Originally posted by Dragon FireI'm unable to easily see the tournaments you're talking about. It sounds like you and somebody else are in the wrong bands. Just because 2 people are in the incorrect bands, doesn't mean that there isn't a problem.
1450-1500 - through to next round;
1500-1600 - through to next round;
1600-1700 - knocked out (my 3 wins and 3 draws lost to 4 wins, 1 draw and 1 loss);
1700-1900 - 50/50 2 games to go against same opponent, 1 won 1 lost. If I win the 1 I am through as is my outstanding opponent if he wins the other 1);
1700-1800 - winning last game whi ...[text shortened]... through but if I lose I am out.
Not all my own way by any means so what is the big deal?[/b]
D
The highest rating ever on the site seems like the easiest solution. Why make this complicated?
If someone is that much worse than their highest rating (how is this possible in one or two years, but whatever) then they can paas on banded tournaments and enjoy the many other opportunities to play on this site.
Originally posted by Dragon FireThat is absolutely false.
A 1750 has exactly the same chance of beating a 2100 as a 1400 has of beating a 1750 as the rating is based on probability theory.
equally the 2100 has the same chance of beating a 2450 and a 1050 has the same chance of beating a 1400.
If chess was throwing dice your statement would be true.
But chess is not about throwing dice.
Rating curve shows that.
Originally posted by rmackenHave a rec. A great argument for using peak rating instead of rating within some arbitrary time period to determine eligibility.
Thread 53253
Need I say more??
EDIT: Hmm. It didn't help your rec. count any! 😉
It is both unfair and demovitvating to use you highest ever rating to establish entry. For example I had a clutch a high ranked palyers resign / timeout in a short space of tim about 3 years ago. pushing me above 1500 I've never been near that since.
Personally I think there is not a big problem with the current system. Except after hundreds of tournies entered I'm without a win.
There is a lot be said for a twin rating system. Current rating and overall average rating since site joining the site. Whilst this may be of limited use for tournies it would be an interesting stat for all concerned when checking out the varisity of your opponent.
- Joined
- 05 Aug 04
- Moves
- 219597
Whichever is higher from.
Top rating ever less 10%. eg. Top rating 1800 banded tourni rating 1620.
This would stop those that resign loads of games to get their rating down and then enter banded tourni's. What's that all about? Oh I get it. The glory of wiinning a tourni. Kind of like Chelsea entering the FA vase.
Last 30 days.
If some one enters a tourni and goes +10% over the top rating limit BEFORE the tourni starts they no longer qualify.
Originally posted by invigorateI agree entirely. Highest rating ever seems inappropriate for these reasons. I still stand by average rating over last 100 games as the best solution.
It is both unfair and demovitvating to use you highest ever rating to establish entry. For example I had a clutch a high ranked palyers resign / timeout in a short space of tim about 3 years ago. pushing me above 1500 I've never been near that since.
Personally I think there is not a big problem with the current system. Except after hundreds of tournie ...[text shortened]... ould be an interesting stat for all concerned when checking out the varisity of your opponent.
I dont really see an issue with the current system.
Here is one for the idea stew though.
Take the highest rating over the last n days
Take the lowest rating over the last n days
Then use the rating in the middle.
Something like
High 1450
Low 1400
So rating to be used 1425
This would hopefully help in 2 ways
1. Stop better rated players entering tourneys they are 2 good for.
2. Stop worse? rated players entering tourneys they are not good enough for (ie rating spike due to T.O)
The question then is n days = what ??
My guess 60 days, but we use the 60 days not from today but 60 from the date they last moved. Hopefully that would solve the problem where they have been inactive for an extended period.
Daniel
Originally posted by RagnorakThankyou Ragnorak for pointing to this thread.
Read this, and the following few posts, for what I think is the best solution(s).
http://www.timeforchess.com/board/showthread.php?threadid=46202&page=2#post_828738
D
It is summed up in a few words:
* We need one solution for "Sandbaggers", people who lost dozens or hundreds of rating points. Here a maximum rating could eysily implemented based on all time maximum rating.
However a good sandbagger would start a game against a weak player via open inivtes and loose this deliberatly (there is no chance to prevent someone who wants to sandbag to do it...).
* We need a solution for new players. A formula was
So, say our New Player's highest opponents were 1600s players.
Count number of 1600s players played = GameCount (say 10)
Get Average opponent rating of these games, AvOppRating (say 1650)
Get record against these players, ie: won 9 out of 10 = 90%
Insert AvOppRating as OppRating in Win Expectancy Formula.
So the formula would look like:
0.90= 1/(10^((1650-NewPlayerRating)/400)+1)
That implies that NewPlayerRating would be guesstimated for banding purposes = 2033.
So New Players band would be 2000-2099.
The implications for a new person playing against only 1200's was adressed also.
I herewith support Ragnorak's suggestion.
Originally posted by rmackenYep, he's done it again in a 15-1750.
Thread 53253
Need I say more??
Originally posted by Marinkatombthen again, it's not like he was able to join tournaments for higher rated players before he got his rating up a bit. and joining tournaments is the quickest way to get a lot of games and get his rating up to his level.
Yep, he's done it again in a 15-1750.
so the real problem seems to be how to compensate rating loss for mass resignations/timeouts, and how to get strong players faster up to their level when they start.
starting level could be easily solved by letting people set their initial rating while joining. maybe there could be fixed choises of 1200, 1400, 1600, 1800 and 2000, with a recommendation (for the newbies) to set 1200 if in doubt. a 1200 setting his rating as 2000 would still be down to (2000+800+800)/3 = 1200 after 3 losses against other 1200's, a little slower if he wins some games. the problem of opponents getting too much gain from the wins against the unrealistically optimistic newbies could be solved by scaling the gain inversely to the initial rating.
mass resignations could be dealed with tournament mods adjusting the ratings of such players by request or 'alert'. the 'bumbed' player would be then dropped off the tournament if necessary, before the tournament starts. they would of course also check against people mass resigning only lost games, to prevent rating abuse.