On second thought -- disregard the bar table comment because that's not exactly the same thing. I conflated a couple different things there.
Seeing FARGORATE stuff makes me wonder
- Thread starter Pete
- Start date
I think Fargo is likely a modified Elo Rating system or something very similar.
https://en.m.wikipedia.org/wiki/Elo_rating_system
Sent from my iPhone using Tapatalk Pro
https://en.m.wikipedia.org/wiki/Elo_rating_system
Sent from my iPhone using Tapatalk Pro
I had a similar thing happen to me and my son when we tried to rank our game using the AccuStats system, I mean if you play at 850 you are making a mistake 15% of the time, that seemed pretty high. So we played a few sets and kept track of things using their sheets. We ended up in like the 600 range LOL. Let me tell you, seeing that and realizing that about 1/3-1/4 of the time you are messing up sure wakes you up as to how much better you need to get.
Many players go by their ability as their top game, not their average game. Fargo gives you your average game, which given enough data will be pretty accurate. I've been a 550-560 rating for a while, I've had times when I felt like a 600 and would run 2 racks, other times I miss ball in hand. Averages to my 550 score.
Last edited:
Quit bashing something that you clearly know nothing about. Both women and Fargo.
It appears to me he is voicing his concerns,not bashing.You are bashing him:sorry:
If he is so "ignorant" perhaps you can educate him:wink:
JC
Coos Cues
sbpoolleague,
I've been a very vocal supporter of FargoRate. I think it may be the best thing that's happened to pool in decades.
How about YOU educate me on this?
What would happen if there was an island that only played 3 ball and their data was coupled with everyone else's by having an outsider come to them to play 3 ball?
If you can't explain that then maybe it's YOU that doesn't really understand this system.
I've been a very vocal supporter of FargoRate. I think it may be the best thing that's happened to pool in decades.
How about YOU educate me on this?
What would happen if there was an island that only played 3 ball and their data was coupled with everyone else's by having an outsider come to them to play 3 ball?
If you can't explain that then maybe it's YOU that doesn't really understand this system.
Last edited:
Of course but this is clearly NOT the case with Siming Chen. And 3-ball results are not accepted by Fargo. You claim to be a proponent of Fargo yet you continue to propose inflammatory, hypothetical, virtually impossible cases to attack Fargo. Please knock it off.
I'm sure Mike Page and Steve Ernst have considered adding a measure of connectivity to Fargo ratings, which would immediately show just how isolated a person is from the overall field. You MIGHT see an isolated island of players in a remote league somewhere, but you will NEVER see an island in the professional ranks.
Last edited:
All right, Original Poster here.
Let's forget about men vs women for a moment. So please explain how FARGO data is collected, what data is used to determine FARGO RATING, and any direct links that explain it please.
Now after we can see this clearly (since so far it still isn't crystal clear to at least me), then we can try to wrap our minds on why it doesn't or does work to rate men vs women in a nongender way.
This is where I was hoping to be by now...
Thank you for your continued posts ladies and gentlemen...
Here’s the gist (over-simplified). All it takes is win-loss information between players. The two players have a rating. Based off statistical models, if Player A is X number of points above Player B then he is expected to win 10 out of 11 games. If he wins one game, he takes 0.1 points from Player B and adds it to his own rating. If he loses, Player B gets 1 full point from Player A instead. If they continue to play each other and Player A continues to win 10 out of 11 games, they break even and their handicaps never move. If Player B improved and becomes better and starts winning 2 out of 11 consistently, their handicaps re-level to where they will steal points at an even pace for that 9 out of 11 pace instead.
That’s about how individual pairings go. Then I assume the system does recursive analysis on all the data to ensure the ratings are balanced to properly model all submissions from all players in the entire database. This is likely done with state of the art statistical models and analytics technology. More likely it could be built by a couple computer science and statistics college kids in a dorm room.
Bubbles of players are a challenge. If an island existed that submitted data in to the system but never ever played anyone outside the island, their ratings on the island wouldn’t be equivalent to ratings off the island. Local league handicaps are like that. The league bubbles get popped thanks to regional and national tournaments. There are always players that bridge their local bubble of players into the larger national and international network of players. The more cross connections that exist, the more predictive the ratings are between two players in different bubbles.
Fargo claims that they have enough links between players at many levels between the male and female bubbles. Enough that a 790 woman should play even up to a 790 man. Challengers acknowledge there are many links between men and women at low, mid and mid-high ratings but assert there are very few links between the very top men and very top women in the database. They think that can cause the top-most ratings to not be equivalent.
Sent from my iPhone using Tapatalk Pro
Last edited:
You do realize that you are like the millionth person to ask this and it has been answered a million times. A simple Google search will yield more info than you could fathom.
Anyways, the HOW...
League operators and tournament directors submit results of matches to Fargo. Some leagues and TDs use software that make this submission automatic. Some send in results manually. There are safeguards set up to reduce the chance of bogus data being submitted.
The WHAT...
Simply wins and losses between two players. The data was originally limited to three games (8/9/10 ball) but I understand that other games either have been or will be included. When you manually submit data you must include (other than player info and match result) the date of play, game type, table size, and a description of the event. Exactly how these data figure into the Fargo ratings is unknown, and probably proprietary. In order for the math to work and generate an accurate rating, a player must have at least 200 games in the system.
If player A plays player B, and player B plays player C, an inference can be made about the relative strengths of player A and C. If these were the only matches in the system, this inference would be weak, but with Fargo's humongous set of data, just about every player can be connected via a vast spider web, and the relative strength (the Fargo rating) can be calculated.
Take my league for instance. Over half our members have never played rated matches against players outside the league, and they have exclusively played 8-ball on bar tables. But many dozens in our league HAVE played other local and national 8/9/10-ball events on a variety of table sizes. Thus those "isolated" players can be compared to the world through their spider web connections to the dozens of players that aren't isolated.
The Fargo ratings are recalculated frequently, and these calculations involve the entire set of games in the system. Thus if you do not play any games this week but the players that you have played against DO play games, then if their ratings change, yours might change, even though you didn't play.
As the overall data set of games grows, the possibility of having an "island" of players that is not connected to this spider web decreases.
JC
Coos Cues
A million and one and guess what?
Those who struggle with math and logic still don't get it.
They never will.
JC
I understand the math just fine. You guys don't understand my point and I don't think I can be anymore clear.
Mike could easily prove or disprove my point IF he was able to pare the data based on game type but that might not be feasible with the amount of data in the system.
I'm not saying the top women don't play great. I just think the game they play and the way the data is coupled could account for a 15-30 point swing in Chen's rating.
Mike could easily prove or disprove my point IF he was able to pare the data based on game type but that might not be feasible with the amount of data in the system.
I'm not saying the top women don't play great. I just think the game they play and the way the data is coupled could account for a 15-30 point swing in Chen's rating.
I understand your concerns.
You do kind of answer your own ?'s with realizing the greater the robustness the more accurate the fargo!