Is the math around this number calculatable or is it a "secret" formula we don't really know?
(Apologize if this has been beat up already.)
(Apologize if this has been beat up already.)
Fargo Rating
- Thread starter BarTableMan
- Start date
How the ratings are determined -- the method of calculating them -- is public. Much of it is described in the FAQ page at https://www.fargorate.com/
Briefly:
Gather a record of match scores of games of eight ball, nine ball and ten ball (edit: and one pocket) for as many players as you can. Joe beat Willie 10-5 on June 3rd, 2017, and so on. Currently the list has 28 million games and 300,000 players with at least one game recorded .
Calculate the most likely relative strengths of all the players in the system that would produce that set of results. That such a calculation is even possible is amazing, but it can be done even for such large numbers of players. There is a similar calculation for college football rankings. "Relative strengths" of two players means the win ratio between them expected for a long set. In the case of Joe and Willie, Joe beat Willie by a 2:1 ratio.
Turn those expected win ratios into ratings on an easy-to-use scale. In the case of Joe and Willie, a 2:1 win ratio means they are 100 rating points apart. Because the ratings are all relative, there is no absolute scale -- say 1000 for a "perfect" player. The top players are set to about 800 and the rest of the universe is set relative to them.
There is a simple formula to turn the rating point difference between two players into the expected win ratio, such as 2:1 for Joe and Willie. Basically every 100 point difference is another factor of 2 in the ratio. If a player is 400 points below SVB, a match between them is expected to have a 16:1 win ratio. There is a simple calculator on the FargoRate website for this.
The ratings can be thought of as a huge map with the towns being players and the roads being win/loss records between the players. Because everything is connected together -- I think there are no islands -- any match anywhere will affect your rating. As a simple example, if Joe has a match with Charlie who is a new player, that will set Charlie's rating. If Charlie now plays Willie, the result will likely change Charlie's rating and that in turn will change Joe's rating because we have a better rating for Charlie. Joe's rating changes even though he didn't add any games to the system. Everything is connected. A match between two players three stops on the map away from you will affect your rating but probably won't change it much.
Old games have less weight than recent games. A match that is three years old has half the weight as a recent match, etc.
New players are handled a little differently with a starter rating that disappears after they have 200 games in the system. (If your only match in the system was a 5-0 win in league, that is an infinite ratio and you are probably not infinitely better than all other pool players.)
There are probably a few other details.
Last edited:
Are they calculating 1P matches, gambling matches? How is it possible?
I think it's a 'Gaff' and in the end will Not help pool.
I think it's a 'Gaff' and in the end will Not help pool.
(Edited)
There are criteria for what data they will accept. It has to be a pre-scheduled event so that the players can't control inclusion of the data after the fact. 14.1 is not included but one pocket has been added recently.
Lots of people find the ratings very useful. The players who want to sneak up on people, not so much.
As far as not helping pool.... I would describe the recent Fargo-rating-limited event in OKC -- $1000 entry, full field of medium speed players -- as creating a lot of interest in pool.
Last edited:
Amateur tournaments would like you to hold their beer
I believe 1 pocket is now included.
Have you actually seen anybody that could run 3 balls, that wasn't trying to sneak up on somebody in the pool room Bob? Maybe I could be less cynical if I could just find 1, myself.
There is no way to stop all cheating , but the Fargo system seems to me to do about as well as you can.
Last edited:
I also believe one pocket is now included as of about one year ago. The fargorate FAQ still states 8, 9, 10 ball only.
Bob, how long do you think the computer calculations take? Is it a split second? A few seconds? A few minutes? etc.
My understanding is that there are two phases. The first phase with the new day's data added is a quick approximation. Sort of small tweaks on the large structure that already exists. The next phase removes the small errors that were introduced by the quick stuff. Sort of like smoothing off rough edges after the major chiseling. I think that takes hours or most of a day. I believe it is an iterative process that notes discrepancies and adjusts the ratings to get smaller errors, and then the process is repeated. The iterations end when the largest discrepancy is less than some small value. I think the effect on your rating due to a Eurotour event doesn't show up until the second phase.
(A similar kind of calculation is done for large electronic circuits with thousands or millions of transistors. Such calculations on large networks have been a topic of intense research since about 1975.)
Last edited:
The idea of Fargo is great, but it's pretty frustrating when tournaments are now having a Fargo rating cap along with minimum robustness requirement. In certain area of the country, local tournaments don't report to Fargo so it's near impossible to achieve an established rating.
It's rather annoying when you have to go out of your way to try and build a Fargo rating, just to play in a specific tournament. I understand the goal is to level the playing field, but there will always be outliers.
It's rather annoying when you have to go out of your way to try and build a Fargo rating, just to play in a specific tournament. I understand the goal is to level the playing field, but there will always be outliers.
Is it true, that a player's FR never goes down?
These days in the rooms where I play there is little action so not much sneaking up.
No. If you read my description above -- maybe it's a little long -- that should be clear. If you have a spell of playing poorly your rating will go down. Your great play 12 years ago is discounted by a factor of 16 relative to your recent games.
I believe Mike said a few years ago a non-active player’s rating might end up effectively frozen if they never played again. It was in response to a question I asked about how could Nick Varner be a 777 at the time (maybe 5 years ago). I just looked up his rating now, and it’s a 781, which would put him on the Mosconi team as a player, not a captain.
Edit, here is the post (and thread) if anyone is interested why Varner is currently a 781, and was a 777 in 2019. https://forums.azbilliards.com/threads/nick-varner-is-a-fargorate-777.502238/post-6516956
Edit, here is the post (and thread) if anyone is interested why Varner is currently a 781, and was a 777 in 2019. https://forums.azbilliards.com/threads/nick-varner-is-a-fargorate-777.502238/post-6516956
Last edited:
Oh man I wish this was the case
I'm unaffiliated with those working on Fargo, but I can guarantee that you rating can most certainly go down.
Can You show an example?Oh man I wish this was the case
I'm unaffiliated with those working on Fargo, but I can guarantee that you rating can most certainly go down.
Need to come down here $100-5K a game one hole any night of the week lol
You've been around forever
(I mean that as a compliment). Have you ever seen a player get better or worse, that was not a kid coming up, or a 60 year old going down? I know I never have.
Here's my progress below.
I started to keep track of my progress in 2020. Wish Fargo would do this instead. Left is Fargo rating and right is games. Everytime it updated with a tournament I played in I recorded it.
You can definitely drop. I see lots of people going down. I also was in 700 range before I started recording it back in 2020. 700 was a bit high for me back then and it dropped to 680 range.
I tried my hardest to break 700 and when I finally did it felt awesome lol. It was quite the grind
Attachments
Last edited: