FargoRate Alex Pagulayan
- Thread starter alphadog
- Start date
If the player with a 70% chance of winning won every time, then he wouldn't have a 70% chance of winning.
We are in infancy in some areas, firmly established in other areas, and somewhere in between in others.
So here is the situation. In all cases, with more data a rating is more reliable. Generally once a player has 200 games or so in the system, the rating is decent--not great but decent, and we call it "established."
This, however, assumes the games are played against opponents of known rating, established opponents. When many of your 200 games are played against opponents for whom we have little information, then your 200 games are not as good for establishing your rating. This is more likely to be the case in the areas or regions where we are light.
That's the bad news. The good news is your rating will get better--more reliable--even if you don't play more. That is because as your former opponents play more, we get a better understanding--looking back--of how to interpret your wins and losses against them. This process is part of what we call the "ab initio Global Optimization" and is why we need so many computers to do it and is why what we do is far superior to the standard ELO-type scheme like FIDE uses for chess.
More data makes us better, always.
Once again thanks for explaining.
Problem is the naysayers will just ignore your post, so you are just preaching to the choir.
Preach on brother
I didnt Start this thread to knock the fargorate. I think such a system is well past due. I wonder,how,when,and why,some info is used,or not used? I questioned the OmegaTour, info came in. Thats great but how do you use info from a handicapped trnmnt?
As FargoRate gets more info, it will be interesting following the players at all levels.
There is old racetrack lore that you bet your cash on horses moving UP in company...
....and lay off on horses moving DOWN.
I think this will apply to pool players also.
pt...hoping to find a player who moves from 550 to 950...I'll be rich..:happydance:
There is old racetrack lore that you bet your cash on horses moving UP in company...
....and lay off on horses moving DOWN.
I think this will apply to pool players also.
pt...hoping to find a player who moves from 550 to 950...I'll be rich..:happydance:
I am currently at 550... you just might get your wish
I'll take the under
I've got a great deal for you: Let's take an upcoming tournament, say the US Open in a few months. In each match, you give me $10 each time Fargo correctly predicts the winner, and I give you $10 each time it doesn't. Deal?
Excellent question. Take this match from the Omega Tour for instance. Robert needed to go to 8, and Ian needed to go to 6. Ian won the match, even though Robert won more games. We actually don't care about this. We don't care what they needed to get to to win, and we don't care who won the match. The information we use from here is the following:
Robert and Ian played 13 games, with Robert winning 7 and Ian winning 6.
So handicapped tournaments are not a problem for us provided
(1) we are not including "games on the wire." In other words we need actual game results.
(2) There are no ball-spots (like wild 8 and the breaks). That makes the game useless to us.
Here is another advantage of what we do over the usual ELO-type schemes. What happens when we recognize we have bad data? (e.g., a forfeited match was recorded as 7 - 0 win, or we now realize a tournament we included earlier used ball spots, or games were assigned to a wrong person). In the other approaches, those problems permeate through the whole system and there is no way to undo their effects other than to include more good data and overwhelm their influence. For us, with the optimization, the next day it is like the data was never there.
Attachments
I'm considering opening a carwash....and stealing the cars.
pt....read The Gang Who Couldn't Shooot Straight
Do you understand what "70%" means?
The problem is when the lessor player decides he can't win and dumps the last couple games , or for instance the better player plays down to his competition , players aren't stupid if thier trying to manipulate the system i don't see a catch fail ,, if a player knows he's on the wire of moving up over a said cut off point that puts him in a higher level how can you trust those matches he loses
1
1
With the money in pool, meaning lack of, the cost does not out way the benefits. If you think otherwise, go for it, and let us know the outcome.
It always amuses me how people always have excuses when they lose. I dumped, I got bad rolls, I play down to the competition, etc.
Bron you may not like or not understand what 1stroke is saying but what he is saying is true and happens every day whether you are privy to it or not.
Things like this happening occasionally have pretty small influence. What we capture is an averaged performance over lots of games/matches. Suppose in a small fraction of your matches you give up or were drunk or didn't care or had something else on your mind. Consider this: In ALL those other matches, the rest of them, you sometimes benefited from playing an opponent who gave up or was drunk or didn't care or had something else on his mind. I'm not saying these things don't happen. It's just that their impact tends tends to be small.
There is no catch fail. A player can dump and have his or her rating influenced by it.
One of the benefits of having done this in my area for almost 7 years now is I can go back and look at the players who in the early going raised the same concerns you are raising now. And there were plenty in the beginning. People talking about hiding in the weeds, people talking about dumping or otherwise finding a way to manipulate the system and so forth. And when I look at those players now, 6 years later, I can ask myself a couple different questions. (1) Are they still talking like that? talking about manipulating the system or accusing others of doing it or having others accusing them of doing it? Answer: No. (2) When we analyze statistically the data for those players (who were going to blow off league to keep their rating low or whatever), do we find irregularities? Is their rating considering only league data or only the low-entry-fee weekly tournament different from their rating considering bigger events? Answer: No.
The last thing you mention is someone who is at 600 doing something to get to 599, or someone who is at 599 doing something to avoid getting the 600. This, I think, is a valid concern if we do something to create SPECIAL ratings. This is the reason we discourage, for example, having a big national event for "under 600." It is fine to have a 540 and under tournament, or a 620 and under tournament, of an under 400 tournament. That is happening more and more as regions are benefiting from getting players established. But these tournaments need to be mixed up. 520 and under this month, 570 and under next month, and so forth.
I think one exception to this caution might be defining an elite level, like having a rating above which a player may not enter an amateur competition like the SBE event. If you said nobody over 720 or 710 or 700 or (or nobody over the level of the US #100 player), then it is only a few players right near each bubble who even conceivably might be in a position to slip under. I say so what. Live with it. I think there probably will be none of it going on, though.
If I could show you my email inbox and show you every email I have gotten from shortstop and pro-level players, you might be amazed that in every single case the concern is why isn't their rating higher. "Do you have that time I double-dipped Sylver Ochoa?"