I've started reading UP2NG's post twice now and can't make it to the end.....I agree with him tho. I know better than not to. I'm gonna go read something a bit more my speed....maybe some Tolstoy.
Thanks for your comments up2ng, they are always well thought out and spot on!
> Dweebo, seat switching is STILL not completely > accurate in D-Pente).
Noted in my bug list
> The only minor changes > I'd make are that when one player presses Play to > begin Game 2, they get that same dialog -- i.e. the > other player can take up to 7 minutes (for a bathroom > break perhaps?) to begin.
I hadn't thought of that option, makes sense though.
> The real issue is that this would really change how > things are stored in the database, and how database > search results would look. When you search for a > certain player's wins, are you talking about a game > or a set? Or how about just the running list of most > recently played games? Should this change to reflect > grouping the games of a set together and somehow > showing the result of the set?
My thinking on the game database is to keep it the way it is mostly. It is a GAME database after all, not a set database. The primary use of the database is to search for a specific game position. I think that in the list of games, there will be additional information about the set (W/L/D, link to other game in the set).
Viewing games in a player's profile however is a different matter. There I think it makes sense to group games together that are in a set, and perhaps even display both games together on the same page when you want to actually view a completed game (not sure about that).
> (A draw of a single game, which should never happen, I would change to just be unrecorded and the game would have > to be "replayed"). For Pente this is certainly true, but for C6 it is very possible to have a draw in a game. The way it is implemented now for turn-based games is that if one game is a draw, then whoever wins the other game wins the set. If both games are a draw then the set is a draw. I think that is fine for live play as well.
I am looking into the details of how to code this currently.
Posts:
1,032
Registered:
Dec 16, 2001
From:
Powell, OH
Age:
37 Home page
Re: Set-based question
Posted:
Oct 1, 2008, 6:47 PM
It just dawned on me the other day, that if a player wins a set of games under a set-based ratings system, that player's rating should go up twice as much as if they had won both games under the current ratings system, right?
For turn-based games I did not do it that way! Winning a set of turn-based games results in your rating advancing the same amount as if you had won a single live game.
I think all that has to change is the K value in the ratings formula from 32 to 64.
Posts:
751
Registered:
Mar 1, 2002
From:
My Own Lil World Mostly
Age:
43
Re: Set-based question
Posted:
Oct 1, 2008, 11:08 PM
I don't think it matters if it doubles the points. It just matters that it is consistent across the board. IMHO. If changing the K value to 64 is the fix then voila!
Re: Set-based question
Posted:
Oct 2, 2008, 3:01 AM
I agree with P. The amount gained per set win is not as important as changing over to set ratings. Changing K to a larger number would allow players to reach their new natural ratings range more rapidly, but it would also tend to increase the size of their natural range somewhat from what it is currently. Personally, I think leaving it at 32 (or perhaps slightly higher- 48 perhaps) would be okay. But again, the switchover is the crucial bit, what the K factor actually ends up being is secondary.
ETA: One thing which to consider as a possible issue under the new set based rating system having to do with K, though-
K may need to gradually decrease in value as a player's rating increases. Otherwise, top ratings will be more a function of how frequently a top player plays than of their true skill level. What follows is exerpted from wikipedia's article on the Elo rating system:
"Certain Internet chess sites seem to avoid a three-level K-factor staggering based on rating range. For example the ICC seems to adopt a global K=32 except when playing against provisionally rated players. The USCF (which makes use of a logistic distribution as opposed to a normal distribution) have staggered the K-factor according to three main rating ranges of:
Players below 2100 -> K factor of 32 used Players between 2100 and 2400 -> K factor of 24 used Players above 2400 -> K factor of 16 used FIDE uses the following ranges[7]:
K = 25 for a player new to the rating list until he has completed events with a total of at least 30 games. K = 15 as long as a player's rating remains under 2400. K = 10 once a player's published rating has reached 2400, and he has also completed events with a total of at least 30 games. Thereafter it remains permanently at 10. In over-the-board chess, the staggering of K-factor is important to ensure minimal inflation at the top end of the rating spectrum. This assumption might in theory apply equally to an online chess server, as well as a standard over-the-board chess organisation such as FIDE or USCF. In theory, it would make it harder for players to get the much higher ratings, if their K-factor sensitivity was lessened from 32 to 16 for example, when they get over 2400 rating. However, the ICC's help on K-factors indicates[8] that it may simply be the choosing of opponents that enables 2800+ players to further increase their rating quite easily. This would seem to hold true, for example, if one analysed the games of a GM on the ICC: one can find a string of games of opponents who are all over 3100. In over-the-board chess, it would only be in very high level all-play-all events that this player would be able to find a steady stream of 2700+ opponents – in at least a category 15+ FIDE event. A category 10 FIDE event would mean players are restricted in rating between 2476 to 2500. However, if the player entered normal Swiss-paired open over-the-board chess tournaments, he would likely meet many opponents less than 2500 FIDE on a regular basis. A single loss or draw against a player rated less than 2500 would knock the GM's FIDE rating down significantly.
Even if the K-factor was 16, and the player defeated a 3100+ player several games in a row, his rating would still rise quite significantly in a short period of time, due to the speed of blitz games, and hence the ability to play many games within a few days. The K-factor would arguably only slow down the increases that the player achieves after each win. The evidence given in the ICC K-factor article relates to the auto-pairing system, where the maximum ratings achieved are seen to be only about 2500. So it seems that random-pairing as opposed to selective pairing is the key for combatting rating inflation at the top end of the rating spectrum, and possibly only to a much lesser extent, a slightly lower K-factor for a player >2400 rating."
My guess as to why this hasn't (so far) been a factor in turn based play is the relatively smaller number of games played and the longer time frame which games take. It could eventually become an issue there as well, I think, but would take quite a bit longer to manifest. Also, having K remain at 32 there has probably helped to prevent it as well. Message was edited by: watsu at Oct 2, 2008 10:21 AM
Retired from TB Pente, but still playing live games & exploring variants like D, poof and boat
Posts:
542
Registered:
May 9, 2002
From:
Northeast USA
Re: Set-based question
Posted:
Oct 2, 2008, 6:50 PM
I would actually keep it to 32. You would either win a set, lose a set or tie (no change). Your rating would be adjusted based on this result.
I see what your saying, but I think that's sort of on the wrong track, although yes, you'd end up with a similar adjustment as to what happens now if someone wins a set.
I think as has been already said, the K value doesn't matter so much as long as it's not changed all the time. It determines how wildly the scores will fluctuate based on a single result. I'd keep it at 32, but I could probably be convinced otherwise.
Re: Set-based question
Posted:
Nov 29, 2008, 5:28 AM
"Winning a set of turn-based games results in your rating advancing the same amount as if you had won a single live game.
I think all that has to change is the K value in the ratings formula from 32 to 64.
Let me know if you don't agree with that. -dweebo"
I've considered the issue of what the K value should be for set based ratings a bit more, and I actually now think it should be switched to 64 instead of staying at 32. The reason for this is that if someone is rated enough lower than their opponent in order to merit a 64 point increase from winning a set, then I think they should receive full value for that accomplishment. Consider how much more difficult an accomplishment it would be for the significantly lower rated player to pull together not only back to back wins from both sides of the board but also achieve them in the same set. Currently, that player could get a 64 point rating bump by winning twice as P1- spread out over any number of games or sets. With set based ratings, that player gets zilch from playing that higher rated opponent unless or until they pass the won set threshhold- no mean feat.
Retired from TB Pente, but still playing live games & exploring variants like D, poof and boat
