Johnny Archer and Fargo Ratings

mikepage

mikepage

AzB Silver Member
Silver Member
Johnny is US #2 by Fargo Ratings with a rating of 789. Some posters have wondered whether FargoRate gives too much credit to older games, that maybe this impressive standing is in part a testimony to Johnny's historical success.

We did an experiment to find out.

FargoRate has about 5,000 games for Archer that go back about a decade. Although older games are valued less as a normal part of the optimization, we decided to do a more explicit test.

So we took the 403 games Johnny has played in the last ONE YEAR and assigned them to a hypothetical player from planet Tralfamadore named Superman Archer. In addition we took the 703 games Johnny played in the last TWO YEARS and assigned them to Batman Archer.

Findings:

Batman Archer is rated 789, the same as Johnny Archer and 2nd in the USA.

Superman Archer is rated 773, tied for 5th in the USA.

Johnny Archer is still top US and world-class player.
 

Attachments

  • archers.jpg
    archers.jpg
    29.3 KB · Views: 1,081
watchez

watchez

What time is it?
Silver Member
Can you list Johnny's last 50 matches (W/L and opponent)....also with dates of the matches to show how much he has or has not played in recent history.

Thanks.
 
Johnnyt

Johnnyt

Burn all jump cues
Silver Member
IMO any pool player stats should only be for one year, like most sports are. Johnnyt
 
mikepage

mikepage

AzB Silver Member
Silver Member
watchez said:
Can you list Johnny's last 50 matches (W/L and opponent)....also with dates of the matches to show how much he has or has not played in recent history.

Thanks.
Click to expand...

OK I don't have the dates on these, but they are all from the last two years, from,

US Open 8ball 2015
Us Open 10 Ball 2015
Memphis Open 2015
2015 Bigfoot Challenge
2014 US Open 9-Ball Championship
2014 Turning Stone Classic
2013 Turning Stone
 

Attachments

  • archermatches.jpg
    archermatches.jpg
    84.2 KB · Views: 952
ideologist

ideologist

I don't never exaggerate
Silver Member
Johnnyt said:
IMO any pool player stats should only be for one year, like most sports are. Johnnyt
Click to expand...

I disagree, no other sport offers 70 year old son the chance to dominate.

Looking at cumulative stats that diminish in significance makes more sense.

If I beat Darren Appleton tomorrow in a race to 100, whatever talent I possessed does not go away in 2 years.

My eyes may degrade in that time, and that would be reflected in more losses piling up and lowering my rating.
 
gutshot

gutshot

AzB Silver Member
Silver Member
Professional Golf uses the last two years. The current years points count double the previous year. They started out using the last three years to determine the world rankings.

Johnnyt said:
IMO any pool player stats should only be for one year, like most sports are. Johnnyt
Click to expand...
 
mikepage

mikepage

AzB Silver Member
Silver Member
watchez said:
Can you list Johnny's last 50 matches (W/L and opponent)....also with dates of the matches to show how much he has or has not played in recent history.

Thanks.
Click to expand...

Our subjective judgment places too big an emphasis on a few key games and a few key matches.

If Pin Yi Ko beats Shane 15 to 14, we tend to make a big deal about Ko's performance, whereas if Ko loses 14 to 15 to Shane not so much. The actual performance is not very different.

Sure there are differences in the way players can handle pressure situations, and FargoRate makes no explicit note of this. But our analyses suggest things like this are a smaller effect than any people think.

At last year's Bigfoot challenge, as one example, Archer lost first round 11-8 against SVB. Probably nobody remembers this. But the fact is no other player got past 8 against Shane for the rest of the tournament. If Archer just happened to have a different draw and got to the finals of that event, we would all be talking about him very differently. When by performance it's not very different.
 
S

(((Satori)))

AzB Silver Member
Silver Member
I have a question about your system.


Can one match drastically change a layers rating?


Like for example his win over Lee Van Corteza. It says that he beat Lee Van 11-1... if he would have beat Lee Van 11-10 would that have changed his overall rating much?
 
watchez

watchez

What time is it?
Silver Member
So if I added up correctly, Johnny is 26-24 in his last 50 matches and this goes back to 2013 (which means he hasn't been drawing his ball much lately). Of those 26 wins, I would give quality wins over Karl Boyes, Lee Van, Jason Klatt, John Morra, Ouschan and Jayson Shaw (6 total -- I'm sure someone will find one or two I missed or want to argue about. )


Johnny has 6 quality wins in his last 50 matches over the past 2 years and he is #2 in the USA?? Not trying to discount your hard work with the Fargo ratings but I'll stand by my original statement that it needs some re-evaluating.

Johnny didn't play in any other DCC events? And what about his recent Young Guns vs Old School challenge where Bergman and Sky beat Johnny and Rodney?
 
mikepage

mikepage

AzB Silver Member
Silver Member
(((Satori))) said:
I have a question about your system.


Caonn one match drastically change a layers rating?


Like for example his win over Lee Van Corteza. It says that he beat Lee Van 11-1... if he would have beat Lee Van 11-10 would that have changed his overall rating much?
Click to expand...

We will release a utility soon that estimates the rating change for any player with a particular match score against any other players. In general, the effect of that match on a newer player (with only 200 games in the system, for example) is quite a bit larger than is the effect on an established player like Archer.

In this case, that match with Corteza was from the 2013 US Open. It is in the mix for Batman Archer (789 and US #2) but not for Superman Archer(773 and US #6). It is good to have a sense of this, but the "but for xxx" thinking it is a danger path. Do you also throw out his 9-2 and 9-3 losses against Strickland and Shaw?
 
S

(((Satori)))

AzB Silver Member
Silver Member
mikepage said:
We will release a utility soon that estimates the rating change for any player with a particular match score against any other players. In general, the effect of that match on a newer player (with only 200 games in the system, for example) is quite a bit larger than is the effect on an established player like Archer.

In this case, that match with Corteza was from the 2013 US Open. It is in the mix for Batman Archer (789 and US #2) but not for Superman Archer(773 and US #6). It is good to have a sense of this, but the "but for xxx" thinking it is a danger path. Do you also throw out his 9-2 and 9-3 losses against Strickland and Shaw?
Click to expand...

I'm not wanting to throw out any matches. I'm just curious how much of a difference one match can make. Can you replace the 11-1 with an imaginary 11-10 and let me know what it changes his overall rating too?
 
Last edited:
Bob Jewett

Bob Jewett

AZB Osmium Member
Staff member
Gold Member
Silver Member
Johnnyt said:
IMO any pool player stats should only be for one year, like most sports are. Johnnyt
Click to expand...
It depends on the purpose of the system. If you want to give the most accurate estimate of each player's ability relative to the other players, I think you want to include all the available data. If you want a ranking system that rewards current participation then you should discard old data to force the player to play in current events. That's what they do in snooker and 3-cushion. Seeding is according to recent performance only.

A difference from some other sports is that teams change through frequent sales and trades while individuals tend to change more slowly.

If Chris Melling doesn't play pool for two years (because he's on the snooker tour) I don't think he should be rated as a 0 if he comes back to pool. I'd rather see a rating that accurately reflects his ability.
 
Jaden

Jaden

"no buds chill"
Silver Member
I think you're missing the point of the fargo rating.

watchez said:
So if I added up correctly, Johnny is 26-24 in his last 50 matches and this goes back to 2013 (which means he hasn't been drawing his ball much lately). Of those 26 wins, I would give quality wins over Karl Boyes, Lee Van, Jason Klatt, John Morra, Ouschan and Jayson Shaw (6 total -- I'm sure someone will find one or two I missed or want to argue about. )


Johnny has 6 quality wins in his last 50 matches over the past 2 years and he is #2 in the USA?? Not trying to discount your hard work with the Fargo ratings but I'll stand by my original statement that it needs some re-evaluating.

Johnny didn't play in any other DCC events? And what about his recent Young Guns vs Old School challenge where Bergman and Sky beat Johnny and Rodney?
Click to expand...

It is designed to find an individual's performance rating and to accurately predict how a single matches outcome will be.

To get it to do what you and I think several others are wanting it to be, there would have to be an algorithm included weighting placement in tourneys as well as games and matches won against specific opponents.

The danger in that though is that draw plays a much bigger role in the distance gone in tourneys and it becomes nigh on impossible to create an accurate algorithm without insane amounts of manual input and opens it up to inputer discretion more.

I think this is a very accurate system and as more people adopt it, it will become more and more accurate.

Jaden
 
mikepage

mikepage

AzB Silver Member
Silver Member
watchez said:
So if I added up correctly, Johnny is 26-24 in his last 50 matches
Click to expand...

you did not. For the 45 matches I posted Johnny won 27 (with an average margin of victory of 5.6 games) and lost 18 (with an average margin of -2.8).

[...]
Not trying to discount your hard work with the Fargo ratings but I'll stand by my original statement that it needs some re-evaluating.
Click to expand...

Message heard.
 
You must log in or register to reply here.
Top