Is this working as designed, a bug, or a weakness?
I've always been puzzled by how board rankings are calculated. I've done particularly well on the Civil War board and was puzzled why everyone ranked higher had lower win percentages and H-ratings (some significantly).
http://www.wargear.net/boards/view/Civil+War/Rankings
I've played 17 games. That's not statistically insignificant, but it is lower than everyone above me. But, then again, the player ranked 7 slots above me has only played 20. Then I realized that I had no tournament or team ratings.
So, my question is, are the team and tournament ratings used in the board rankings or are they just there for general info? If they are used, are they based on the standard 1000 points assigned to those playing the board for the first time? If so, would my ranking actually go up if I played a tournament and team game on this board even if I lost every game? Again, presuming these are used, maybe instead of zero, 1000 should be used as the default ratings instead of 0.
Thoughts?
It looks like you just need to play a lot more games. Someone who has triple your game count and an 85% H rating should be ahead of you in the rankings (IMO) even if your H rating is a little bit higher. Trying to get to the top of a board's ranking isn't a flash bang achievement but more of a conquest. With the popularity of that board it makes for quite the long conquest at that.
http://www.wargear.net/wiki/doku.php?id=general:help:rankings
Board rankings are explained in the link above.
Two more things to note. One factor that needs to be considered is the board ranking of the players you are playing against. The top 10 players on that map have played an average of 57 games.
it's probably a combination of # of players in a game and board ranking of the players in the games.
you have a higher win% than most everyone (almost twice as high), but 1/2 to 1/3 of the games as the low leaders. you keep up the winning with more games and you'll likely be with berick.
the more big games you win, the less games you'll have to play to gain those points.
Abishai wrote:It looks like you just need to play a lot more games. Someone who has triple your game count and an 85% H rating should be ahead of you in the rankings (IMO) even if your H rating is a little bit higher. Trying to get to the top of a board's ranking isn't a flash bang achievement but more of a conquest. With the popularity of that board it makes for quite the long conquest at that.
Thanks for the reply. Am I correct that you're speculating more than basing your ideas on definitive knowledge?
I very much understand and appreciate the difficulty of maintaining a high win percentage as the number of games played increases. (That's why I noted that 17 is not necessarily statistically insignificant... because I recognize that it's easier to win 75% of one's games when one has only played 4). It's unlikely that if I were to triple my current number of games on that board, I'd be able to maintain a75% win average. So, I have no problem weighting results to correlate for that and, in fact, would argue for it.
The reason I'm questioning number of games played is because it doesn't seem to be a consistent factor in ratings. For a while, I was ranked first on the Crystal Caves map with only 8 games played. Granted, that board hasn't been played as much as Civil War, but at the time I was top ranked, the 2nd ranked player, IRsmart, had played at least 5 times as many games as I. Harry Chest is ranked 5th in the global rankings with only 95 games played (relatively very low compared to most others).
I read the wiki page that describes rankings but I'm not seeing how it addresses the notion of number of games played except indirectly at best. And there's no mention of whether team and tournament ratings are a factor in board ratings... unless by not mentioning them, it implies they are not a factor. So, I'm still looking for a more definitive answer.
It's mostly the # of players in your games.
In the Civil War games you have won the # of players were 10, 4, 5, 4, 4, 4, 3, 6, 3, 3, 7, 9, 5 which is an average of 5.15 (your record is 13-4)
Marsaic has a higher score (1933 to your 1754) despite a worse record (12-8) but his wins came with these #'s of players: 4, 16, 12, 6, 6, 6, 8, 5, 7, 6, 6, 8. That's an average of 7.
Or, put another way - you have earned points from beating 54 players and Marsaic has earned points from beating 72 players.
Mostly Harmless wrote:.And there's no mention of whether team and tournament ratings are a factor in board ratings... unless by not mentioning them, it implies they are not a factor.
They are not.
No, there is no speculation on how rankings are calculated.
Team and tournament games only count towards team and tournament rankings not GR or board rankings (note board 'ratings' are something different).
In addition to how many games you win and your win percentage, you also need to include game size and the rankings of your defeated opponents. If someone wins a large game against high ranking opponents, their score will go up a lot more than if they win a small game against low ranking opponents.
The exact formula is given in the link above. If you think some of the rankings are incorrect, you can calculate them by hand to check.
Mostly Harmless wrote:Am I correct that you're speculating more than basing your ideas on definitive knowledge?
There is no doubt that Team and Tournament play are not part of the G.R. It may not be specifically stated in the Wiki - but then the wiki is written by players like us. It is implied.
As far as diffusing doubt is concerned, all you have to do is look at your ranking history and note that there are no team or tournament games counted in your G.R. fluctuations - If you have never played in a team or tournament, all you have to do is check the ranking history of one who does play in such game types.
I would note that in your case, game size (# of players per game) seems to play into the nature of how slow your ascent to the top is. With an average game size of 3.1 players, you barely have 2 opponents in your average game. Most of the high ranking players on this site see an average of 5+ players in a game.
Yeah... I would play more 9 and 10 player games if I were u......harry chest average game size...9.1
Mad Bomber wrote: Yeah... I would play more 9 and 10 player games if I were u......harry chest average game size...9.1
Actually i was curious a couple of years ago and created a spreadsheet of all the games played by the top 5 ranked players and analyzed the points gained/lost compared to game size. It was pretty clear that the players earned the most points by playing 9-10 player games.
All you guys know this better than I, so correct me if I'm wrong, but:
You also gotta beat the good players. If you win against 19 players with a score of 900, you get waaaay less points than if you beat 16 players with a score of 1500.
ratsy wrote:All you guys know this better than I, so correct me if I'm wrong, but:
You also gotta beat the good players. If you win against 19 players with a score of 900, you get waaaay less points than if you beat 16 players with a score of 1500.
Correct but overstated. It is not "waaay". It's not even "way" for small games.
Jake has a score of 1800. He wins a game against 4 other players who each have 900. He earns 4*(900/1800)*20=40 points. He wins a game against 4 other players who each have 1500. He earns 4*(900/1800)*20=67 points.
Jake has a score of 1800. He wins a game against 9 other players who each have 900. He earns 9*(900/1800)*20=90 points. He wins a game against 9 other players who each have 1500. He earns 9*(1500/1800)*20= 150 points.
So I would say "way" more for big games vs small games.
Also since most games have a mix players the "player strength" sort of evens out and it's really game size that is the dominating factor.
Mostly Harmless wrote: Is this working as designed, a bug, or a weakness?
I was kind of indirectly answering the first two questions and giving my opinion on the third.