8-Ball Ratings (long)

K

Keith Buck

AzB Silver Member
Silver Member
The thread on the Fargo Ratings made me do some thinking on comparing players from different areas. The problem, as I see it, with most of the rating systems based on results is the difficulty of comparing the ratings on results against completely different opponents.

We use scoring average in our VNEA league to calculate handicaps and that works OK since everybody is facing the same players. The problem with ratings comes up when large tournaments are held with players from many charters like the provincial championships.

A player who has an 8.00 average in the top league in Toronto is a much better player than the player with a 9.00 average in the mixed league in the small city where I live. I read the same thing in the U.S. as far as the APA where people say that an APA 7 in NY is usually superior to an APA 7 from a more rural area.

I started to think about a possible way to rate players where the opponent would be irrelevant. Total ERO's (running the table on your first visit with 15 balls on the table) would be one way but the data set would be pretty small in a lot of leagues where the best players only get 1 ERO per ten games played and the majority of the players get 0,1 or 2 for the whole season.

The idea I came up with last night was to count the number of balls made per ERO attempt (an ERO would be eight, missing the eight ball for your ERO would be 7, etc.). I was thinking that you would only count the attempts where at least 4 balls were made so the minimum possible score would be 4.0 and the top possible score would be 8.0. I would eliminate the attempts under 4 as there are situations where you might want to just get control of one set of balls and then play safe or you might not have a possible shot after the break and these situations would skew the results to the low side.

The league organizer would still have some leeway to rate players higher than their number in some situations. I'm thinking about a situation where all but one of the top ten players in a league had a number over 6 except for say the third place player who was at 4.5. Obviously something about his/her style of play would be making the number lower than the actual skill level (maybe he/she likes to sink four balls and then play safe in a high percentage of games).

Anyway, I just thought of this last night and would just like to hear any opinions and maybe spark some discussion. The stats would have to be kept for a league or tournament for a while to get a big enough dataset to see if a handicap system or a player rating system could be devised from this.
 
Last edited:
subscribing to thread

Very curious to see what kind of opinions and/or suggestions this thread raises.

At the very least, it seems that you put some time and thought into this topic, so TAP TAP :grin:
 
It is an interesting thought, I like your line of thinking. The only downside is defining an ERO "attempt" ... that sounds subjective. If the guy runs 4 balls, falls near a cluster, then bunts into that cluster and plays a perfect safe... was he trying to run into the cluster, then resorting to plan B? Or was plan B really plan A all along?

The VNEA has some complicated formula that rates a player by their power average. It factors ERO's, safeties, and of course win rate. I don't know the formula but I found myself agreeing with it... the top 5 or so guys by power average were exactly who I would have considered the top 5 in our league.

I do think that when played "correctly" 8b is a game of runouts and the really good players don't even try to make a single ball until the out is possible. And at the highest levels, the out is always possible. Watch some of the old IPT videos. They still have stats on each pro's break and run percentage. I don't care how good a guy is at moving, safeties, etc... it's hard to fade someone who breaks and runs 40% of his games (which doesn't even factor rack'n'runs, or 2nd inning outs from an opening safety).

The other thing is that different tables are more susceptible to outs than others. Actually I never firmly established that 8b runouts are easier on a barbox or not. I hear scary stories of top level players who never won a major but can run 15+ racks on the box. But my personal experience is it seems to be harder, with the beat cloth and funny rolls ...and lousier spreads and tiny sides... and un-drawable mudball and crowded lanes.

A fairly solid indicator should be not so much runouts, but how many innings the shooter gives to his opponent. If he intentionally gives up an inning on a safe, but it's Super Effective™, he should then be out when he gets to the table next. His strong safety play will show through in terms of reducing innings vs. some other guy's mediocre safeties that get him back to the table... but then he has to safe again.
 
I agree with OP's idea in general, but it sounds a bit too complicated to use in practice. Collecting stats other than just the match results is pain in the ass, even if it's set up separately. There are people doing this for the pro matches, but I guess they are paid to do it or have other motivations. For the vast majority, no one wants to deal with all the hassle.

I think something like Fargo Ratings is the way to go. It sure has the problems that you describe, but there are ways around the problems. And if not, say there are completely separated player pools, it's just not applicable and that's ultimately all there is to it.

I hope I don't sound too pessimistic, because I think you have a fair point and I personally wouldn't mind a system like this. But ultimately I think it is not realistic enough. I would love to be proven wrong, though.
 
CreeDo said:
It is an interesting thought, I like your line of thinking. The only downside is defining an ERO "attempt" ... that sounds subjective. If the guy runs 4 balls, falls near a cluster, then bunts into that cluster and plays a perfect safe... was he trying to run into the cluster, then resorting to plan B? Or was plan B really plan A all along?

That would certainly skew the result lower but if the number of games was high enough I don't think it would happen too often where a good player is going to sink 4 balls and then look for the safety so hopefully it wouldn't change it by more than .1 or .2. Most books I've read say "Don't sink more than two balls unless you think you can run out."

The VNEA has some complicated formula that rates a player by their power average. It factors ERO's, safeties, and of course win rate. I don't know the formula but I found myself agreeing with it... the top 5 or so guys by power average were exactly who I would have considered the top 5 in our league.

I like the power average better than the score average for handicapping as well. The average score makes a player who wins all his games by an average of 10-6 equal to the player who wins all his games by an average of 10-4 if their win percentage is the same.

I do think that when played "correctly" 8b is a game of runouts and the really good players don't even try to make a single ball until the out is possible. And at the highest levels, the out is always possible. Watch some of the old IPT videos. They still have stats on each pro's break and run percentage. I don't care how good a guy is at moving, safeties, etc... it's hard to fade someone who breaks and runs 40% of his games (which doesn't even factor rack'n'runs, or 2nd inning outs from an opening safety).

The other thing is that different tables are more susceptible to outs than others. Actually I never firmly established that 8b runouts are easier on a barbox or not. I hear scary stories of top level players who never won a major but can run 15+ racks on the box. But my personal experience is it seems to be harder, with the beat cloth and funny rolls ...and lousier spreads and tiny sides... and un-drawable mudball and crowded lanes.

I agree completely there. If one charter plays under pristine conditions while the other charter is using decrepit equipment that would make a difference.

A fairly solid indicator should be not so much runouts, but how many innings the shooter gives to his opponent. If he intentionally gives up an inning on a safe, but it's Super Effective™, he should then be out when he gets to the table next. His strong safety play will show through in terms of reducing innings vs. some other guy's mediocre safeties that get him back to the table... but then he has to safe again.

That leads back to the level of your opponent again, though. If you were playing a D player you might get ball in hand, a B player might hit the ball but leave you a shot and Efren might hit it and return the safe.
Click to expand...

Thanks for the good reply.
 
JarnoV said:
I agree with OP's idea in general, but it sounds a bit too complicated to use in practice. Collecting stats other than just the match results is pain in the ass, even if it's set up separately. There are people doing this for the pro matches, but I guess they are paid to do it or have other motivations. For the vast majority, no one wants to deal with all the hassle.

I think something like Fargo Ratings is the way to go. It sure has the problems that you describe, but there are ways around the problems. And if not, say there are completely separated player pools, it's just not applicable and that's ultimately all there is to it.

I hope I don't sound too pessimistic, because I think you have a fair point and I personally wouldn't mind a system like this. But ultimately I think it is not realistic enough. I would love to be proven wrong, though.
Click to expand...

Thanks for the reply.

I agree that keeping track would be a bit of a pain but it is certainly easier than the scoring that is required for APA now. Maybe the best thing would be for a bunch of the players who are planning to play in tournaments to get together early in the season and have each player break thirty or forty games and play the "ERO attempt" after each break (whether they sink a ball or not) to get a "score" and then have the league organizers compare the scores with the other charters to see where the players should rank on a relative basis.
 
These rating have been a hot topic here in Minnesota. When trying to find a rating system, there are always going to be flaws, no matter what.

Mike has a great rating system in Fargo. The more players that begin in that leagues, and continue to accumulate games, the more accurate his system is going to work.

I agree with the ERO idea that you have brought up. I believe there should be a factor based on an Eight-Ball RunOut (in the exact terms you have described, with the 15 ball approach.) Mike's system analyzes on a game by game basis. The more games you have in the system, the more accurate that will be. But if you could use an adjustment factor based on an ERO, I think the ratings would be even more accurate faster.

I have always played in a points format (VNEA style leagues, etc.) and I like the idea of a points format versus a win/loss format. Players that have a lot of fire power and run multiple racks can close the gap much quicker than just winning a couple of games.

MVP points: points scored minus points given up
I have always believed that MVP points are the best item to look at when analyzing the strength (rating) of a player. Over a period of time, the stronger players will achieve many more points scored than points given up.

In a 25 game, 5 player, round robin format, a 40-10 week is very strong. Most likely you had 4 tables runs for you and one against you. Some people with argue and say that a 50-35 week is stronger because you won every game. I strongly disagree, you managed to only open a 15 point gap between your opponents where the other player opened a 30 point gap.

A buddy of mine created a site to originally joke around about this type of stuff with a small group of friends. Throughout the years, we have kept our stats in a database and can always go back and look through them. We have gone as far as breaking out the games you break versus the games you rack.

Here is what the site originally started as:

http://www.erostats.com/erostats/stats.cfm

Eventually, he decided to market the website for leagues. Feel free to take a look at the main site as it is today.

www.erostats.com

The site is nice where you can see an abundance of statistics. You can even go back and look at an actual matchcard between teams.

The rating system will always be tough to come by. It is difficult to gauge one's speed as compared to others. There will always be flaws, but over time, hopefully those iron themselves out.
 
The BCA uses amount of balls. THis may sound good but there are too many things affect the ratings.

Not all 8 or 9 are the same. Joe does a B&R gets 10 for the win and the opponent gets 0.

Next round. Opposing player breaks comes up dry. Joe runs out. Joe gets 10 opponent 0

Joe does a B&R. Joe gets 10 opponent 0

Opponent breaks and makes a few balls. Joe runs out 6 balls and misses a breakout. Opponent wins.

Joe get 36 and opponent gets 10 for the night. Joe does something like this every week. So he averages 9 balls and is ranked 9

John wins 3 out of 4 every night and averages 36 points but he is not a runout player. The only reason that he wins is because his opponent clears off the table but fails to run out. So his opponents usually average 28 to 31 points against him. John does this almost every week. So he is ranked a 9.

John and Joe play in different division. The one John plays in is weaker players that are C and B. Joe plays in a stronger division where they are mostly B and A.

A lot of the players in Joes division are also victims of B and R. Each time they shoot the breaker either does a run out or the other player run outs. Their scores for a night could 10-0, 0-10, 10-0, 1-10. In this situation they are averaging 21 balls for 4 games and could be ranked a 5.

When the players from each division for playoffs those from the weaker division could end up giving a spot to stronger players.
 
You must log in or register to reply here.
Back
Top