FargoRate?

jtompilot

jtompilot

AzB Silver Member
Silver Member
I haven’t played in a 9/10 ball tournament in a long time. I’ve had a few trips to Austin recently and no one wanted to play 1P so I got in some of their 9 ball tournaments. After three of them my FargoRate was 647, what ever that means. Two weeks ago I’m in Panama City and go downtown to practice on the only Diamond table I know of. The room was deserted, the bar lady asked if I was here for the tournament. She said there’s a tournament downstairs.

So I do downstairs to check it out. It’s a 10 ball tournament on Dynamo bar boxes😝There’s only five players and they talk me into playing. $20 entry fee and they report to FargoRate. OMG that heavy mud ball gave me fits. I get a bye then win my first match 3-0. The next guy is pretty decent, he must be a 3C player because I hook him several times and he manages to make some nice hits. It’s 2-2 and I’m struggling to figure out that mud ball and I lose. On the loser side I win a match 3-0 then another 3-1. Now I have to double dip the guy who put me in the losers bracket. I win 3-2 then 3-0. I get $60 and he gets $40, I guess the bar got $20. A week later my FargoRate is 625. Oh well, I’m sure I wasn’t a 647 anyway😜
 
647-625 is about as even as rating as you can get. It's maybe a 5-4 ratio of wins to looses between those two ratings if they were two different players. Maybe even 6-5. Unless you have several 100 matches in, the rating does not have a very solid meaning, and it's based on the rating and games won on who you played, the equipment is not taken into account. It will even out at the end of a few 100 matches on different equipment and skill tends to translate good enough to other areas and equipment. A 500 is very likely to beat a 400 even on unfamiliar equipment.
 
Your rating is based on how well you perform against a player of whatever their ability is, so if you under perform, whether you win or lose, you will go down, and vice versa. For example, if you keep going hill hill with SVB in long races but lose every time, your rating is going to get real high because it shows that you are almost as skilled as an 830 rated player (Shane's current rating). And if you barely manage to beat a beginner player your rating is going to go down because you barely performed better than a beginner. Over time your rating ends up accurately reflecting your average game, not your best or worst game. At first your rating is going to bounce around a lot, but the more games you get in the system the more accurate it becomes and the less it will move with each new match. Once you have 200 games in the system (not matches, just total games) your rating is considered to be pretty accurate.
 
Last edited:
skip100 said:
Is there a questions about Fargo ratings in this post or is it just a diary of your tournament?
Click to expand...
I think the implication is it is problematic a rinkydink tournament on small sub-par unfamiliar equipment can influence his rating. He has in the past questioned whether games played on easier-than-usual equipment, like 7-foot tables, are helpful for the ratings.

In Austin, 31 of his games were against Eric Aicinena (724), Daniel Schneider (722), Junior Jueco (709), Jason Harkrider (669) and Chris Morris (659). These opponents average about 700, and JTom won 12 of the 31 games, 39%. That is in the long haul what is expected of a player rated 632.

The problem is, JTom's opponents could have caught a gear in those 31 games and strung some racks together such that JTom only won 8 of them. And then he'd look like a 546. Or JTom could have gotten some rolls and won 16 of these games, and then he'd look like a 707. If JTom had several trips to Austin, these swings would tend to average out. So if JTom won 120 games out of 310 (instead of 12 out of 31), we'd feel more confident in the ratio.

Our experience is the ratio of games JTom wins against opponents averaging 700 is pretty insensitive to lots of things, including whether they are playing 10-Ball, 9-Ball, or 8-Ball, whether they are playing on a loose table or a tight table, fast slippery cloth or a rug, and on and on. This means the fact of getting more data is the bigger deal than is the subtle differences that come from the equipment or the game.

He is a bit over a third of the way to having the minimum number of games to have a "Fargo Rating."
 
mikepage said:
I think the implication is it is problematic a rinkydink tournament on small sub-par unfamiliar equipment can influence his rating. He has in the past questioned whether games played on easier-than-usual equipment, like 7-foot tables, are helpful for the ratings.

In Austin, 31 of his games were against Eric Aicinena (724), Daniel Schneider (722), Junior Jueco (709), Jason Harkrider (669) and Chris Morris (659). These opponents average about 700, and JTom won 12 of the 31 games, 39%. That is in the long haul what is expected of a player rated 632.

The problem is, JTom's opponents could have caught a gear in those 31 games and strung some racks together such that JTom only won 8 of them. And then he'd look like a 546. Or JTom could have gotten some rolls and won 16 of these games, and then he'd look like a 707. If JTom had several trips to Austin, these swings would tend to average out. So if JTom won 120 games out of 310 (instead of 12 out of 31), we'd feel more confident in the ratio.

Our experience is the ratio of games JTom wins against opponents averaging 700 is pretty insensitive to lots of things, including whether they are playing 10-Ball, 9-Ball, or 8-Ball, whether they are playing on a loose table or a tight table, fast slippery cloth or a rug, and on and on. This means the fact of getting more data is the bigger deal than is the subtle differences that come from the equipment or the game.

He is a bit over a third of the way to having the minimum number of games to have a "Fargo Rating."
Click to expand...
Thanks for explanation, I can see how the rating swings with a low game count. Jason Got a rolI or two but at the end of the day I just couldn’t get the job done. I need some more trips to Austin to get my game count up.
 
mikepage said:
I think the implication is it is problematic a rinkydink tournament on small sub-par unfamiliar equipment can influence his rating. He has in the past questioned whether games played on easier-than-usual equipment, like 7-foot tables, are helpful for the ratings.

In Austin, 31 of his games were against Eric Aicinena (724), Daniel Schneider (722), Junior Jueco (709), Jason Harkrider (669) and Chris Morris (659). These opponents average about 700, and JTom won 12 of the 31 games, 39%. That is in the long haul what is expected of a player rated 632.

The problem is, JTom's opponents could have caught a gear in those 31 games and strung some racks together such that JTom only won 8 of them. And then he'd look like a 546. Or JTom could have gotten some rolls and won 16 of these games, and then he'd look like a 707. If JTom had several trips to Austin, these swings would tend to average out. So if JTom won 120 games out of 310 (instead of 12 out of 31), we'd feel more confident in the ratio.

Our experience is the ratio of games JTom wins against opponents averaging 700 is pretty insensitive to lots of things, including whether they are playing 10-Ball, 9-Ball, or 8-Ball, whether they are playing on a loose table or a tight table, fast slippery cloth or a rug, and on and on. This means the fact of getting more data is the bigger deal than is the subtle differences that come from the equipment or the game.

He is a bit over a third of the way to having the minimum number of games to have a "Fargo Rating."
Click to expand...
Are the actual matches/ subsequent scores leading to players' ratings available anywhere, either publically or via private request?
 
Black-Balled said:
Are the actual matches/ subsequent scores leading to players' ratings available anywhere, either publically or via private request?
Click to expand...
You can see your own, and they now give you the option to make your own matches public. It shows who you played, where and when, and the score.

I know with Facebook and other services there’s been a lot of attention to privacy on the internet, and I’m sure these guys don’t want to go anywhere near those issues. I can imagine, however unlikely, that they could fear someone stalking someone because they figure out that they play in a weekly tournament.

But I think they should consider at least making all scores and opponents public. Maybe they could remove the where/when from public view unless you opt in.
 
BRussell said:
But I think they should consider at least making all scores and opponents public. Maybe they could remove the where/when from public view unless you opt in.
Click to expand...
Something which may or may not be a consideration for Mike is that as soon as you do this it becomes possible to work out all the finer details to the exact algorithm behind FargoRate.
 
Poolplaya9 said:
Something which may or may not be a consideration for Mike is that as soon as you do this it becomes possible to work out all the finer details to the exact algorithm behind FargoRate.
Click to expand...
It's on wikipedia so the math is not really a trade secret: https://en.wikipedia.org/wiki/Bradley–Terry_model

the bigger issue is protecting the true "killer app" of Fargo which is the massive game database they have spent years collecting.
 
skip100 said:
It's on wikipedia so the math is not really a trade secret: https://en.wikipedia.org/wiki/Bradley–Terry_model
Click to expand...
While the basis of what he is doing is no secret (and obvious), and he has in fact shared it, he is also doing things differently than any previous system or algorithm has done. On top of that, there are TONS of things where you must make choices about how you want to do them or handle them or what "setting" you want to use (and where infinite choices are available and no two people would inherently be making the exact same decisions about how to handle these tons of things). Between those two above things there is still quite a lot of secret sauce there that could in fact be worth protecting.
skip100 said:
the bigger issue is protecting the true "killer app" of Fargo which is the massive game database they have spent years collecting.
Click to expand...
The database does indeed also have value.
 
If by "have value" you mean "is unique in the pool world and all but impossible to be replicated by anyone else".

Code is cheap. Data is expensive.

APA could potentially equal or better Fargo's game database but that's unlikely to happen for many reasons.
 
You must log in or register to reply here.
Back
Top