Is there FargoRate rating inflation?

[mikepage]

Is there FargoRate rating inflation?​

Six years ago, spring of 2019, 24 players had an established Fargo Rating over 800. That number is now 68. Further, the average rating for that 2019 group has risen 15+ points over those years. Is this rating inflation, as in the same level of play that earned 800 then earns 815 now? Or have those players and their new elite friends upped the ante?

While many point to this expanding 800-club and suggest we’re seeing significant inflation, we disagree. Rating inflation appears to be small, around two points total in the last six years. The rest is, yes, the best players are actually playing better.

Evidence for this view is shown in the graphic. The 24 rated-over-800 players from 2019 average 826.1 now, up from 810.6. We could not support the view that this increase is inflation without also making the less credible claim that the thousands of players at lower ratings here have on average gotten 13 or so points weaker. Essentially The 24 elite players who were back in 2019 beating the players in the 675 rating group (350 of them) by a score of 21 to 9 are now beating those same opponents by a bigger margin, 21 to 8. They’ve widened the gap.

Don’t Elo ratings, like in chess, have known inflation/deflation issues
They do. FargoRate is fundamentally different from Elo, though, and the causes of drift for Elo approaches don’t apply to FargoRate.

What does FargoRate do to keep the skill level of a rating roughly the same over time?
Fargo Ratings can all be shifted up or down any arbitrary amount because predictions refer only to rating differences and not to absolute ratings. We now use this flexibility to require the average rating for the top 4000 players with established rating and activity in the last two years remains fixed from today to tomorrow. Though the comparison group can change over time, the averages are for a common group of players.

Why have the world’s top players improved?
We can only speculate: more opportunities for elite players to compete with one another and greater awareness for top players of where they fit in with top players remote to them.

 

[sjm]

I'm not convinced here. Yes, Fedor and Josh are better than they used to be, maybe better at 9ball than anything we've ever seen over the green felt, but, based on the eye test, I do not agree that the Top 10 as a group are playing significantly better than they used to, and without naming names, I can think of numerous high-Fargo players whose Fargo rates have gone up while the quality of their play has gone down.

Yours, Stu (Fargo's greatest advocate)

 

[Fatboy]

I'm not convinced here. Yes, Fedor and Josh are better than they used to be, maybe better at 9ball than anything we've ever seen over the green felt, but, based on the eye test, I do not agree that the Top 10 as a group are playing significantly better than they used to, and without naming names, I can think of numerous high-Fargo players whose Fargo rates have gone up while the quality of their play has gone down.

Yours, Stu (Fargo's greatest advocate)

It’s possible I’m just dead wrong here. Then again maybe I’m not, so I’ll mention my thought

Gotta frame it accurately, I have 10% of your experience sweating top players. Ofc I watch lots of matches and we see the same thing 90% of the time.

However where we kinda are thinking different is I think the top 6-7-8-10 guys. Not more than 10 are playing better than 5 years ago. SVB included. Those guys are looking better and better to me. Even when I watch old matches now and see prime “who ever” they don’t look as sharp pocketing balls as the guys now.

The 750’s meh, they are still 750’s. Even below that things feel pretty much the same.

It’s interesting as most models do have inflation. They have used a few backgammon models which aren’t the same as chess and those as well have a bit of inflation. I don’t have the maths prowess to explain all that. And yes the top 10 BG players in the world are without a doubt playing much better than 5 years ago-but we have computers to analyze that so it’s not subjective as pool is.

Stu, I hope you have a great summer! If I get to the city I’ll let you know! Dinner is on me! So pick a good place

Best
Fatboy<———offsetting inflation with my -200 Fargo drop

 

[iusedtoberich]

Does not pass the sniff test. Idk why you are the only person on the planet, Mike, who does not believe the creep.

 

[skogstokig]

i guess one could run it against the TPA stats of those top players throughout the years. MR also have match stats but they are less credible.

also i think both things could be true here, the best players have become better, but it doesn't necessarily match the jump in fargo numbers. i remember when filler was 825. he is better now, but is he 30 points better?

 

[MitchAlsup]

I do not know the "math" behind Fargo rates::

But, let us postulate 16 players with Fargo Rate of exactly 800 each with exactly 1000 Fargo rate qualified matches each,
and these are the only players in "joe random tournament".

After the tournament, what would be the resulting Fargo Rates ??

 

[mikepage]

i guess one could run it against the TPA stats of those top players throughout the years.

That's a good idea. Challenges are getting enough data to be meaningful and hoping equipment changes isn't biting you. But again, good idea.
[...]

 

[mikepage]

Does not pass the sniff test. Idk why you are the only person on the planet, Mike, who does not believe the creep.

lol

I'm that one guy at the Thanksgiving dinner table with a different answer for "what is the highest mountain in South America?" Reasonable to think I'm probably wrong..... until you notice I'm the only person at the table with a phone attached to the internet...

 

[mikepage]

Does not pass the sniff test. Idk why you are the only person on the planet, Mike, who does not believe the creep.

One problem is you're doing too much of your sniffing in the Philadelphia area, where some tournament data started coming in years ago that was largely uncoupled from the rest of the world. The area was quite out of whack. And as coupling data came in the whole tide raised up to match with the rest of the world.

 

[markjames]

Even when I watch old matches now and see prime “who ever” they don’t look as sharp pocketing balls as the guys now.

seems to me that the level of play continues to improve

imagine how different efren would have been if he was always playing against fedor

 

[iusedtoberich]

One problem is you're doing too much of your sniffing in the Philadelphia area, where some tournament data started coming in years ago that was largely uncoupled from the rest of the world. The area was quite out of whack. And as coupling data came in the whole tide raised up to match with the rest of the world.

When I have some time, I’ll put a graph of all my players here. From top pros, old and almost dead pros, local “decent” players. They almost all went up a mile. Creep is the only thing that makes sense.

 

[Fatboy]

seems to me that the level of play continues to improve

imagine how different efren would have been if he was always playing against fedor

1P would be his only shot to be competitive

 

[Island Drive]

I'm not convinced here. Yes, Fedor and Josh are better than they used to be, maybe better at 9ball than anything we've ever seen over the green felt, but, based on the eye test, I do not agree that the Top 10 as a group are playing significantly better than they used to, and without naming names, I can think of numerous high-Fargo players whose Fargo rates have gone up while the quality of their play has gone down.

Yours, Stu (Fargo's greatest advocate)

Especially with 4'' pockets

 

[SBC]

Samuel Clemens, also known as Mark Twain, is famously associated with the phrase "lies, damned lies, and statistics." This quote highlights the potential for misrepresentation and manipulation within statistical data. Clemens himself noted his own difficulty with interpreting figures, and the idea that statistics can be persuasive even when used incorrectly. He recognized the power of statistics to be used in misleading ways.

I work in cancer care....the pharmaceutical industry runs studies on drugs and can produce "statistically significant" results to fit desired outcomes.

I'm a Fargo 610....should be 680....but I don't even play much.
If I played only 610's for the cash, I won't even need a day job.

I saw Shaw vs Eklent Kaci the other day playing and they are 1 Fargo point apart. Even post his accident I'll bet high on Kaci because he is on the improve and for anybody paying attention, Shaw isn't winning like he was. Owning a pool hall and having a family takes a toll on your game.

 

[Island Drive]

That's a good idea. Challenges are getting enough data to be meaningful and hoping equipment changes isn't biting you. But again, good idea.
[...]

I'm unable to feel this way about challenge$.... for cash.

 

[SBC]

i guess one could run it against the TPA stats of those top players throughout the years. MR also have match stats but they are less credible.

also i think both things could be true here, the best players have become better, but it doesn't necessarily match the jump in fargo numbers. i remember when filler was 825. he is better now, but is he 30 points better?

Yes at least 30.

If he played more events you'd see him winning more than he does


Nobody even wants to play him in long races because he's the favorite against all of them.

 

[highkarate]

Could you run a simulation to prove/disprove the inflation theory? Something like, make a field of players that all have a known rating. Say 10 players that all have a true rating of 700 (or any ratings will work if they have infinite robustness but I’ll say 700 for ease). Assign some amount of variance and appropriate standard deviation, then simulate a large number of matches with them all playing each other and Fargo doing its math on all their games. If, after thousands (or millions?) of games, the total number of their combined ratings is significantly over 7,000, there is inflation.

 

[skip100]

I'm not convinced here. Yes, Fedor and Josh are better than they used to be, maybe better at 9ball than anything we've ever seen over the green felt, but, based on the eye test, I do not agree that the Top 10 as a group are playing significantly better than they used to, and without naming names, I can think of numerous high-Fargo players whose Fargo rates have gone up while the quality of their play has gone down.

Yours, Stu (Fargo's greatest advocate)

This is the strongest evidence for inflation.

Overall this is a weird post. The contention that FargoRate (i.e. the Bradley-Terry pairwise comparison model https://en.wikipedia.org/wiki/Bradley–Terry_model) is "fundamentally different than Elo" is simply not true. The optimization process is different but the underlying probability model is the same.

 

[iusedtoberich]

Even Nick Varner who is half dead has gone up.

 

[alphadog]

Is it possible that having the top players beat each other on a regular basis has caused the top players to go up?

Seems logical to me that a 820 beating other 800+ players has to raise the 820's fr.