Interesting FB post about FargoRate

[sjm]

I learned everything I needed to learn about handicapping and sandbagging as a golfer for the past 50 years.

I admit, when FargoRate first came out I was pretty leery because of my experience with golfers and their ability to manipulate their handicaps.

However, over the past couple of years, I’ve come to appreciate that it actually does a pretty good job for the most part. Matches handicapped by FargoRate tend to be pretty fair, even with players of significantly different skill levels, and I‘ve found that most ratings pretty accurately reflect the relative ability of the players.

Having said that, I’m starting to see cases of sandbagging. Not necessarily so that one player has an advantage over another in a Fargo rate handicapped match/league, but because there are a lot of tournaments that limit entries based on FargoRate.

Some of these tournaments are pretty significant. An “under 600” tournament can easily have 64 players with payouts in the thousands of dollars and Calcuttas to match. When there’s several thousand dollars available for the top finishers, there’s a lot of incentive to keep your FargoRate under that 600 (or whatever) threshold if you can so you’re not excluded from the tournaments. That’s not too hard to do by dogging a few league matches. It can even be done pretty easily without affecting your league team standings with a little bit of care.

Should/could a league operator notice and manage these things? Probably. Maybe. Again, it can be pretty subtle.

Overall, I think FargoRate is the best handicapping system I’ve seen in pool. It pretty much does what it‘s supposed to do, and the ratings travel pretty well. Certainly much better than any kind of regional rating system that I’ve ever seen over the years. Sadly, any kind of system can be manipulated by an individual intent on doing so. About all we can do is keep an eye out for those individuals, and address them as aggressively as possible when they‘re found.

Take a deep bow, Dave, for a superb post. Fargo isn't perfect, and I don't really have a strong opinion about it below pro level. As you suggest, I'm sure there's room for manipulation.

At pro level, Fargo is the best system we've had in a long time. This is, in part, due to the absence in recent years of a credible set of rankings for pool. Matchroom is in the process of changing this, however, as the Matchroom nine-ball rankings are the best set of rankings we've had in our sport in many years, and it is those rankings that determine invitations and seeding in their events. Matchroom is gradually transforming pro pool into a meritocracy, and their new tour offers pro players the greatest opportunity they have had in a very long time.

To her credit, Emily Frazer has been quite candid about the fact that, because these rankings are new, it will take a little time before the Matchroom rankings will fully reflect where things really stand in the world of pool, but Matchroom has us on a winning course.

I've said it before, and I'll say it again. Fargo ratings have and continue to add something of great value in pro pool.

 

[sjm]

The t.p.a is a better sample of actual ability. It doesn't rely on your opponent's ability being accurate and is closer to golf handicaps. The problem is accurately, easily and consistently recording those results.

Someone who averages in the .900s, would be top pro level, someone averaging in the mid to high .800s would be low pro short stop, so on and so on.

It would be similar to a bell curve, bunched up in the middle with few at the top or at the bottom.

Jaden

Well said, although TPA also has its limitation because it doesn't measure one's proficiency at the break, defense, kicking, or general tactical conceptualization (pushouts, two-way shots, etc.). As you say, though, anyone carrying a TPA of .900 is probably an elite player.

 

[Black-Balled]

There is more mis information about fr then covid

Bet, mother fuc....!

 

[alphadog]

Bet, mother fuc....!

I will stack them deep and my cousin can rule on it

 

[telinoz]

Most of this makes zero sense.

All good, some people can't read too many words and keep track of context.

 

[u12armresl]

Take a deep bow, Dave, for a superb post. Fargo isn't perfect, and I don't really have a strong opinion about it below pro level. As you suggest, I'm sure there's room for manipulation.

At pro level, Fargo is the best system we've had in a long time. This is, in part, due to the absence in recent years of a credible set of rankings for pool. Matchroom is in the process of changing this, however, as the Matchroom nine-ball rankings are the best set of rankings we've had in our sport in many years, and it is those rankings that determine invitations and seeding in their events. Matchroom is gradually transforming pro pool into a meritocracy, and their new tour offers pro players the greatest opportunity they have had in a very long time.

To her credit, Emily Frazer has been quite candid about the fact that, because these rankings are new, it will take a little time before the Matchroom rankings will fully reflect where things really stand in the world of pool, but Matchroom has us on a winning course.

I've said it before, and I'll say it again. Fargo ratings have and continue to add something of great value in pro pool.

The unfortunate thing is that Matchroom only covers the top 1% of players.
What about the majority of people who just want to be labeled fairly and not be penalized because they won 5-4 instead of 5-1

 

[David in FL]

The unfortunate thing is that Matchroom only covers the top 1% of players.
What about the majority of people who just want to be labeled fairly and not be penalized because they won 5-4 instead of 5-1

I’m not sure that I understand.

What do you mean that someone is “penalized” by winning a match 5-4 instead of 5-1?

 

[hang-the-9]

The t.p.a is a better sample of actual ability. It doesn't rely on your opponent's ability being accurate and is closer to golf handicaps. The problem is accurately, easily and consistently recording those results.

Someone who averages in the .900s, would be top pro level, someone averaging in the mid to high .800s would be low pro short stop, so on and so on.

It would be similar to a bell curve, bunched up in the middle with few at the top or at the bottom.

Jaden

Issue is that it takes someone sitting there marking down all the events in the rack to do the TPA scores, which only happens in leagues, and the leagues have their own handicapping system that don't run on Fargo ratings. I know when I'm in a random tournament I am not going to keep track of mine or my opponents every action and shot. Won/Loss stats and scores are simple.

 

[hang-the-9]

The unfortunate thing is that Matchroom only covers the top 1% of players.
What about the majority of people who just want to be labeled fairly and not be penalized because they won 5-4 instead of 5-1

Well a 5-1 score and a 5-4 score shows different abilities. A C player may be able to win a single game against an A player but it is very unlikely they would get to 4 vs one. But a B+ or another A player can certainly get to hill hill with an A player, so the 5-4 score recorded in Fargo is perfectly fair, as is the 5-1 score. It's statistics over time, a single match here and there does not matter, you need many dozens of matches and games for the ratings to even out the random bumps in very good or very bad results to get the average of the skill rating. Even if both players were A level and the score was 5-1, it is very unlikely that both of them would lose say 20 matches by such a blow out score, so it all just goes into the average.

 

[Jaden]

Well a 5-1 score and a 5-4 score shows different abilities. A C player may be able to win a single game against an A player but it is very unlikely they would get to 4 vs one. But a B+ or another A player can certainly get to hill hill with an A player, so the 5-4 score recorded in Fargo is perfectly fair, as is the 5-1 score. It's statistics over time, a single match here and there does not matter, you need many dozens of matches and games for the ratings to even out the random bumps in very good or very bad results to get the average of the skill rating. Even if both players were A level and the score was 5-1, it is very unlikely that both of them would lose say 20 matches by such a blow out score, so it all just goes into the average.

Exactly, but that's what I was saying about chip tournaments where you're only playing 1 game against every opponent. Entering those scores into fargo is ridiculous, it's not going to be good for anyone in terms of accuracy. You should beat them 5-1, yet it may record them as beating you 1-0 if they happen to get that first game, and you beating them 1-0 is less than what it should be just to maintain your rating.

Jaden

 

[mikepage]

Exactly, but that's what I was saying about chip tournaments where you're only playing 1 game against every opponent. Entering those scores into fargo is ridiculous, it's not going to be good for anyone in terms of accuracy. You should beat them 5-1, yet it may record them as beating you 1-0 if they happen to get that first game, and you beating them 1-0 is less than what it should be just to maintain your rating.

Jaden

Single games are all that FargoRate ever knows about. A "match" that ends 5-to-3 is, to FargoRate, just 8 single games against the same opponent. It contains a comparable amount of information as does 1 game each against 8 different opponents. Chip tournaments are fine.

Also it's fine to play all your games against just weaker opponents or all against just stronger opponents or a mixture. It doesn't matter.

 

[easy-e]

If you're supposed to win 4-0 against a player, how can a 1-0 score NOT lower your rating? You're only 1 game ahead where as you should be 4 games ahead.

Jaden

I see it the other way. You won infinite more games than they did, how can your score NOT go up? From my experience, my score goes up when I win these 1-0 matches. This happens during league nights and chip tournaments. I don't play the singles chip tournaments anymore because I have to win something like 35-3 playing bar table 9-ball giving up the break. It's rough.

 

[Jaden]

Single games are all that FargoRate ever knows about. A "match" that ends 5-to-3 is, to FargoRate, just 8 single games against the same opponent. It contains a comparable amount of information as does 1 game each against 8 different opponents. Chip tournaments are fine.

Also it's fine to play all your games against just weaker opponents or all against just stronger opponents or a mixture. It doesn't matter.

Not when the matches are handicapped. If you are ONLY playing handicapped matches against lesser opponents, there WILL be discrepancy.

That stems from the times when they get to two games before the better player gets the games they should. The better player's opportunity to flatten out the discrepancy ends prematurely. How much of a statistical difference it makes in the long run is debatable. It would greatly depend on how often they are getting the games and ending the match prematurely.

If they are even races, then yes I would agree that it would be completely accurate regardless as to what level of player you're playing against, but because most of those matches are handicapped allowing for early endings to matches, it will never be totally accurate.

Jaden

 

[sixpack]

Not when the matches are handicapped. If you are ONLY playing handicapped matches against lesser opponents, there WILL be discrepancy.

That stems from the times when they get to two games before the better player gets the games they should. The better player's opportunity to flatten out the discrepancy ends prematurely. How much of a statistical difference it makes in the long run is debatable. It would greatly depend on how often they are getting the games and ending the match prematurely.

If they are even races, then yes I would agree that it would be completely accurate regardless as to what level of player you're playing against, but because most of those matches are handicapped allowing for early endings to matches, it will never be totally accurate.

Jaden

I think what you are stuck on is the matches vs games. FR doesn’t care about matches. Winning 5 games against 5 different opponents with an average rating of 500 counts the same as beating a 500 5-0.

Edit: if you are playing a race where you have to win 9 and your opponent has to win 3 - and you lose 7-3. FR sees this as you having 7 wins and 3 losses against a 350? And a 350 as having 3 wins and 7 losses against a 700?.

So even though you lost the match the FR really isn’t that far out of line.

 

[mikepage]

Not when the matches are handicapped. If you are ONLY playing handicapped matches against lesser opponents, there WILL be discrepancy.

That stems from the times when they get to two games before the better player gets the games they should. The better player's opportunity to flatten out the discrepancy ends prematurely. How much of a statistical difference it makes in the long run is debatable. It would greatly depend on how often they are getting the games and ending the match prematurely.
[...]

Those matches will end "Prematurely" in a way that balances the times when they end "Overmaturely." There is no statistical effect here.

 

[Jaden]

Those matches will end "Prematurely" in a way that balances the times when they end "Overmaturely." There is no statistical effect here.

They can't end overly maturely though. If a 650 is playing a 450 and the race is 6-2, The best the 650 can do is win 6-0. Where he may have been able to win 11-0, or 11-1. The match can't end overly maturely. This prevents a true metric from being created when the premature ending can only happen or is more likely to happen in one direction.

At the upper levels where they're more likely to NOT being playing within a handicap, it absolutely will balance out in the end. In a handicap. It can't because it literally can only end prematurely in ONE direction.

Jaden

Mike, you have the best system that has been used to date, I REALLY am not shitting on it at ALL. Any system that would be more accurate is not implementable in a realistic way.

 

[mikepage]

They can't end overly maturely though. If a 650 is playing a 450 and the race is 6-2, The best the 650 can do is win 6-0. Where he may have been able to win 11-0, or 11-1. The match can't end overly maturely. This prevents a true metric from being created when the premature ending can only happen or is more likely to happen in one direction

Jaden

It happens in both directions.

Let's say the expectation for 8 games is 6-to-2

If the players do a straight race to 6, then the scores might look like
6-2
6-0
6-1
6-3
6-0
6-4
6-2
6-3
6-4
6-1

for a total of 60-20

If the same 10 matches were handicapped 6-2, then the scores might look like

5-2
6-0
6-1
3-2
6-0
0-2
3-2
5-2
4-2
6-1

Now the total score is 44-14, still at that 3-to-1 ratio.

In the cases where the weaker player got to 2, the stronger player is, as you say, "cut short" of his expectation. But it's also true that in the cases the stronger player got to 6 in the handicapped match, the weaker player was cut short of his expectation. It balances out.

 

[Jaden]

It happens in both directions.

Let's say the expectation for 8 games is 6-to-2

If the players do a straight race to 6, then the scores might look like
6-2
6-0
6-1
6-3
6-0
6-4
6-2
6-3
6-4
6-1

for a total of 60-20

If the same 10 matches were handicapped 6-2, then the scores might look like

5-2
6-0
6-1
3-2
6-0
0-2
3-2
5-2
4-2
6-1

Now the total score is 44-14, still at that 3-to-1 ratio.

In the cases where the weaker player got to 2, the stronger player is, as you say, "cut short" of his expectation. But it's also true that in the cases the stronger player got to 6 in the handicapped match, the weaker player was cut short of his expectation. It balances out.

Except you're not allowing for the other cut shorts. the 2-1, 2-0, even 2-2 or 2-3. The difference in it being cut short can be 4-6 games, versus only 2 games MAX in the other direction.

Now I conceded it would depend how often those types of cut shorts happen as to how much of a statistical effect it has, but the difference could be substantial.

Jaden

 

[mikepage]

Except you're not allowing for the other cut shorts. the 2-1, 2-0, even 2-2 or 2-3. The difference in it being cut short can be 4-6 games, versus only 2 games MAX in the other direction.

Now I conceded it would depend how often those types of cut shorts happen as to how much of a statistical effect it has, but the difference could be substantial.

Jaden

There is a 2-0 and two 2-3's in my example.

There is no effect at all like the one you're suggesting. It is not a matter of how big a statistical effect; no effect.
I suggest you create a simple spreadsheet and simulate both handicapped and unhandicapped matches to convince yourself.

 

[Jaden]

There is a 2-0 and two 2-3's in my example.

There is no effect at all like the one you're suggesting. It is not a matter of how big a statistical effect; no effect.
I suggest you create a simple spreadsheet and simulate both handicapped and unhandicapped matches to convince yourself.

I would agree that if it's truly random there would be little to no effect. I'm just not sure IIRC that it IS truly random. I would want to see a real world sample of how often it happens to know for sure. There would be mitigating circumstances if say, there are several sandbaggers that AREN'T truly 450's, which we KNOW is the case.

BTW, I commend you for ferreting out those statistically anomolous and obviously contrived entries that this thread is based on.

Jaden

p.s. I'll capitulate the point for this for now as I just don't have the time to invest right now to truly test it.