A Questoin For FargoRate
- Thread starter watchez
- Start date
I do, but based on your posts....I suspect English is not your native language.
As for the pretenders... You know what they say...takes one to spot one.
Usually hundreds. Several hundred people will reflect votes/opinions of a hundred million people very closely.
But I don't think sampling is an issue with these Fargo ratings. They probably want every tournament entered into the system, and they'll probably get them.
If it works for chess then it will work for pool. That's pretty much the bottom line here IMO.
And I do believe that it will work.
I looked up the ratings of a bunch of people that I know from my years traveling around and so far the ratings seem to be pretty accurate as to how I stack up against them based on what I know first hand of how they play.
So this tells me that I can probably count on them to be fairly right when I am looking at a stranger's rating.
And I do believe that it will work.
I looked up the ratings of a bunch of people that I know from my years traveling around and so far the ratings seem to be pretty accurate as to how I stack up against them based on what I know first hand of how they play.
So this tells me that I can probably count on them to be fairly right when I am looking at a stranger's rating.
Agreed, new technology (such as a scoring app) will reduce the marginal cost of Data collection, but when people start talking about implementing an accustats type rating system for all major events...this is a silly pipe dream.
A scoring app, as far as players being able to enter their own scores, with no verification that the match ever took place, will completely destroy the integrity of the data. Even if you had a way for both players to verify the match happened, there could still be fake accounts, collusion between players, ridiculous matches that may have not even involved pool at all, etc....
Getting fargo to be a respected and universal rating means the matches recorded need to be official and verified in some way. ELO ratings would be a joke in chess today if the FIDE allowed players to enter their own games.
You don't need, or really want, every single game ever played, recorded. Quality and accuracy is much more important. Over time, there will easily be more than enough data to accurately represent people's relative skill levels.
Getting fargo to be a respected and universal rating means the matches recorded need to be official and verified in some way. ELO ratings would be a joke in chess today if the FIDE allowed players to enter their own games.
You don't need, or really want, every single game ever played, recorded. Quality and accuracy is much more important. Over time, there will easily be more than enough data to accurately represent people's relative skill levels.
Pipe dream or not. it is a way more accurate way to rate players. If you want to rate you have to look at the performance not results. This is not chess and each player does not get the same luck. You can play perfect and lose or play bad and win. That is why nobody matches up based on results... not if they are smart.
Ranking, which is something different, can he done from results but again in order to be accurate you need all of the results. WPA already has a ranking system that ranks with the wpa data. Is it going to tell who is truly the best? It will get close but not as close as something like accustats. Is it going to tell who has been winning the events and accumilating points? Yes and that is all pro pool needs anyway imo.
Last edited:
I meant in conjunction with tournaments. As in they enter their scores in real time or something like that.
Yes, data integrity is a very important issue to us. And of course we cannot ever guarantee every score is accurate.
There will be no self reporting. We are working on automated mechanisms to bring in data. Generally, expect data that comes in to be publicly available somewhere. If data from your weekly tournament is to come in, the brackets for your weekly tournament with the scores should also be displayed online. That in-the-sunshine-with-your-peers approach will go a long way to thwart or dissuade fraudulent reporting.
As to the marginal cost of bringing new data, this is the opposite of the situation for a survey where the first 100 is easy, the next 100 a little harder, etc. There have at this point been several thousand person-hours getting the first million games. I suspect the next several thousand person-hours will get in ten or more times that number of games.
Wouldn't it be great if we had an accu-stats number for every pro in all of their major tournaments, like we have free throw percentage or batting average? But the accu stats ratings are tough to calculate - you have to watch every shot of the whole game to calculate it.
Agreed, there can be a significant amount of luck in any result. However, the basic premise behind Fargo ratings (along with every other statistical estimator) is that under an appropriate set of conditions, for any given sample size, we can place reasonable boundary on how much influence luck (or randomness, or noise, etc...) can have.
Yes but the idea is that it's a cumulative picture of performance over many games. So while you can play bad and win and play good and lose over time your average performance shows in the results you get consistently against the crowd of people who are also doing the same task as you, playing to win.
If you play only opponents who are not rated, then FargoRate has no way to connect you to the rest of the world--cannot give you a rating that is meaningful outside your little group. However, if a few of those opponents later play against established people, then not only do they come into line, but so do you.
This is an important distinction between what we do and what a run-of-the-mill implementation of an ELO scheme does.
Interesting perspective. Sounds like a prescription for homeless "smart" people to me...
Yes. When Justin Bergman gets to the finals of an event against John Morra, a lot of people are watching and paying attention to individual shots and decisions and the like. But that is just not very much information for players that are close.
There is much more information buried in the matches these players played to get there. The differences between these players is based on subtle differences in shotmaking, subtle differences in a tendency to overrun position, subtle differences in the ability to kick at the right part of the ball, subtle differences in decision making, and a host of other things.
These differences in the long haul manifest themselves as tendencies to beat the lesser pros 11-7 rather than 11-8 and the shortstops 11-5 rather than 11-6--things that happen when we're not watching.
Judging by results can get you broke also.
You have to know how the play when you know that you know if they played over their head to beat so and so or played real bad to get beat by some other guy.
Last edited:
Banks
Banned
I think i understand what he's saying and i agree. I've checked the ratings of several people and they're not quite in line with how they would match up. It could be due to a low-ish robustness, but it's still there. The other side to that is FR is about tournaments, not money games. They're two different beasts and it's not a bad thing to keep them separate.
Watchez, Johnny has played tournaments near me in Canada that aren't on Mike Pages list, as well as who knows how many more. He's an active player all right, but the money just isn't there to be on the road all the time.
Sent from my C811 4G using Tapatalk
So he is sneaking around - that makes sense.