Mike Page, FargoRate, Greg Hogue, and the Mosconi Cup

JAM said:
I don't have time to read the entire thread, but I am a little miffed at some of the posts I did read.

Mike Page has always been a strong supporter of pool. I remember when he traveled to Maryland in my neck of the woods and was checking out pool rooms for future reference. He is one of the good guys in pool. Not sure why folks are upset about Fargo Rate. It's a rating system that was needed. Better than BCA's old rating system, which is now defunct, and the WPA rating system is kind of not up to date.

Matchroom looks to be creating a series of events and will have a rating system, I guess, based on income?

I'm out of the loop on ratings systems, I guess. That said, Mike Page is one of the good guys in pool. That's my story and I'm sticking with it.
Click to expand...
We all like Mike Page. And most of us like FargoRate.

But the guy runs this thing terribly. It took years for an app. The app barely functions. He operates on mostly closed systems with limited accessibility for leagues and tournaments. He has a closed door policy on game and player data. He doesn’t have the algorithm open for a performance calculation. He addresses unimportant critiques with in depth videos while if ignoring legitimate questions. It’s just a mess. I feel like the quality of the algorithm itself along with Mike being a “good guy” are the only things keeping this afloat. You could quadruple usage in a year with a few SIMPLE changes.

Nobody is asking Mike for anything free. We’re asking him (begging him at times) for a BETTER more transparent product.
 
FargoRate is not a version of Elo? All this time I thought it was. Learn something new everyday. Also did not know that the equation was private.

On a different note, has anyone done any research as to how FargoRate stacks up to other ranking systems such as TruSkill and Glicko/Glicko2?
 
jtompilot said:
I played a 9 ball tournament at Bangin Ballz in Vegas and a 10 ball tournament in Spokane. They didn’t show up in my FargoRate. I assume the brackets were not public?
Click to expand...
A good percentage of the people I know that run tournaments are not really interested in reporting them to Fargo. Unless they contact Fargorate and use the systems that Fargo monitors, the tournaments are not going to count. If the brackets were on paper, it's a good bet they were never meant to be included.
 
Sheldon said:
A good percentage of the people I know that run tournaments are not really interested in reporting them to Fargo. Unless they contact Fargorate and use the systems that Fargo monitors, the tournaments are not going to count. If the brackets were on paper, it's a good bet they were never meant to be included.
Click to expand...
And then there is the rest of the world, with players who train religiously, and would love to have a datapoint that tells them exactly how they currently stack up with top tier pros.

There's a reason we currently only have one world class player on the pool scene.
 
Positively Ralf said:
FargoRate is not a version of Elo? All this time I thought it was. Learn something new everyday. Also did not know that the equation was private.

On a different note, has anyone done any research as to how FargoRate stacks up to other ranking systems such as TruSkill and Glicko/Glicko2?
Click to expand...
As others have said, FargoRate is pretty different from Elo for how the ratings are determined. But once you have ratings, interpreting them and using them is the same as Elo. FargoRate is much better than Glicko/Glicko2 and also much harder to do. The blog post below attempts to explain the difference. What we do is not secret or hidden.

As skip100 says, it's the same thing as the Bradley-Terry model (aka maximum likelihood) with the additional feature that older games are considered to be more chaotic and contain less information--more so the older they are.

FargoRate -- a look behind the curtain - FargoRate Blog

News and usage hints for FargoRate.
fargorate.com
 
I ran into Spanky late in a one pocket tournament right after he got sober. He's always been good but that field has 8 or 10 players rated 700-800.
He drilled me and he played incredibly well and it was probably 2 balls better than what I had been seeing out of him for the last 20 years. Way to be Greg! I am honestly impressed and reinvigorated about the game as another middle aged, under 700 rated player.
 
Looks like posts 117 and 119 mean a way forward for German league.

@ShortBusRuss
Just cc the league board members, owners on the email back to FargoRate.
Seems like that is all you were missing to let them know it is a group effort.
Still, a simple reply to you explaining where you fell short would have resolved it a while ago.
Lessons learned.

Generally, also working with Bryan at FargoRate, it is pretty straightforward getting data in.
Having weblink to a bracket, in Challonge , Cuescore, DigitalPool etc and a brief explanation is good enough.
For 1v1 matches, we use Salotto in between tournaments and leagues when there is nothing on, or we are getting new players established.

Before end of 2022, New South Wales state in Australia will have all major associations, league, private run tournaments linked to FargoRate in some shape or form.
No more convoluted local handicaps, vulnerable to manipulation.

Another observation Russ, I see DigitalPool has a league function.
I have not tried it out.
But, DP is the only site I have seen that automatically sends results to FargoRate.
Don't know if that applies to the League module on it.
 
ShortBusRuss said:
And yet.. When measuring expected outcome between two "legit" Fargorated players, it is accurate to within a game or two, no matter how far apart the skill levels are.
Click to expand...
Over how many games? I don't think Fargorate is inaccurate, just that pool results can vary wildly. I've taken more games off strong players than I should have, and given up many more games against weak players.
 
tomatoshooter said:
Over how many games? I don't think Fargorate is inaccurate, just that pool results can vary wildly. I've taken more games off strong players than I should have, and given up many more games against weak players.
Click to expand...
The statistical word for this is "variance". Doesn't change the total percentage of games you can expect to win against either group in the long term.

I am quite sure there are times you do much better against the weak players than expected, and do much poorer against the strong players than expected.

Fargorate doesn't state that you win 100% of matches against players 100 points below you. It says on average, you will win x% of your games against those players. Which could means you get drilled the very first match, scream "SANDBAGGING!!!!!!", or "The RATING SYSTEM IS BROKE!!!!", and then win the next 3 matches or whatever, with the total won games exactly as expected.

And for a more precise answer to your "over how many games?" question... Perhaps you should look at some of the Fargorate analysis of some of the long race format gambling matches... Donny Mills got almost EXACTLY the amount of games Fargorate expected him to against that Chinese female champ he played. All while some of the incels on here were insisting she'd never beat him, because "Females and Males are in two different playing populations, so their ratings cannot be compared". I am betting Donny is a little bit more convinced in Fargorate these days after losing a bunch of money to her.

Most of the grumbling about Fargorate comes from people who don't have the first clue about how statistics, expected value, and variance work.
 
Some may recall this conversation from 2.5 years ago. [might want to review the initial post]
Greg (FR 699 at the time) was in the conversation for Mosconi Cup based on success in then-recent ranking events, and the issue was whether his skill might be outrunning his Fargo Rating. There was speculation his skill ascent might be unusually rapid because not only was he working hard on his game but he also had turned his life around more broadly.

You may have seen we recently showed that for players with many thousands of games, their Fargo Rating was more predictive of future performance than was a performance rating based on 500 recent games.

We can now go back and look at Greg's performance up to that point and also for the 1240 games he's played since then. The blue dots here are performance ratings each based on 620 games. You can see there is some statistical scatter and also a trend. The red X's are his Fargo Rating at the time of the podcast (699) and his Fargo Rating now (708). The Fargo Rating reflects all the games in a player's history but gived more credence to games played more recently, Information from old games is progressively ignored, i.e., more so the older they are.
1745936806044.png
 
You must log in or register to reply here.
Back
Top