Updated FargoRates are out

[Poolplaya9]

To make this short......would you take Siming even against Sky Woodward, Rodney Morris, Ralph Souquet, Nick Vandenberg, Corey Deuel, Thorsten, Earl, Mika, Archer, John Morra, Kiamco, Billy Thorpe.........etc??

They are all within 10 points of Siming's fargo rate.

According to your post, you have a lot of faith there are enough connections in which to accurately asses her rating, therefore she's even money against all those players. So would you honestly be comfortable wagering a large amount of money(hypothetically, I'm not making this about action) on her against all those within 10-20 fargo points?

This is going to be long. Let's start by listing their fargoratings as of this moment to make the discussion easier.

Siming Chen 778

Warren Kiamco 788
John Morra 788
Mika Immonen 788
Sky Woodward 785
Ralf Souquet 784
Nick VanDenBerg 783
Rodney Morris 779
Earl Strickland 778
Thorsten Hohmann 777
Corey Deuel 776
Johnny Archer 772
Billy Thorpe 764

1. Fargorate has a margin for error that is usually probably less than 5 or so points but that can in some specific cases be much more, such as when a player's abilities are rapidly improving or declining, or when a player doesn't have enough games in the system, and other cases such as those.

2. Unless I have current extraneous info that I believe will make a difference, such as who has the flu, who is jet lagged or tired and doesn't play well tired, who seems to have who's number, who had their dog die yesterday, who plays better under a certain type of pressure etc, then I don't tend to bet on the player who is less favored to win so no I wouldn't be betting on Siming against any of those people who are rated higher than she is, which is everybody from Rodney on up.

3. I don't tend to bet on coin flips either because coin flips are for suckers. With the margin for error taken into consideration that rules out everybody from Johnny on up.

4. That leaves Thorpe, who is rapidly improving and because of this he is underrated a bit by fargo IMO. The way fargo works, anybody rapidly improving or declining in skills will always have fargo trailing behind a little trying to "catch up" to what their true rating really is due to the nature of how they average all performances over the last ten years if they have history that goes back that far. In Thorpes case, results from when he was not nearly as good as he is now are still being averaged in and his true rating is not quite being reflected yet. Fargo largely addresses this issue by more heavily weighing results the more recent they are but the old results from when he wasn't nearly as good are still able to have a little effect and I am guessing his true level of play is mid 770s-ish now, and with the margin for error that puts him in the coin flip category with Siming as well and as said before I don't tend to sucker bet on coin flips.

5. Because of the small margin for error in ratings, and because of the variances in how people perform on any given day, you probably need to see something like at least a 10 point or more difference in ratings in my estimation before you can even begin to feel that someone has even a slight edge in a match up enough to be worth betting on, particularly if it isn't a really long race or ahead set.

So would I bet any of these matches? Nope, but not because I don't believe their ratings are pretty accurate, but because all those matches are either coin flips where either could win on any given day or they show her as the clear underdog who is going to lose a bit more often than she wins.

 

[Poolplaya9]

How do the top 100 lists work?

If I click on the top 100 American players, the lowest is at 707. But there are other American players (e.g. Chris Bartram, Keith Bennett) that have higher Fargo scores and aren't in the top 100.

To be included on one of the "Top xyz" lists a player has to have a minimum of 300 games in the system, and at least 150 of those games have to have been in the last 2 years.

 

[Bob Jewett]

Kristina Tkach RUS 686 who is 16 or so, is rated higher than every US female player of any age. I think there is a statement there about other countries organized billiards programs.

Poland has an official pool program in high schools. At the American 14.1 Championships which just finished in Richmond, Konrad Juszczyszyn (YOOSH-ta-shin, more or less) who is 23, beat Dennis Orcullo in the finals.

 

[donuteric]

I'm assuming here, but maybe they don't have enough games required to be considered a reliable score?

Bartram has plenty of games in and he should be up there but perhaps he hasn't been playing much in Fargo "enabled" tournaments recently.

However, my recent experience with some of these regional tournaments using Fargo is that matches never get submitted.

 

[one stroke]

This is going to be long. Let's start by listing their fargoratings as of this moment to make the discussion easier.

Siming Chen 778

Warren Kiamco 788
John Morra 788
Mika Immonen 788
Sky Woodward 785
Ralf Souquet 784
Nick VanDenBerg 783
Rodney Morris 779
Earl Strickland 778
Thorsten Hohmann 777
Corey Deuel 776
Johnny Archer 772
Billy Thorpe 764

1. Fargorate has a margin for error that is usually probably less than 5 or so points but that can in some specific cases be much more, such as when a player's abilities are rapidly improving or declining, or when a player doesn't have enough games in the system, and other cases such as those.

2. Unless I have current extraneous info that I believe will make a difference, such as who has the flu, who is jet lagged or tired and doesn't play well tired, who seems to have who's number, who had their dog die yesterday, who plays better under a certain type of pressure etc, then I don't tend to bet on the player who is less favored to win so no I wouldn't be betting on Siming against any of those people who are rated higher than she is, which is everybody from Rodney on up.

3. I don't tend to bet on coin flips either because coin flips are for suckers. With the margin for error taken into consideration that rules out everybody from Johnny on up.

4. That leaves Thorpe, who is rapidly improving and because of this he is underrated a bit by fargo IMO. The way fargo works, anybody rapidly improving or declining in skills will always have fargo trailing behind a little trying to "catch up" to what their true rating really is due to the nature of how they average all performances over the last ten years if they have history that goes back that far. In Thorpes case, results from when he was not nearly as good as he is now are still being averaged in and his true rating is not quite being reflected yet. Fargo largely addresses this issue by more heavily weighing results the more recent they are but the old results from when he wasn't nearly as good are still able to have a little effect and I am guessing his true level of play is mid 770s-ish now, and with the margin for error that puts him in the coin flip category with Siming as well and as said before I don't tend to sucker bet on coin flips.

5. Because of the small margin for error in ratings, and because of the variances in how people perform on any given day, you probably need to see something like at least a 10 point or more difference in ratings in my estimation before you can even begin to feel that someone has even a slight edge in a match up enough to be worth betting on, particularly if it isn't a really long race or ahead set.

So would I bet any of these matches? Nope, but not because I don't believe their ratings are pretty accurate, but because all those matches are either coin flips where either could win on any given day or they show her as the clear underdog who is going to lose a bit more often than she wins.

Id bet all of them in a long race against her in winner break 10 ball all day long

1

 

[hang-the-9]

To be included on one of the "Top xyz" lists a player has to have a minimum of 300 games in the system, and at least 150 of those games have to have been in the last 2 years.

Maybe the last 2 years is the issue, he has an 1802 robustness so has plenty of games.

 

[fastone371]

U can get action on a case by case match.. I have no proof but I have a feeling fargo is starting to fall apart.

Thats a pretty pessimistic view, what makes you feel that Fargo is starting to fall apart? Fargo may be new to these forums but it has been in use long enough that it is not new anymore. If Fargo was going to flop I would guess it woulda happened in the first couple of years.

 

[fastone371]

So what does the top 100 list really tell us? My bet is, if you stop 1,000,000 people at random on the streets of America, the only pool player they will mention is "The Black Widow", Jeanette Lee! So who is number one? You know my answer. Fargo be damned.

Lyn

It seems like you are confusing player rankings with pool promotion. I find it highly unlikely that Fargo was ever intended to promote the game of pool. If the average American can not name the top pool players that is not the fault of Fargo, it is the fault of whomever it is that gets paid to promote pool. Oh wait, that position probably does not exist, but whomever is responsible for promoting pool is failing incredibly. It probably starts with all of the organizers that tried to get fat from the players prize money and pushed them in a poor direction telling them they would earn more if they listened to the guy lining his pockets.

 

[Poolplaya9]

So to say, Karen beat earl, and Simeng beat Karen, therefore Simeng would beat Earl is not only cheapening what Karen did, but it's frankly misleading to the entire world to where folks believe Simeng is as good as Earl (because Fargo Rate has them identical).

I grossly simplified my explanation of how Fargorate works so that people could understand the basic concept of some of the ways in which it works, but it gets far more complicated than that and there are far more reliable connections that what you are able to conceive. You are also painting it out above almost as if Karen's rating has been determined based on how she performed against Earl in one match and of course that isn't remotely close to the case. She is rated based on how she has played against everybody she has ever played, for all the times they have played (that fargo is aware of), and against all the people they have played, and so on until essentially everybody is included. Not to mention that just in regards to Earl alone Karen has probably played him more than once, and she has certainly played lots and lots of people that have played Earl lots and lots of times so there is lots and lots of data through lots and lots of connections just between Karen and Earl alone but of course she isn't rated based only on her connections to Earl and the many connections between her and Earl is but a very minuscule percentage of how her rating has been derived.

Similarly, Siming's connections to Karen are just massively extensive, as are her connections to Earl through this absolutely massive web of connections that are "averaged" together to very precisely and accurately tell you how she compares to Earl (and everyone else in the system).

Let's put it this way, Ko pin Yi played at Turning Stone and didn't even make it into Sunday. That's a difficult tournament and Karen has performed well in it, there are times she doesn't do so well. But she plays in it and Simeng doesn't.
How would Simeng fare? We will never know until she plays.

Well we do know quite well actually exactly how her skill level compares to the players there. Now could other extraneous factors such as being intimidated by the men or the mixed nature of the event cause her not to play to at least her average level? Of course, but this applies to anyone, not just females, and not just Siming, and at her level of play with her mental fortitude I doubt she would be drastically affected but certainly anything is possible.

Keep in mind though, Fargo is there to rate someones average skill level across a variety of conditions and circumstances so that it is useful under all conditions, not just on how they would do under more pressure than normal, or how they would do if grandma just died, etc, yet that is exactly what you are essentially arguing that it should be able to do. We could have a Fargo that told us only how players did under the highest pressure situations but then it wouldn't be as accurate or useful for the other 95% of circumstances. The information that is overall most useful is to know how someone does on average against a variety of opponents in a wide variety of conditions and circumstances, and that is exactly what Fargo provides. How an atypical pressure situation or grandma dying or any other specialty situation might cause any deviation from their "average performance level under a variety of conditions and circumstances" is for us humans to contemplate when considering specific match ups.

But if Simeng only matches up against Karen in tournaments with magic rack and alternating break, how can any hypothetical of Simeng versus Thorsten or Earl be taken seriously when she doesn't even compete like Karen.

I know you don't understand and accept it yet but alternate breaks does not ever change who wins over time. It only affects by what margins somebody wins. As for magic racks, I don't think they probably give a lot of advantage or disadvantage to most people over time either, and whatever exists gets averaged in with all the other varying conditions and circumstances that can affect things.

That's where I'm still not all in with Fargo. To lump all pool matches and outcomes together and derive results without proper context is just not valid in my opinion.

As explained a bit above, having an average of performance under a variety of conditions and circumstances is exactly what you do want if you only have one rating system. Now Mike could have 50 (or even thousands) different FargoRating systems each specifically tailored to specific conditions and circumstances if he wanted, like he could have one tailored to show who performs best under the highest pressure situations, who is performing best only in the last two months, who performs best in 9 ball on the 10 foot bigfoot table when the pockets are 4.25 inches when the temperature is a constant 68 degrees with a relative humidity of 64% and when a close family member has died within the past week, and another 47 more specialty fargorating systems (he could even do hundreds, or even thousands of them to make them specialize in every conceivable variable we can think of). But for a single general purpose system it is by far most useful if it gives the average performance level under a variety of conditions and circumstances and we can just apply our knowledge of any special circumstances or conditions to determine how much affect we think it will have on someone in a specific match up.

 

[mikepage]

How do the top 100 lists work?

If I click on the top 100 American players, the lowest is at 707. But there are other American players (e.g. Chris Bartram, Keith Bennett) that have higher Fargo scores and aren't in the top 100.

There is a "current player" requirement to be on these lists.

A player must have 300+ games total and must have 150 or more recorded in the last two years.

So it is not the we think Bartram or Bennett don't play at the speed of the people on this list, I'm quite sure they do. But they just don't meet the current-player requirement

 

[AtLarge]

... He claims there are thousands of games between men and women in which Fargo rate uses the connections. I rarely hear of many matchups between men and women. He uses Corr as an example, but she is one of the few that you hear about. So I’d be curious where all of these thousands of games come from, unless they come from a very small pool(no pun intended), of women who compete against men. If so, I would think that would be a very fragile number to be basing the connections on. ...

Back in August Mike Page did another Babu/Bibi comparison -- for Jennifer Barretta, who has played a lot of open events. Mike reported:

"If we only look at Jennifer Baretta's 900 games against men, she is shooting at 651 speed.

In her 1100 games against women, she is playing at 648 speed --the same."

 

[vapoolplayer]

And this is the problem I have with Fargo rate that I'm still trying to wrestle with.
To say Karen played Mika is not placing it in the right context. She most likely played him on the joss tour.
When Karen defeated Earl Strickland at the Turning Stone Classic in the quarterfinals, I watched her charge through the field for 4 days, and then go through the one loss bracket which means more matches played (and remember, this is wooden rack, opponent racks and winner breaks). It was an example of her persistence, fitness, and skill that she could get there. When she took down Earl, at that point, on a Sunday at turning stone, that was a HUGE accomplishment.

So to say, Karen beat earl, and Simeng beat Karen, therefore Simeng would beat Earl is not only cheapening what Karen did, but it's frankly misleading to the entire world to where folks believe Simeng is as good as Earl (because Fargo Rate has them identical)

Let's put it this way, Ko pin Yi played at Turning Stone and didn't even make it into Sunday. That's a difficult tournament and Karen has performed well in it, there are times she doesn't do so well. But she plays in it and Simeng doesn't.
How would Simeng fare? We will never know until she plays.

But if Simeng only matches up against Karen in tournaments with magic rack and alternating break, how can any hypothetical of Simeng versus Thorsten or Earl be taken seriously when she doesn't even compete like Karen.

That's where I'm still not all in with Fargo. To lump all pool matches and outcomes together and derive results without proper context is just not valid in my opinion.

Pretty simple to test. I'm sure Mike and company keep track of how often their rating predict the winner of a set.

While I may not totally agree with the connections between men and women being enough to rate them side by side, I have to trust their math is correct since it seems like a fairly refined analysis.

I'm thinking they likely predict the outcome of a match much more often than they miss the outcome. If they didn't, I'm sure they would have went back to the drawing board.

If they do predict the winner much more often than not, then they have proved your opinion you can't lump games together as wrong. If they predict it at a coin toss or lower, then that would prove your opinion correct I believe.

 

[vapoolplayer]

Back in August Mike Page did another Babu/Bibi comparison -- for Jennifer Barretta, who has played a lot of open events. Mike reported:

"If we only look at Jennifer Baretta's 900 games against men, she is shooting at 651 speed.

In her 1100 games against women, she is playing at 648 speed --the same."

It would seem that I am wrong then.

I'd be curious as to the percent of matches correctly predicted by Fargo in the last six months to a year

 

[Bob Jewett]

...
I'd be curious as to the percent of matches correctly predicted by Fargo in the last six months to a year

It's not clear what is meant by "correctly predicted" here. If two players are 10 rating points apart, Fargo will predict that each is close to 50% to win the match. What outcome would count as a "correct" prediction?

Even if you look at only matches that are 100-point mismatches, Fargo will predict that the underdog will win (for example) 10% of the time. I think that a "correct" prediction would be if the actual fraction of victories by all the underdogs was about 10%.

 

[mikepage]

Poland has an official pool program in high schools. At the American 14.1 Championships which just finished in Richmond, Konrad Juszczyszyn (YOOSH-ta-shin, more or less) who is 23, beat Dennis Orcullo in the finals.

Yooshtashin had one of the top few specific performance ratings (SPRs) at the US Open this year (some combination of tough draw, losing close matches, and winning matches decisively), though I don't have those numbers in front of me. Jayson Shaw had the best. Warren Kiamco was up there pretty high too.

 

[mikepage]

Billy Thorpe did not budge up much.. For beating Wu and I think Biado maybe.

He was one of the bigger movers... Here is a facebook note that has a chart of some of the key changes

 

[vapoolplayer]

It's not clear what is meant by "correctly predicted" here. If two players are 10 rating points apart, Fargo will predict that each is close to 50% to win the match. What outcome would count as a "correct" prediction?

Even if you look at only matches that are 100-point mismatches, Fargo will predict that the underdog will win (for example) 10% of the time. I think that a "correct" prediction would be if the actual fraction of victories by all the underdogs was about 10%.

Correct, evenly matched players would be 50/50 so it’s a wash. As you said, it would need to be something where someone was an obvious favorite, but not by a landslide.

 

[Poolplaya9]

Correct, evenly matched players would be 50/50 so it’s a wash. As you said, it would need to be something where someone was an obvious favorite, but not by a landslide.

How closely matched they are is immaterial really. What is important is what percentage of the time people won compared to what percentage of the time FargoRate had predicted they would win.

 

[vapoolplayer]

How closely matched they are is immaterial really. What is important is what percentage of the time people won compared to what percentage of the time FargoRate had predicted they would win.

Makes sense, as 50/50 matches wouldn’t be considered a predicted win. I’d imagine anything within the margin of error that puts it inside the 50/50 area would also be omitted.

 

[donuteric]

It's not clear what is meant by "correctly predicted" here. If two players are 10 rating points apart, Fargo will predict that each is close to 50% to win the match. What outcome would count as a "correct" prediction?

Even if you look at only matches that are 100-point mismatches, Fargo will predict that the underdog will win (for example) 10% of the time. I think that a "correct" prediction would be if the actual fraction of victories by all the underdogs was about 10%.

The probability of winning given 2 players' fargo rates is calculated in an even race setting. The algorithm should be able to estimate a "fair" handicapped match. Whatever the actual score ends up to be, we should be able to gauge the correctness of the prediction if a match ends up hill-hill (or close enough) in the in-silico predicted fair match.