point Differential only contribute to half of each players ranking (with Points per Game--yes, points awarded for the place you finish--being the other half).  I didn't want rankings that went solely on what place you finish, because, as you stated, there can be lots of luck involved in a 4 or 5 player game.  I also wanted close finishes to be more valuable than blowouts.

As for not including 2 player games, that was a personal choice because the point of the league is to encourage group Dominion nights, not two people playing 10 games in one day. The scoresheet could be adjusted to include two-player games.

One more thing, my playgroup tends to stay away from ridiculous mega-turn combo engines as we find those kind of games a little un-fun for some of our less experienced players, so I wasn't too worried about the scenario you gave as an example being an issue.  I toyed with the idea of making the VP Differential stat be based on percentages as opposed to averages (as in, what percentage of all VP you take in a given game), but had trouble trying to figure out how it would work with negative VP scores.

[Ozle]

Using points from the game is not really a useful measure.  Suppose it's a 3p game and there is a really great mega-turn engine on the board.  Players 1 and 3 recognize this and both pursue that strategy.  Player 2 is not really paying attention and decides to just buy as many Estates as he can.  Eventually, both p1 and p3 are just about ready to trigger their mega-turns.  Thanks to first player advantage, p1 does it first and is able to pile drive all 12 Provinces.  Meanwhile, p2 ended up collecting all 12 Estates.  Final scores are 75, 15 and 3.  The average score is thus 31.  The differentials are +44, -16 and -72.  This makes it seem like p1 was far better and p3 far worse than the other players, but actually p1 and p3 were really close and p2 was terrible.

For a simpler example, consider the point spread that can happen in Goons games.

I agree with you, but for an entirely different reason.

in a game where the end result is how many VP's you have it doesnt matter how and when you get them, the player who ends up with the smallest amount of points IS how you rank people, so in the example you have given P1 IS far better from the end result and P2 has a P3 has a far worse result.
The scoring doesnt care how 'well' you have played during the game or how close you were to the mega turn, tough.  (P3 has essentially gambled and lost)

However, where I think using points is bad, is because of the variable game possibility

You could have one game with Alt VPs, goons or something else that produces a high variance, while another has just plain old vanilla, so low variance.
Therefore someone who only played in the second game is massively handicapped against someone who only played in the first game.

[Ozle]

P.S Someone should tell Jeff Botter that he is off my Christmas card list!

[eHalcyon]

Well, using the differential encourages winning by more points.  In the example, one player has very few points and another has a whole bunch.  The differential in that game creates a gap of 72 points between first and third.  In a different game, maybe one player isn't paying attention and a skilled play sees that he can end the game in a win with a small lead.  Maybe the final scores are just 3/2/1, with the 3VP player exhibiting tremendous skill by pulling off a surprise 3-pile ending.  But in this case, the winner only gets a measley +1 while the losing player only gets -1.

Using percentages can help, I suppose?  Not counting the in-game VP is probably best though (or at least give it much less weight).

[sudgy]

Well, using the differential encourages winning by more points.  In the example, one player has very few points and another has a whole bunch.  The differential in that game creates a gap of 72 points between first and third.  In a different game, maybe one player isn't paying attention and a skilled play sees that he can end the game in a win with a small lead.  Maybe the final scores are just 3/2/1, with the 3VP player exhibiting tremendous skill by pulling off a surprise 3-pile ending.  But in this case, the winner only gets a measley +1 while the losing player only gets -1.

Like the game where I three-piled and bought one Estate?

Also, it encourages winning players to win more.  Sometimes I three pile just to end my opponents misery, but in this case I would probably buy all the points just to get better.

I think the only way to make a good rating is when someone wins they get the same amount of points in the rating system each time.

[Ozle]

Well, using the differential encourages winning by more points.  In the example, one player has very few points and another has a whole bunch.  The differential in that game creates a gap of 72 points between first and third.  In a different game, maybe one player isn't paying attention and a skilled play sees that he can end the game in a win with a small lead.  Maybe the final scores are just 3/2/1, with the 3VP player exhibiting tremendous skill by pulling off a surprise 3-pile ending.  But in this case, the winner only gets a measley +1 while the losing player only gets -1.

Like the game where I three-piled and bought one Estate?

Also, it encourages winning players to win more.  Sometimes I three pile just to end my opponents misery, but in this case I would probably buy all the points just to get better.

I think the only way to make a good rating is when someone wins they get the same amount of points in the rating system each time.

Yep, agreed

And that way, you encourage people to play more games, because they play more games, they get more chances at points!
(you could do least amount of games played as a tie breaker)

[serakfalcon]

Probably could fix the point differential by dividing by the zero-adjusted total amount of points gained. Aka positive proportional points.
(you have to adjust for negative scores, and who has the time for standard deviation with casual games?)
What i mean is:
if at least one person has negative points, give everyone an amount of points equal to the most negative points person, then divide each person's score by the total amount of points, that's the proportion of points won.
That scores low point games the same as high point games, but otherwise captures what you're trying to look for.

Therefore winning 1-0 (100% ratio) gives the same bonus as winning 80-0 and almost the same bonus as winning 530 - 10, (98%) and the same bonus as winning -4 to -6. However winning 24 - 25 only gives a ratio of 51% to the winner.

For the example above with 75, 15 and 3 points the differentials would be 80.64% 16.13% and 3.23%
Which would be the exact same ratios you would get if you won 25 / 5 / 1

Edit: you could probably also adjust for negative points by always adding the amount of curses to everyone's score (so +10 points to both sides in 2 player game, +20 to everyone in 3 player games) which has the effect of making the ratio for low-point games closer, and also ratios for multiplayer games closer. So it naturally compensates for a bit of the randomness of those games.

[DG]

You really shouldn't go into points differential at all. It only needs a game where one player can score vp chips each turn and the rest can do no more than buy out coppers to end the game, and you'll realize that win/loss/tie is all you can reasonably measure. You absolutely need a scoring system that encourages players to win (or tie) as quickly as possible in order to keep the games fun.

[flies]

If the players are playing directly for the rating, then rewarding points differential encourages players in a winning position to draw the game out and get more points.  This is something I occasionally do against bots (on Adventures this _is_ how you get 3 stars), but against people it's not very sportsmanlike.

However, since you say this is a friendly group who've instituted a rating system for kicks, the above isn't really a concern.  The greater concern is that winning by one point is still winning, and that's what the rating system should reward.

[Stealth Tomato]

Point differential also warps the strategy. If I see two strategies, one of which has a 60% chance of winning but would only win by a few points, and the other of which has a 40% chance but by an astronomical margin, my strategy decision changes based on how valuable point differential is.