what exactly is "H Rating?"

http://www.wargear.net/wiki/doku.php?id=general:help:rankings#h_rating

"50% H rating means you are winning 1/2 of your 2-player games, 1/3 of your 3-player games, 1/4 of your 4-player games and so on"

Larger games mean a lower chance of winning - so this is your likely hood of winning a statistically "normal" number of games for any given game size. 

thank you

ratsy wrote:

Larger games mean a lower chance of winning - so this is your likely hood of winning a statistically "normal" number of games for any given game size. 

I don't think this is a fully correct explanation. Mathematicians (I'm not one) help out please.

An h-rating of 50 means you are winning at a statistically normal number. (You would win the same number if we just drew out of a hat).

An h-rating below 50 means you are winning less than that.

An h-rating above 50 means you are winning more than chance -- and the higher the rating the better.

H-rating factors in the size of the games anf number of games played, so you can tell that a player with a 10 wins and 20 losses in 30 8-player games is at least as good if not better than a guy with 110 wins and 99 losses in a bunch of 2-player games. 

Bottom line, H-rating is my favorite way to tell who is good at board.

BTdubs wrote:

H-rating factors in the size of the games and number of games played, so you can tell that a player with a 10 wins and 20 losses in 30 8-player games is at least as good if not better than a guy with 110 wins and 99 losses in a bunch of 2-player games. 

Bottom line, H-rating is my favorite way to tell who is good at board.

I've also always preferred the H-rating as the best way to rate a player's ability because it factors out luck (presuming a statistically significant of n-player games have been played) and it seems the ranking score least likely to be "gamed".   Unless I'm ignorant to some aspects of how it is calculated, I think it has the following weaknesses, most of which could be addressed relatively easily:

1.  Luck is only factored out when a "statistically significant" number of games is played.   So, the H-rating would be better if it only counted n-player games which exceeded that threshold.  I'm not sure how to determine that number, but I would guess that a player who plays regularly could quickly achieve that number for the smaller n-player games (2 to 8 players).   Since there are fewer games hosted with larger numbers of players (more than 8 ) and those games take longer to play, it might be more difficult to reach that threshold for larger n-player games.    But, not including large n-player games in the H-ranking might not really make a difference in the practical value of the H-rating.  By that I mean, a  player who does well in larger games will probably also demonstrate that in games with 8 or less players.
2. Luck plays a much larger factor in smaller games, particularly for 2-player games.     A better H-ranking would factor that in somehow.    In a way, a 2 player game is a different beast altogether and one could argue that you could separate those out.   An H-rating that only included a statistically significant number of 3 to 8-player games, even though it might seem incomplete, at a practical level, might more closely reflect the skill level than the current H-rating.
3. The H-rating can still be gamed by players who only play certain maps in certain situations, particularly scenarios which puts them in a game with less-experienced players on a map unfamiliar to them.  But, I don't know how to prevent this unless it is specified that H-ranking is only derived from games played on a predetermined set of maps/games.   I don't imagine there are many players though who would fall in this category, particularly if you took out the 2-player games.