She is the better player.
Top 57 Players on Earth
- Thread starter Rhea
- Start date
She is the better player.
See. Here is another difference that comes to mind.
Most of the games that use ELO type algorithms are games that involve 2 participants.
In chess,you have two people battling it out over a chessboard.
In sports, you have one team vs another.
The thing about all those games, is that there were two active participants that each had opportunity.
Pool isn't always like that.
In rotation pool, you have situations where only one player is at the table for a game won, when they break and run out. How is it fair to take away points from the loser of that game if they were never at the table?
Is it really fair to give that game the same value as a game where both players were at the table and had opportunities to win?
That is why I feel that to be more accurate, if they are going to use all games and not make a distinction of 1 person vs 2 people having opportunity in each particular game, that the actual winner of the set should be factored in somehow.
And then different values of sets being determined by the length of the races of each set. Race to 7 counting less the race to 9, and race to 11 counting more then 9, race to 13 counting more then 11, etc etc.
And that the win or loss of the set, should have some type of value on its own, differing from the individual games it was comprised of.
You follow me?
Prima facie that sounds logical
Except that a set (whether race to 7 or 9 or 11) is equivalent to 1 game in chess that is if it is alternate break . Which is probably why alternate break matches are better to be included cos both players will have chance at table
Including winners break is tricky coz player A and player B may be same skill but player A gets to break and run pack of 7 to win 7-0. The rating will then show that player A is much better than player B even though player B never had chance at table
Prima facie that sounds logical
Except that a set (whether race to 7 or 9 or 11) is equivalent to 1 game in chess that is if it is alternate break . Which is probably why alternate break matches are better to be included cos both players will have chance at table
Including winners break is tricky coz player A and player B may be same skill but player A gets to break and run pack of 7 to win 7-0. The rating will then show that player A is much better than player B even though player B never had chance at table
How often does a player break and run the set out?
Is it really enough to make a difference?
Prima facie that sounds logical
Except that a set (whether race to 7 or 9 or 11) is equivalent to 1 game in chess that is if it is alternate break . Which is probably why alternate break matches are better to be included cos both players will have chance at table
Including winners break is tricky coz player A and player B may be same skill but player A gets to break and run pack of 7 to win 7-0. The rating will then show that player A is much better than player B even though player B never had chance at table
Exactamundo!
It's the core concept that counts.
Games that only involve one player. Or situations where one player has more opportunity then another.
That's all.
Yes I follow.
The break and player opportunity argument makes sense. (As an aside, you have said many time over the years alt break format stinks....). I think the break format really doesn't matter as much as we all think it does though. No one is running a 7 pack. It happens once a year, if that, on a tournament. Both players get up there and shoot plenty of times in a match.
As far as giving extra points to the winner of the set, in addition to only individual games, that also is another possibility. I can see it both ways though... In chess, there might be a championship match that is the best of 7 games (I'm making this up, I don't know the chess formats...). So, the eventual champion might win 3 games, lose 2, and draw 2 games. Should winning this "chess set" transfer more points between the two players compared to if they played the same 7 games over 7 individual tournaments spread over time? I don't know that it really makes any difference in the long run.
We can debate the best formula to use. So many variables: the weighting of recent matches compared to older matches, how many points are transferred between players, how many points are transferred between new players with less total games played, adding points for overall set winner, etc. All of these would change the list order to some extent. I kind of think though (guessing) that any of these changes would not be that significant to the overall list, and that the ranking order would be mostly the same. Its like in real tournaments, the same few players always seem to find the winner's circle. Does not matter the game, the break format, the table size, the cloth used, etc. The same guys end up on top because they are simply better.
Well sure 7-0 whitewash B&R is uncommon but it is the psychological mental impact when someone runs a pack of 4 or 5 (which is quite common) in race to 7 . The question to be asked is how often does player behind comes back and wins the match. I am guessing small minority
Fargo follows ELO of chess and in chess besides player getting alternate turns , playere also to play alternate whites and blacks . Playing white in chess is like breaking first in pool . Besides being the almost perfect game with almost zero luck , chess is well suited for ELO theory coz it is almost frictionless and more likely to be closer to theoretical statistical computations. For other sports like rotation pool ( which has probably one of highest % of luck) , obviously there has to be adjustment of many variables in practice to bring it closer to theoretical computations. :smile:
Agree. Alt break would make much more sense to the core concept. Whether it makes a difference in the real life results is debatable. But certainly to the core concept of a per game rating system it would be preferable.
What you are essentially saying is that there is a problem with Fargo because it doesn't help to determine who will win a match between two evening skilled players. You might want to think that one out a little bit more. Of course it isn't going to be able to say who is going to win if two players are evenly skilled, lol. It tells you the truth, that since they are evenly skilled, it is a coin flip. And of course the system can actually predict one of the players to be X percentage more likely to win that the other when they are not evenly skilled, because now that they are not evenly skilled one is actually more likely to win than the other.
I think you have this kind of backwards. Actually, Fargo is the best tool for rating all players of all abilities from all around the world against each other. To do so there are some trade offs that had to be made and it might not be as good at any one specific things as it could have been were it designed for only that one specific thing. If the system was only for rating professional players, Mike could have tweaked it so that it would have been slightly more accurate at rating the pros. But his intention was not to have the all around most accurate system for rating pros. His intention was to have the best all around most accurate system for rating everybody against everybody at any level for all players in the world, not the absolute most accurate system for rating only the pros (you can't have both in one system, but you can have one system that does it all exceptionally well, and that is what he has).
And obviously the larger the sample size of data, the more accurate the rating will be. I am sure if there were a way to ensure you only received legitimate data that Mike would LOVE to get access to all matches that ever occurred between any two people. But one glaring problem with getting "all data down to the local level" is the trustworthiness of the data. You can't just leave it open for anyone to be able to submit results for anything because invariably people are going to send in totally made up stuff just to try to manipulate things for one reason or another.
I would suspect that Mike would love to have and probably has or soon will be working on trying to get all the data from all BCA and APA and TAP and every other major national league and even from all the legitimate city leagues as well, but actually getting them I'm sure is the issue. Most companies or people are not going to spend the time and effort to compile that stuff and give it to Mike unless they are getting something out of it to offset their time and effort. Also, when have you ever seen everybody in pool come together and cooperate for a common goal? So yes I am sure Mike would love to have all data from every league and every tournament that ever occurred at any level, but actually getting it and getting it accurately is an issue I'm sure.
Unless I am remembering incorrectly, Fargo has essentially the same type process in place because it isn't rating solely by who won or lost, but by what the final score was compared to what it should have been based on your rating. If you were to play Orcullo in a tourney, the expected score might be say 9-3 based on your abilities/Fargo rating, so 9-3 is essentially seen as the "draw" by Fargo. If the actual score for the match is you lose 9-8 Orcullo, you did way better than your rating indicates you should have done and your rank is going to go up, and vice versa. I simplified a bit but this is the gist as I recall.
I am guessing here based on what I know about how the system works. We know that the system uses the past rolling ten years of game history, with the most recent being much more heavily weighted than the older. But if there is no recent data I think you essentially stay exactly or relatively near where your rating was on the date the data for your last match was input and only really changes as the earliest data gets to be over ten years old and rolls off completely and is no longer used in the calculation. Maybe Mike will confirm or clarify.
Yes, there is. Obviously a single pool game doesn't mean much, and means less than single game of chess, but when you get hundreds or thousands of those "couple minute games of pool" it starts to mean something, and pretty accurately I might add. Yes, a chess game takes much longer, and means much more than a single game of pool, but there are fewer of them. What a single pool games loses in quality because it doesn't mean as much as a single chess game, it makes up for in quantity and larger sample size because there are a lot more games of pool played for every game of chess. The bottom line is that what really matters is sample size, not the length or meaning of a single game. If you have a large enough sample size it will be accurate regardless of the length or "tellingness" of a single game.
Again, I think Fargorate essentially "prorates" already, because it isn't just using who won or lost the match. It is using the exact score the match was won or lost by and losing 9-8 to Orcullo is going to affect your rating differently than losing 9-1 to Orcullo.
I DO think alternate break format stinks.
That will never change.
But at the same time, I'm not trying to make a rating system either.
I don't think alternate break would be a solution either simply because you still have situation where the game involves only one player.
That being said, games that have only one participant, should not be given the same value as games that have 2.
The only real solution to eliminate that would be to have Accu Stats rules, where someone breaks, and regardless of if a ball is made or not, the person who didn't break has to push out before the game can progress.
But the pool world is just never going to do that. LOL
So in light of that, the only solution is to have different values for different games won, or something to that effect, if they truly want to make a system that is accurate as possible, IMO.
Not that the system is ridiculously flawed, cause I don't think it necessarily is, but it could probably benefit from some tweaking here and there.
I'm sure appropriate algorithms could be figured out.
Only issue if they did tweak it is that more data has to be entered, which might be too time consuming, and too difficult when you consider that past data doesn't have all those statistics in the first place.
Food for thought anyway.
That will never change.
But at the same time, I'm not trying to make a rating system either.
I don't think alternate break would be a solution either simply because you still have situation where the game involves only one player.
That being said, games that have only one participant, should not be given the same value as games that have 2.
The only real solution to eliminate that would be to have Accu Stats rules, where someone breaks, and regardless of if a ball is made or not, the person who didn't break has to push out before the game can progress.
But the pool world is just never going to do that. LOL
So in light of that, the only solution is to have different values for different games won, or something to that effect, if they truly want to make a system that is accurate as possible, IMO.
Not that the system is ridiculously flawed, cause I don't think it necessarily is, but it could probably benefit from some tweaking here and there.
I'm sure appropriate algorithms could be figured out.
Only issue if they did tweak it is that more data has to be entered, which might be too time consuming, and too difficult when you consider that past data doesn't have all those statistics in the first place.
Food for thought anyway.
Yawn......:boring2:
I think a solution to the alternate break vs winner break problem would be to have a parameter that gives more weight to racks won in alternate break over winner break. It would have to be fine tuned, but eventually it would come out closer to the truth.
You do know Carl is backing me now days .
She could play a set for a lot of money .
It would be worth it if she can win .
Can't ya feel either a new nationwide or worldwide league coming soon?
Well I guess its a good thing we are in America on an American forum owned by an American then huh?
I think what he may be saying is A lot of folks don't care about who can beat someone one short set in a tourney, what really matters is who leaves with the chicken and if so I agree.
Ya never know
could be a cash cow down the road. Doesn't !matter to me cause I don't do league but that would kinda make sense. Hey BTW, I've been coming thru your neck of the woods recently and haven't stop in Che's Beach since like 06 but I should be back down that way within next month or so. Can you recommend a good place for lunch? Pool table would be cool too but def not necc. Thanks!