Only a little idea, but to explain, i'm playing somebody who plays a lot of games concurrently but due to access problems hasn't been online for a while. As a result, many folk have taken TOs, my opponents' point score has tumbled, and now should i win or loose, the RHP point system is effectively against me (i stand to win nada, or loose a lot from playing someone whom i suspect is better than me!).
This problem would not occur if the points to be gained/lost (i.e. the ratings input to the points score change alogrithm) were set/stored at the start of each game.
It makes even more sense for those (not me) who care about their rating and put only accept challenges from... limits onto games, though granted, many of them just dont feel in the mood to slaughter amatures like myself!
So what do you folks think? And to Russ, is it feasible?
- Joined
- 26 May 02
- Moves
- 72546
Originally posted by CrowleyNo, it's calculated from the ratings at the end of the game.
I think it already works like that.
The ratings are calculated on the start of game ratings of both players.
I looked at the help, but didn't find any info about that, but I think I read about this somewhere in the forums.
That's good if you're playing somone new to the site, who is likely to improve, as his/her rating should be significantly higher by the end of the game.
But if someone decides to leave the site and resigns all their games, the order in which he/she resigns them has a big effect.
So there are pros and cons for either method of calculating ratings. It's easier to stick with the present system.
-Dave
good point about the equivalent upwardly mobile dave!
i suspect that of the two cases, your example of the resigantion is likely to have greater system impact and so i would bias the downward case above the upward one.
The main reason for thinking so is that (most) new people will not be pawn stars and so only have 6 games on the go at any time anyway.
However, established players leaving / going inactive are more likely to have more concurrent games (though i'd hope somebody leaving would play them out before heading for the door!).
My initial point started because i now have a situation of choosing whether to TO that is biased by the points system. As it happens, because i know my opponent isn't able to get online, i shan't TO anyway, but if i didn't (and he didn't tell me: i only heard rumour in a forum), then i might have been tempted to take the TO before his rating fell too much.
hmmm.... on a bit of reflection, even I now think my initial idea is rubbish.
Points / site abusers eh? It would be too easy to let your rating fall, encourage LOTS of games with low rated players, then kill 'em all and gain masses and masses of points.
So I'm now totally in agreement with Dave: DON'T do this. It's a shockingly thought out, poorly argued idea that should forever vanish and never rear its' ugly head again.
Originally posted by ToeWhy would someone want to lose a lot of games to win again afther it....sounds pretty stupid to me!😕
hmmm.... on a bit of reflection, even I now think my initial idea is rubbish.
Points / site abusers eh? It would be too easy to let your rating fall, encourage LOTS of games with low rated players, then kill 'em all and gain masses and masses of points.
So I'm now totally in agreement with Dave: DON'T do this. It's a shockingly thought out, poorly argued idea that should forever vanish and never rear its' ugly head again.
Olav
Originally posted by LivingLegendwell, imagine putting yourself down to say 1000 points.
Why would someone want to lose a lot of games to win again afther it....sounds pretty stupid to me!😕
Olav
loose loose loose time: looks good to, 'cos your win/loose thing shows lots of losses: a good lure tothe unwarey.
Then start challenging people in the 1200 point range: non-pawn stars for preference. Don't beat any till you've got 100 on the go.
then get 100 * 30 points or so when you thrash 'em all, or most of them anyway.
Bingo, a 4000 point rating in a month or two.
Originally posted by ToeIt doens't work like that though because the amount you get for winning is inversely proportional to the difference between you and your opponent. For instance, I won a game today in Quick Fire V against someone who had a rating in the 1200s and my rating was in the 1600s. I didn't gain any points and my opponent didn't lose any.
well, imagine putting yourself down to say 1000 points.
loose loose loose time: looks good to, 'cos your win/loose thing shows lots of losses: a good lure tothe unwarey.
Then start challenging people in the 1200 point range: non-pawn stars for preference. Don't beat any till you've got 100 on the go.
then get 100 * 30 points or so when you thrash 'em all, or most of them anyway.
Bingo, a 4000 point rating in a month or two.
The idea is that you're supposed to be able to beat someone 400 points below you so you don't gain anything for doing so.
So, taking your example, the amount you gained from each successive victory would diminish until you didn't gain anything once you hit somewhere around 1600.
Paul.
Originally posted by chewieIndeed you are right chewie. That is how it works now.
It doens't work like that though because the amount you get for winning is inversely proportional to the difference between you and your opponent. For instance, I won a game today in Quick Fire V against someone who had a rating in the 1200s and my rating was in the 1600s. I didn't gain any points and my opponent didn't lose any.
The idea is that you'r ...[text shortened]... tory would diminish until you didn't gain anything once you hit somewhere around 1600.
Paul.
The discussion was about if it changed to defining the points to be won/lost at the start of the game rather than at the end, then playing the 'cheat' above would define your winning points according to your initial rating (1000 in the example), not your current one (increasing per victory).
That's why it was such a howler of an idea in the first place! Keep it as it is is the conclusion of this little thread!