Handicapping a match ?

DynoDan

AzB Silver Member
Silver Member
The owner of a private billiard club I used to frequent became quite familiar with the speed of regular members who entered his 14.1 tournaments, and would assign an appropriate spot to the lesser players. I started many games to 150 already 75 points behind.
 

DynoDan

AzB Silver Member
Silver Member
P.S. I assume the only fair way to calculate a handicap would be to record the scores of all previous games (over an extended period), and average out the difference #s for each matchup.
 

John Biddle

AzB Silver Member
Silver Member
Straight Pool Handicapping

I don't claim to have the BEST way to handicap straight pool but I do have a way that has worked well for my 14.1 league. We're about to start our 13th season on Monday, so it's had a bit of exposure to the real world and I think it has done well.

Everyone has a handicap, a number from 200 to 1300 (if we get even better players we would just raise the top number). There are 22 Skill Levels within this handicap range, each one 50 points. The names are not tied to actual pro levels or what others may think an "A" is, they're just the names we use.

1250 1300 Ranked Pro
1200 1249 Tour Pro
1150 1199 Pro
1100 1149 House Pro
1050 1099 Open
1000 1049 ShortStop
950 999 AAA
900 949 AA
850 899 A+
800 849 A
750 799 A-
700 749 B+
650 699 B
600 649 B-
550 599 C+
500 549 C
450 499 C-
400 449 Learner
350 399 Student
300 349 Novice
250 299 Tyro
200 249 Beginner


Each person was assigned a handicap when thay first joined the league, in the middle of the most appropriate Skill Level based on what several of the more knowledgeable players (knowledge of people and their skills, not knowledge of 14.1) thought. If players join in a season after having played with us before, they come in at the handicap they left with.

After every match the two players handicaps are recalculated based on a formula, so other than the initial assignment of handicap there is no subjective input. If you win you go up, if you lose you go down. The amount is determined by how much, in percentage terms, the loser missed their target. The bigger the miss, the more the players hcps go up/down. Hcps move more quickly in the 1st 6 matches a player is in the league to help the hcp become accurate if we've not made an accurate initial assessment. There is also a maximum amount hcps can move up/down for any one match (40 & 26 for 1st 6 and after).

Targets are determined based on the two player Skill Levels. The higher skilled player will go to 75, 100 or 125 depending on which Skill Level he/she is in. The lower skilled player will go to a lower target based on how many skill Levels they are below the higher skilled player, with each level worth 10%. For example, say a "B-" plays an "A". The "A" is the higher skilled player, and goes to 100. They are 4 levels apart so the lower skilled player goes to 90%x90%x90%x90% of 100, or 66. We have a table posted at thew pool hall which eliminates the need for anyone to understand the system or do any math, I just wanted to explain how it works.

Over many seasons the system has proven to be quite fair and accurate. To test that I look at how the stronger players do against weaker opponents. If one side is winning most of the matches the handicapping would be favoring them. In our league the percent of stronger player wind varies from 45% to 55% season to season. Give our moderate sample size I don't think we could do much better.

That doesn't mean there are never any complaints about handicaps, that's a perennial hazard of handicapping, but I am confident that our players all feel the system is fair and accurate, except right after they've just lost.

This is all much easier to do than to explain.
 

Bob Jewett

AZB Osmium Member
Staff member
Gold Member
Silver Member
What is the best and most fair way to handicap a match ?

Thx in advance !
Steve
The system John Biddle described above is more or less the system I described in a column in Billiards Digest that is also used in several rooms in this area. Some of the details are different but the main idea is the same: players are automatically adjusted up or down in rating according to whether they win or lose their matches. Higher rated players spot lower rated players according to how far apart their ratings are.
 

Seth C.

AzB Silver Member
Silver Member
The system John Biddle described above is more or less the system I described in a column in Billiards Digest that is also used in several rooms in this area. Some of the details are different but the main idea is the same: players are automatically adjusted up or down in rating according to whether they win or lose their matches. Higher rated players spot lower rated players according to how far apart their ratings are.

Bob fed the details of his system to me last year when I started a small 14.1 league among friends. It has worked well.

One thing that I did -- given that we had a number of novices -- was to open the season with a get together where everyone was given 10 attempts at "offense only" runs, starting with any break ball they wished. Like at the Derby City. From 10 attempts, everyone had an average, which could be compared against the averages of the others. Then, I played a preseason match against each of them, to see how they did against me. A common opponent. These two exercises allowed me to set initial handicaps for players about whose games I knew very little. Interestingly, the handicaps held up pretty well.
 

John Biddle

AzB Silver Member
Silver Member
An all offense "test" along with everyone playing a common known opponent seems like an excellent way to get good starting handicaps, but would take a lot of time.

Many of our players were new to the game but not new to pool. We made our initial estimates for hcp based on their skill in other games, assuming, I think correctly, that they would come up to speed in 14.1 fairly quickly. Not that they would know it as well as their strong game, but their skill would carry them a long way.

It has worked pretty well, although we've made a mistake or two along the way that Seth C's system might have minimized or even eliminated.
 

stevekur1

The "COMMISH"
Silver Member
I don't claim to have the BEST way to handicap straight pool but I do have a way that has worked well for my 14.1 league. We're about to start our 13th season on Monday, so it's had a bit of exposure to the real world and I think it has done well.

Everyone has a handicap, a number from 200 to 1300 (if we get even better players we would just raise the top number). There are 22 Skill Levels within this handicap range, each one 50 points. The names are not tied to actual pro levels or what others may think an "A" is, they're just the names we use.

1250 1300 Ranked Pro
1200 1249 Tour Pro
1150 1199 Pro
1100 1149 House Pro
1050 1099 Open
1000 1049 ShortStop
950 999 AAA
900 949 AA
850 899 A+
800 849 A
750 799 A-
700 749 B+
650 699 B
600 649 B-
550 599 C+
500 549 C
450 499 C-
400 449 Learner
350 399 Student
300 349 Novice
250 299 Tyro
200 249 Beginner


Each person was assigned a handicap when thay first joined the league, in the middle of the most appropriate Skill Level based on what several of the more knowledgeable players (knowledge of people and their skills, not knowledge of 14.1) thought. If players join in a season after having played with us before, they come in at the handicap they left with.

After every match the two players handicaps are recalculated based on a formula, so other than the initial assignment of handicap there is no subjective input. If you win you go up, if you lose you go down. The amount is determined by how much, in percentage terms, the loser missed their target. The bigger the miss, the more the players hcps go up/down. Hcps move more quickly in the 1st 6 matches a player is in the league to help the hcp become accurate if we've not made an accurate initial assessment. There is also a maximum amount hcps can move up/down for any one match (40 & 26 for 1st 6 and after).

Targets are determined based on the two player Skill Levels. The higher skilled player will go to 75, 100 or 125 depending on which Skill Level he/she is in. The lower skilled player will go to a lower target based on how many skill Levels they are below the higher skilled player, with each level worth 10%. For example, say a "B-" plays an "A". The "A" is the higher skilled player, and goes to 100. They are 4 levels apart so the lower skilled player goes to 90%x90%x90%x90% of 100, or 66. We have a table posted at thew pool hall which eliminates the need for anyone to understand the system or do any math, I just wanted to explain how it works.

Over many seasons the system has proven to be quite fair and accurate. To test that I look at how the stronger players do against weaker opponents. If one side is winning most of the matches the handicapping would be favoring them. In our league the percent of stronger player wind varies from 45% to 55% season to season. Give our moderate sample size I don't think we could do much better.

That doesn't mean there are never any complaints about handicaps, that's a perennial hazard of handicapping, but I am confident that our players all feel the system is fair and accurate, except right after they've just lost.

This is all much easier to do than to explain.


Thank you John,

We tried the up/down weekly adjustment. And we all feel that it is not fair when even players meet each other and one would have to spot the other maybe a large amount all depending on how there schedule was vs the other.

Steve

What I am currently doing when an upper player plays a lower handicap player I go by the BPI (average inning count) to calculate the game. It has proven to be quite an accurate gauge. Though it takes players to record every inning each game on paperwork. Which is just as easy as putting a slash on a piece of paper for each player every time they go to the table. This method has gotten me within 10 points on final game score most every time I have used it. Unless there was an extreme variable. The way I figure it, the player that plays to their standard or better should win based on the numbers. This helps keep everyone on their toes.
What I am looking to solve is those extreme variables that I hinted at, to see if their were a better way to calculate a game.

I have been doing this for 12 seasons myself, and i am always striving to improve every aspect of the league .

Thanks so much
Steve
 

John Biddle

AzB Silver Member
Silver Member
We tried the up/down weekly adjustment. And we all feel that it is not fair when even players meet each other and one would have to spot the other maybe a large amount all depending on how there schedule was vs the other.

Sounds like you trust your subjective estimates of player ability much more than the ones based on the adjustment formulas. I get that, most people do.

One possibility is to limit the amount that a hcp can rise or fall because of one match. I make it fairly large for the 1st 6 matches in the league and then a smaller amount after that, and smaller yet as the handicaps become stabilized because of a much more extensive history.

For example, each Skill Level is 50 points. If a match were truly lopsided and the loser only got 50% of their target, and they were new to the league (within 1st 6 matches) they would drop 40 points and depending on where they were in their Skill Level, probably drop one level.

If they had been a player more than 6 matches, the maximum would be 26, or ~half an SL. For a seasoned player, this maximum would be 15. You could have more gradations or different maxima if you found that works better.

Let's say two players were equal SLs at the start of a season but weren't scheduled to play until the end of that season. If one was playing poorly and the other well over a 3 month season, why do you think it's unfair that the one playing well should have to spot the other? Don't you think based on this recent history that she/he's more likely to win?

Also, if matches are not blowouts the hcps change less. And, people tend to both win and lose, so hcps don't get far off course before being self-corrected.

Our players were not interested in keeping detailed info about ther matches. Good detailed data would possibly enable a better algorythm to adjust the hcps, but I'm not sure. I am sure they don't want to gather the data. It would be quite complicated, I think, to develop a system that would use all the collected info like average run lengths, safety success ratios etc. My guess is that all that info is contained within the win/loss scores. It doesn't really matter much why or how a victory was achieved, but that it was.

Your current system, in place for quite awhile, sounds like it's working well. I get that you're open to improve, but is there something that's driving your original request? Some nagging issue, minor perhaps, that gives you pause about the way you're currently doing it? If so, let's explore that and see if we all might benefit.
 

Seth C.

AzB Silver Member
Silver Member
An all offense "test" along with everyone playing a common known opponent seems like an excellent way to get good starting handicaps, but would take a lot of time.

Many of our players were new to the game but not new to pool. We made our initial estimates for hcp based on their skill in other games, assuming, I think correctly, that they would come up to speed in 14.1 fairly quickly. Not that they would know it as well as their strong game, but their skill would carry them a long way.

It has worked pretty well, although we've made a mistake or two along the way that Seth C's system might have minimized or even eliminated.

John, your observations are understandable. It might seem that the offense only test could take a long time. But unless you are talking about pretty high level players, the runs end pretty fast. Also, players can do this on their own, using the honor system. But the main point I want to make is that I only used this element because I was working with a small group of friends most of whom had played pool in the past but never at a high level and not for some years. So, I didn't have the benefit of knowing their 9 ball or 1P games -- because they didn't have any!

For anyone thinking to use the offense only test, I might suggest having each player start each run with the balls nicely spread on the table and with BIH. I found, as you might suspect, that the break shot had too much influence on the outcome (how well or poorly the balls were spread had too much effect on the ultimate average run length).

What took a lot of time was the common opponent matches.

I agree that if the person running the league has a good handle on the speed of the incoming player, that's a good basis to set the initial handicap.
 

Bob Jewett

AZB Osmium Member
Staff member
Gold Member
Silver Member
As far as adjustments go, I think they have to be automatic and by a public formula. I think it is a mistake to make them subjective.

In a league I ran for several seasons, the largest problem was accurately placing new players. Since it was a drop-in league -- show up on any Monday night and you will be matched up in one or two or even three league matches -- new players could appear at any time.

I guessed as well as I could, but many of those guesses were rough and some of them were based on bad info. My usual procedure was to adjust them a little extra immediately based on their first full match, and then after four or five matches I would calculate the equivalent of their 14.1 FargoRate based on all of their scores. That seemed to work pretty well.

In my system, the match lengths are arbitrary, but usually the weak players would play to 50 or 60 (for the better of the two) and strong players would go to 100 to 140 depending on how quickly they played. One strong player who was quite slow usually went to 80 and finished after the other matches. The point was to get the first round matches to finish at about the same time so that players could have a second match if they wanted one.
 

John Biddle

AzB Silver Member
Silver Member
@Seth C I like your system but was especially concerned about the time for lots of matches with the same player. By now it would probably work pretty well for the relatively few new players I get each season. But, they tend to be 8 or 9 ball players and usually have an APA or other handicap that I can use if we don't know their play.

It might be interesting to get as many people who've played before to take a turn at all-offense and self-report their 10 rack totals. Could be very useful as a base for future initial hcp estimates. I might give this a try.

@Bob Jewett Yeah, slow players can indeed be a problem. Thankfully, most of the slowest players are the lower skilled ones.

The drop-in league is cool, though I don't think it would ever work for us, there just aren't enough players. But our use of anytime scheduling has made it so that just about anyone interested can find a way to play.
 

Pushout

AzB Silver Member
Silver Member
An all offense "test" along with everyone playing a common known opponent seems like an excellent way to get good starting handicaps, but would take a lot of time.

Many of our players were new to the game but not new to pool. We made our initial estimates for hcp based on their skill in other games, assuming, I think correctly, that they would come up to speed in 14.1 fairly quickly. Not that they would know it as well as their strong game, but their skill would carry them a long way.

It has worked pretty well, although we've made a mistake or two along the way that Seth C's system might have minimized or even eliminated.

I did something similar in a league I started for people I worked with who liked to play pool but usually in a bar setting. I can't remember for sure now but it seems to me that we worked with some kind of adjusted formula that Bob Jewett came up with and wrote about it Billiards Digest. Or, I may have done it on my own, I'm not sure now. I THINK we used Equal Offense scores We played so many weeks of 8 ball, 9 ball, and Straight Pool and handicapped based on the system. Worked ok and I DID NOT WIN!! Gave myself a tough handicap because I played better than everyone else.
 
Top