Go Back   AzBilliards.com > Main Category > Main Forum
Reload this Page A chink in the armor of Fargorate
Reply
Page 2 of 5 12 34 Last »
 
Share Thread Tools Rate Thread
Old
  (#16)
mikepage
AzB Gold Member
mikepage has a reputation beyond reputemikepage has a reputation beyond reputemikepage has a reputation beyond reputemikepage has a reputation beyond reputemikepage has a reputation beyond reputemikepage has a reputation beyond reputemikepage has a reputation beyond reputemikepage has a reputation beyond reputemikepage has a reputation beyond reputemikepage has a reputation beyond reputemikepage has a reputation beyond repute
 
mikepage's Avatar
 
Status: Offline
Posts: 3,654
vCash: 500
iTrader: 0 / 0%
Join Date: Dec 2004
Location: Fargo, ND
   
01-12-2018, 10:16 AM

Quote:
Originally Posted by JC View Post
A 100 point spread between players predicts that the higher player will win two games for every one game the lower player wins.

But all games are not created equal and all 100 point spreads are not linear.

[...]
This is an interesting issue, and in principle this concern is legitimate. That is, there is nothing inherent that requires the ratings work the same way for different games.

In fact it is not hard to invent a game for which they clearly do not. Here is an example. I'm going to invent a game called SIETE. A single game of SIETE consists of a race to 7 9-ball. The first one to 7 in 9-Ball wins a single game of SIETE. The 100-point-lower player using current Fargo Ratings will not win a third of the games of SIETE as expected but rather will hardly ever win a game. So mixing 9-Ball data and SIETE data would be apples and antelopes.
So what then is the justification for mixing 8-Ball and 9-Ball?

Fargo Ratings started out described as a 9-Ball rating system (2002 Billiards Digest article). Then by 2010 a simplified version of it was being used as an 8-Ball rating system, with 13,000 games amongst 320 players. Over the next several years we started collecting some 9-Ball data but didn't consider combining the data for two reasons: (1) we bought into the conventional wisdom that people's 8-Ball and 9-Ball speeds were two completely different things, and (2) the kinds of things that John mentions here.

By the fall of 2014, we had 20 times as much data and 20 times as many players, and we were doing separate ratings for 9-ball (and 10-Ball) and for 8-Ball. And in fact we were doing separate ratings for different table sizes. We started noticing that the players largely had the same ordering and largely had the same rating gaps. So we started investigating. Did 8-ball and 9-Ball work the same? Did 7-foot and 9-foot ratios work the same? We had enough data to determine that remarkably so they did.

There is a huge incentive to combine data if it is at all reasonable to do so because many of your opponents who would be unestablished on, say, 9-Ball on 7-foot tables, are established when we consider other games. Think about this. Suppose you are offered a 7-foot 9-Ball big gambling match against some unknown opponent named Vilmos, and you ask a friend how Vilmos plays. Your friend says he has never seen Vilmos play on a 7-foot table, but on a 9-foot table he plays even with Oscar Dominguez. Are you going to think this is irrelevant knowledge? Of course not. Combining games and table sizes allows the FargoRate optimization to take advantage of similar situations, and these situations are the rule rather than the exception.

JC mentioned that I often use data from pro players in my analytical examples. There are two reasons for this. First we have a lot of data on top players. And second, people are familiar with the names and have seen them play.

What I've done here is to look at my own data. I am rated 623. So I think to investigate John's point, I should compare myself to someone rated 523 in both 8-Ball and 9-Ball. It is hard to get enough numbers to be statistically meaningful. But here is what I did. I looked at my own record against opponents within 20 points of 523 (from 503 to 543) in both 8-Ball and 9-Ball. This should average about a 100 point gap. At 100 points, I am expected to win 67% (two thirds) of the games.

I played 366 games of 9-Ball against opponents 503 to 543 and 981 games of 8-Ball.

I won 66.9% of the 9-Ball games
I won 68.0% of the 8-Ball games

It works out about as expected.

Now, with 6.4 million games in the system, we are up another factor of 20 from the Fall of 2014. We have plenty of data to analyze all sorts of things. We are not wedded to any particular thing we are doing. If the data suggests there is a better way, there is no reason we wouldn't just switch to that way.

Taking deep dives into our data is part of the fun for numerophiles like me ;-)


********
mike page
http://www.fargorate.com
player ratings for the world
********
  
Reply With Quote

Old
  (#17)
jasonlaus
Rep for Smorg
jasonlaus has a reputation beyond reputejasonlaus has a reputation beyond reputejasonlaus has a reputation beyond reputejasonlaus has a reputation beyond reputejasonlaus has a reputation beyond reputejasonlaus has a reputation beyond reputejasonlaus has a reputation beyond reputejasonlaus has a reputation beyond reputejasonlaus has a reputation beyond reputejasonlaus has a reputation beyond reputejasonlaus has a reputation beyond repute
 
jasonlaus's Avatar
 
Status: Offline
Posts: 6,146
vCash: 500
iTrader: 3 / 100%
Join Date: Nov 2007
Location: Las Vegas
   
01-12-2018, 10:17 AM

Quote:
Originally Posted by cleary View Post
Since Mike seems to be the only one who can access the data, he's able to cherry pick situations that work in his favor...
Well, at least you're consistent on bashing him.

I'd put you at an 815 Fargo for bashing
Jason
  
Reply With Quote
Old
  (#18)
mikepage
AzB Gold Member
mikepage has a reputation beyond reputemikepage has a reputation beyond reputemikepage has a reputation beyond reputemikepage has a reputation beyond reputemikepage has a reputation beyond reputemikepage has a reputation beyond reputemikepage has a reputation beyond reputemikepage has a reputation beyond reputemikepage has a reputation beyond reputemikepage has a reputation beyond reputemikepage has a reputation beyond repute
 
mikepage's Avatar
 
Status: Offline
Posts: 3,654
vCash: 500
iTrader: 0 / 0%
Join Date: Dec 2004
Location: Fargo, ND
   
01-12-2018, 10:18 AM

Quote:
Originally Posted by cleary View Post
Since Mike seems to be the only one who can access the data, he's able to cherry pick situations that work in his favor...
That takes too much time. It's faster to just make up data


********
mike page
http://www.fargorate.com
player ratings for the world
********
  
Reply With Quote
Old
  (#19)
alphadog
AzB Silver Member
alphadog has a reputation beyond reputealphadog has a reputation beyond reputealphadog has a reputation beyond reputealphadog has a reputation beyond reputealphadog has a reputation beyond reputealphadog has a reputation beyond reputealphadog has a reputation beyond reputealphadog has a reputation beyond reputealphadog has a reputation beyond reputealphadog has a reputation beyond reputealphadog has a reputation beyond repute
 
alphadog's Avatar
 
Status: Online
Posts: 2,030
vCash: 500
iTrader: 2 / 100%
Join Date: Nov 2007
   
01-12-2018, 10:24 AM

Quote:
Originally Posted by jasonlaus View Post
Well, at least you're consistent on bashing him.

I'd put you at an 815 Fargo for bashing
Jason
Damn Jason just when I thought... you come up with this. Priceless. I think 815 is probanly about right.
I will say this,about cleary,he does seem to a sceptic about many things,fargorate doesnt garner all his cynicism.


A bull without horns is still dangerous.
  
Reply With Quote
Old
  (#20)
mikepage
AzB Gold Member
mikepage has a reputation beyond reputemikepage has a reputation beyond reputemikepage has a reputation beyond reputemikepage has a reputation beyond reputemikepage has a reputation beyond reputemikepage has a reputation beyond reputemikepage has a reputation beyond reputemikepage has a reputation beyond reputemikepage has a reputation beyond reputemikepage has a reputation beyond reputemikepage has a reputation beyond repute
 
mikepage's Avatar
 
Status: Offline
Posts: 3,654
vCash: 500
iTrader: 0 / 0%
Join Date: Dec 2004
Location: Fargo, ND
   
01-12-2018, 10:43 AM

Quote:
Originally Posted by cleary View Post
Another one of Fargo's major flaws. The skill gap between a 450 and a 550 is rather large. The skill gap between a 550 and a 650 can be massive but the skill difference between a 650 and 750 is pretty small. Mostly mental and the 750 has a higher gear but the 650 can run a set out on a bar table. It's not linear. And the table size means A LOT. But whatever... I'm sure I'm wrong.


I would say your claim doesn't rise to the level of being wrong.
Without defining "skill gap" it has no meaning and cannot be right or wrong.

Most 200s can become 300s by playing an hour a day for a few months
Most 300s can become 400s by playing an hour a day for a year.
Few 600s will become 700s even by playing an hour a day for a lifetime
No 700s will become 800s by playing an hour a day for a lifetime

Clearly these rating gaps, numerically the same, mean vastly different things when it comes to acquiring skills.

The RATING GAPS are well defined. In each case the higher rating wins 2 to 1 over the lower player. Skill gap is a vague concept. You could equate it to a rating gap if you want. Or you could define it in a a number of different ways. Or you could not define it and be #notevenwrong


********
mike page
http://www.fargorate.com
player ratings for the world
********
  
Reply With Quote
Old
  (#21)
skip100
AzB Silver Member
skip100 has a reputation beyond reputeskip100 has a reputation beyond reputeskip100 has a reputation beyond reputeskip100 has a reputation beyond reputeskip100 has a reputation beyond reputeskip100 has a reputation beyond reputeskip100 has a reputation beyond reputeskip100 has a reputation beyond reputeskip100 has a reputation beyond reputeskip100 has a reputation beyond reputeskip100 has a reputation beyond repute
 
skip100's Avatar
 
Status: Offline
Posts: 1,343
vCash: 500
iTrader: 0 / 0%
Join Date: Dec 2013
Location: Golden State
   
01-12-2018, 11:36 AM

Any comments on the winner breaks vs. alternate break issue?
  
Reply With Quote
Old
  (#22)
Bob Jewett
Northern California

Bob Jewett has a reputation beyond reputeBob Jewett has a reputation beyond reputeBob Jewett has a reputation beyond reputeBob Jewett has a reputation beyond reputeBob Jewett has a reputation beyond reputeBob Jewett has a reputation beyond reputeBob Jewett has a reputation beyond reputeBob Jewett has a reputation beyond reputeBob Jewett has a reputation beyond reputeBob Jewett has a reputation beyond reputeBob Jewett has a reputation beyond repute
 
Status: Offline
Posts: 13,150
vCash: 1700
iTrader: 10 / 100%
Join Date: Apr 2004
Location: Berkeley, CA
   
01-12-2018, 02:18 PM

Quote:
Originally Posted by skip100 View Post
Any comments on the winner breaks vs. alternate break issue?
Let's assume that you are considering only applying probability and statistics and not some unknown, unproven and unquantified psychological factor....

The winner/alternate break choice does not change the probabilities for who will win the match. This is a rather amazing result, but that's the way it works out. In fact you have the same chances in loser breaks matches and for having the winner of the lag/flip take N breaks first followed by N-1 for his opponent in a match to N games.

However, the format does change the match score expectations some. If both players are expected to break-and-run a lot (like 80%, which is not the case, yet), then alternate will give lots of hill-hill matches while winner breaks will give more lop-sided results.

For a specific example suppose player A has a "wins-from-break" percentage of 80% and B has a WFB of 60%. In alternate breaks, the most likely match score for a race to 9 is 9-5 at 14% while if they play winner breaks the most likely score is 9-2 at 11%. Amazingly, the chance that A will win either format just given the WFB percentages is 82.2%.

The largest WFB percentage that resident statistics guru AtLarge has observed was in a TAR 10-ball match between SVB and Mike Dechaine, with WFBs of 76% and 55% respectively. On average, the top players are around 56% WFB.

So to more directly answer the question I think you were asking, with the present level of play the choice of format makes a small difference in the statistics of match scores but probably not enough to worry about. Technically it means that in the Fargo calculations the game score with alternate-break format should be counted slightly more heavily than the game score in winner-break format games, but this is true for only the top-ranked players.


Bob Jewett
SF Billiard Academy
  
Reply With Quote
Old
  (#23)
tucson9ball
AzB Silver Member
tucson9ball has a reputation beyond reputetucson9ball has a reputation beyond reputetucson9ball has a reputation beyond reputetucson9ball has a reputation beyond reputetucson9ball has a reputation beyond reputetucson9ball has a reputation beyond reputetucson9ball has a reputation beyond reputetucson9ball has a reputation beyond reputetucson9ball has a reputation beyond reputetucson9ball has a reputation beyond reputetucson9ball has a reputation beyond repute
 
tucson9ball's Avatar
 
Status: Offline
Posts: 1,962
vCash: 175
iTrader: 2 / 100%
Join Date: May 2008
   
01-12-2018, 02:57 PM

The only discrepancy I have personally found is switching from 7ft to 9ft tables/games.
If you take somebody who plays 8 ball on a bar box all the time and let's say their rating is around 550. Now, you move that player to a 9ft table and make them play 9 ball. I think you will find that they will play closer to 520 speed in most cases.
The bar table players tend to have lil imperfections in their stroke that get magnified on the 9ft table. Also, when playing 8 ball, they can get out of line and just decide to shoot a different ball of the same group. This option does not work for 9 ball....get out of line and your inning is more than likely over.
Players with higher Fargo, let's say 650+, usually can play the same on either the 7ft or 9ft. A better stroke equals better position play.
Anyhoo, just my .02



Chris
  
Reply With Quote
Old
  (#24)
skip100
AzB Silver Member
skip100 has a reputation beyond reputeskip100 has a reputation beyond reputeskip100 has a reputation beyond reputeskip100 has a reputation beyond reputeskip100 has a reputation beyond reputeskip100 has a reputation beyond reputeskip100 has a reputation beyond reputeskip100 has a reputation beyond reputeskip100 has a reputation beyond reputeskip100 has a reputation beyond reputeskip100 has a reputation beyond repute
 
skip100's Avatar
 
Status: Offline
Posts: 1,343
vCash: 500
iTrader: 0 / 0%
Join Date: Dec 2013
Location: Golden State
   
01-12-2018, 03:32 PM

Quote:
Originally Posted by Bob Jewett View Post
So to more directly answer the question I think you were asking, with the present level of play the choice of format makes a small difference in the statistics of match scores but probably not enough to worry about. Technically it means that in the Fargo calculations the game score with alternate-break format should be counted slightly more heavily than the game score in winner-break format games, but this is true for only the top-ranked players.
I agree with this. In practice it will only affect top ranked players but those are also the ratings people care about most.
  
Reply With Quote
Old
  (#25)
Bob Jewett
Northern California

Bob Jewett has a reputation beyond reputeBob Jewett has a reputation beyond reputeBob Jewett has a reputation beyond reputeBob Jewett has a reputation beyond reputeBob Jewett has a reputation beyond reputeBob Jewett has a reputation beyond reputeBob Jewett has a reputation beyond reputeBob Jewett has a reputation beyond reputeBob Jewett has a reputation beyond reputeBob Jewett has a reputation beyond reputeBob Jewett has a reputation beyond repute
 
Status: Offline
Posts: 13,150
vCash: 1700
iTrader: 10 / 100%
Join Date: Apr 2004
Location: Berkeley, CA
   
01-12-2018, 04:12 PM

Quote:
Originally Posted by skip100 View Post
I agree with this. In practice it will only affect top ranked players but those are also the ratings people care about most.
But the ranking order should not change assuming the top players are in roughly the same mix of formats. I think the absolute errors due to format are going to be pretty small until the players get a lot better.

Where the format can make a difference is in the over/under for total number of games played. In the example of SVB/Dechaine above, in a race to nine with alternate breaks, the over/under split is 14.5 while the winner breaks split is 13.5 games. That's not much and it is the extreme observed WFB percentage.


Bob Jewett
SF Billiard Academy
  
Reply With Quote
Old
  (#26)
poolscholar
AzB Silver Member
poolscholar has a reputation beyond reputepoolscholar has a reputation beyond reputepoolscholar has a reputation beyond reputepoolscholar has a reputation beyond reputepoolscholar has a reputation beyond reputepoolscholar has a reputation beyond reputepoolscholar has a reputation beyond reputepoolscholar has a reputation beyond reputepoolscholar has a reputation beyond reputepoolscholar has a reputation beyond reputepoolscholar has a reputation beyond repute
 
Status: Offline
Posts: 200
vCash: 500
iTrader: 1 / 100%
Join Date: Sep 2010
Location: Minneapolis, MN
   
01-12-2018, 04:40 PM

Quote:
Originally Posted by Bob Jewett View Post
Let's assume that you are considering only applying probability and statistics and not some unknown, unproven and unquantified psychological factor....

The winner/alternate break choice does not change the probabilities for who will win the match. This is a rather amazing result, but that's the way it works out. In fact you have the same chances in loser breaks matches and for having the winner of the lag/flip take N breaks first followed by N-1 for his opponent in a match to N games.

However, the format does change the match score expectations some. If both players are expected to break-and-run a lot (like 80%, which is not the case, yet), then alternate will give lots of hill-hill matches while winner breaks will give more lop-sided results.

For a specific example suppose player A has a "wins-from-break" percentage of 80% and B has a WFB of 60%. In alternate breaks, the most likely match score for a race to 9 is 9-5 at 14% while if they play winner breaks the most likely score is 9-2 at 11%. Amazingly, the chance that A will win either format just given the WFB percentages is 82.2%.

The largest WFB percentage that resident statistics guru AtLarge has observed was in a TAR 10-ball match between SVB and Mike Dechaine, with WFBs of 76% and 55% respectively. On average, the top players are around 56% WFB.

So to more directly answer the question I think you were asking, with the present level of play the choice of format makes a small difference in the statistics of match scores but probably not enough to worry about. Technically it means that in the Fargo calculations the game score with alternate-break format should be counted slightly more heavily than the game score in winner-break format games, but this is true for only the top-ranked players.
Also if there are games spotted on the wire, math changes and then alternate favors the person getting the spot. I wrote a computer program to verify this a while back. I'm sure Mike could gather this data and look at the difference...
  
Reply With Quote
Old
  (#27)
Black-Balled
Look!Everything'sSpinning
Black-Balled has a reputation beyond reputeBlack-Balled has a reputation beyond reputeBlack-Balled has a reputation beyond reputeBlack-Balled has a reputation beyond reputeBlack-Balled has a reputation beyond reputeBlack-Balled has a reputation beyond reputeBlack-Balled has a reputation beyond reputeBlack-Balled has a reputation beyond reputeBlack-Balled has a reputation beyond reputeBlack-Balled has a reputation beyond reputeBlack-Balled has a reputation beyond repute
 
Black-Balled's Avatar
 
Status: Offline
Posts: 33,361
vCash: 1200
iTrader: 13 / 100%
Blog Entries: 1
Join Date: Mar 2004
Location: This Toilet Earth
  Send a message via AIM to Black-Balled  
01-12-2018, 04:42 PM

Quote:
Originally Posted by cleary View Post
Another one of Fargo's major flaws. The skill gap between a 450 and a 550 is rather large. The skill gap between a 550 and a 650 can be massive but the skill difference between a 650 and 750 is pretty small. Mostly mental and the 750 has a higher gear but the 650 can run a set out on a bar table. It's not linear. And the table size means A LOT. But whatever... I'm sure I'm wrong.
But i think a flat measure having disparate impact on the ends of the spectrum is also true in most sports. At least is singular effort ones.

Cycling? A hack vs a low level pro, over a 20mi race? Might be 20min difference. Low pro vs top pro? Pprob only a handful of minutes.


Two thirds from the top,
Another third for roads and cops.
Don't kid yourself that it's ever gonna stop


Quote:
Originally Posted by westcoast View Post
...he loves pool, but he lost his love for pool halfway through the one pocket...
  
Reply With Quote
Old
  (#28)
alphadog
AzB Silver Member
alphadog has a reputation beyond reputealphadog has a reputation beyond reputealphadog has a reputation beyond reputealphadog has a reputation beyond reputealphadog has a reputation beyond reputealphadog has a reputation beyond reputealphadog has a reputation beyond reputealphadog has a reputation beyond reputealphadog has a reputation beyond reputealphadog has a reputation beyond reputealphadog has a reputation beyond repute
 
alphadog's Avatar
 
Status: Online
Posts: 2,030
vCash: 500
iTrader: 2 / 100%
Join Date: Nov 2007
   
01-12-2018, 04:53 PM

Quote:
Originally Posted by Black-Balled View Post
But i think a flat measure having disparate impact on the ends of the spectrum is also true in most sports. At least is singular effort ones.

Cycling? A hack vs a low level pro, over a 20mi race? Might be 20min difference. Low pro vs top pro? Pprob only a handful of minutes.

Somebody hack your account and bring up cycling as a guise?


A bull without horns is still dangerous.
  
Reply With Quote
Old
  (#29)
Black-Balled
Look!Everything'sSpinning
Black-Balled has a reputation beyond reputeBlack-Balled has a reputation beyond reputeBlack-Balled has a reputation beyond reputeBlack-Balled has a reputation beyond reputeBlack-Balled has a reputation beyond reputeBlack-Balled has a reputation beyond reputeBlack-Balled has a reputation beyond reputeBlack-Balled has a reputation beyond reputeBlack-Balled has a reputation beyond reputeBlack-Balled has a reputation beyond reputeBlack-Balled has a reputation beyond repute
 
Black-Balled's Avatar
 
Status: Offline
Posts: 33,361
vCash: 1200
iTrader: 13 / 100%
Blog Entries: 1
Join Date: Mar 2004
Location: This Toilet Earth
  Send a message via AIM to Black-Balled  
01-12-2018, 04:55 PM

Quote:
Originally Posted by alphadog View Post
Somebody hack your account and bring up cycling as a guise?
No...i post a lot about bikes...?


Two thirds from the top,
Another third for roads and cops.
Don't kid yourself that it's ever gonna stop


Quote:
Originally Posted by westcoast View Post
...he loves pool, but he lost his love for pool halfway through the one pocket...
  
Reply With Quote
Old
  (#30)
cleary
Honestly, I'm a liar.
cleary has a reputation beyond reputecleary has a reputation beyond reputecleary has a reputation beyond reputecleary has a reputation beyond reputecleary has a reputation beyond reputecleary has a reputation beyond reputecleary has a reputation beyond reputecleary has a reputation beyond reputecleary has a reputation beyond reputecleary has a reputation beyond reputecleary has a reputation beyond repute
 
cleary's Avatar
 
Status: Offline
Posts: 9,251
vCash: 8500
iTrader: 18 / 100%
Join Date: Nov 2006
Location: nyc/nj
  Send a message via AIM to cleary  
01-12-2018, 07:35 PM

Quote:
Originally Posted by mikepage View Post
That takes too much time. It's faster to just make up data
That definitely true but I would expect you’re better than that.
  
Reply With Quote
Reply
Page 2 of 5 12 34 Last »

Thread Tools
Rate This Thread
Rate This Thread:

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump



Powered by vBulletin® Version 3.8.9
Copyright ©2000 - 2018, vBulletin Solutions, Inc.
vBulletin Security provided by vBSecurity (Lite) - vBulletin Mods & Addons Copyright © 2018 DragonByte Technologies Ltd.