Is your handicap league fair?

[bltuneup]

Hi guys. I’m new to this forum, but I’ve been a league bowler for decades. I’ve always wondered whether my handicap leagues were truly fair. That is, did all teams have a realistic chance of success? That’s obviously true when the handicap percentage is 100% (and the handicap base is high enough), but what about the majority of leagues that use a lower percentage to give a deserved advantage to teams with higher averages?

In recent years, my team finished near or at the bottom of the standings for 10 straight seasons in my company’s bowling league, which used a 90% handicap percentage. Even when we bowled well relative to our capabilities, our results were always the same. This seemed to be strong evidence our league was not fair. But could I prove it? I decided to put my background in math and computer programming to work to see if I could.

In researching this subject, I found no evidence anyone had ever taken a scientific approach to determining the proper handicap percentage to assure a particular league’s fairness. Instead, I found huge amounts of misinformation and cockamamie theories being disseminated on a multitude of bowling websites. (No…allowing negative handicaps or setting the handicap base number too high does not punish good teams! No…putting an upper limit on handicaps or setting the handicap base number too low does not promote fairness…it destroys fairness!)

I was particularly horrified to find that the worst offender in the misinformation department was the U.S. Bowling Congress itself, with that ridiculous study of theirs that concluded the playing field isn’t completely leveled until the handicap percentage goes all the way up to 116%. If anyone were ever silly enough to set up a league this way (I sure hope no one has!), the poorest teams would win almost every season.

After five years of research and coding, I am proud to have created an app called Bowling League Tuneup that does what I believe no software has ever done before. It evaluates a league’s fairness and helps find a league’s optimal handicap formula, given the attributes of the league and its specific mix of bowlers and teams. It determines each team’s odds of success (you know your odds when you play blackjack or craps…you should know them when you enter a bowling league too!), and it can also detect sandbagging and dumping, help explain league outcomes, and produce bowler performance sheets that can be distributed to league bowlers.

My app showed that, as I suspected, our company league was grossly unfair. Poorer teams like mine would have needed superhuman performances to have had a fighting chance against the higher-average teams. After using my app to evaluate many other leagues, I can say the league fairness problem appears to be widespread.

I of course want to make Bowling League Tuneup known to league bowlers far and wide, and I invite everyone to get more information about it at http://www.bltuneup.com. But my primary goal in writing here is to find out if the bowling community shares my perceptions and concerns about league unfairness and its effects on the popularity of league bowling. I would also love to help dispel the misinformation floating around in cyberspace about the proper ways to set up handicap leagues. I’d greatly value the thoughts and questions any of you might care to share. Thanks!

[bowl1820]

Okay basically ad for bowling software.

Give us a example of a league and the fair handicap your software came up with and explain why it is fair?

[bltuneup]

Okay basically ad for bowling software.

Give us a example of a league and the fair handicap your software came up with and explain why it is fair?

OK, now I’m sorry I put in my URL. Please don’t consider this to be an advertisement. I am not a company trying to find a free place to advertise. I am genuinely a league bowler who has been obsessed (to an unhealthy degree) with the issue of league fairness since I found my league to be grossly unfair way back in 2009.

But all the work I put into creating software seems to have been the easy part. Getting the bowling community to know about what I’ve done is the hard part. This looks like a very popular bowling site, so I thought I’d post here.

I know this is a problem at my own bowling center here in the Los Angeles area. They said they can no longer fill their leagues because people are sick of seeing the same teams winning season after season. That should not happen in a handicap league. What team would join a league knowing they have almost no chance of success?

To answer your request for an example, I’ll use the small 16-week 2009 summer league I was in that served as my motivation for beginning my project. Our handicap formula was 90% of 220. But one bowler got his average up to 233. Since negative handicaps were not allowed, his handicap bottomed out at zero, giving him and his team a tremendous unfair advantage. I was stunned to find that most league bowlers thought he was actually at a disadvantage because we all had handicaps and he didn’t. I found that trying to explain basic arithmetic to my fellow bowlers was futile.

There were eight teams in the league with averages of: 769, 719, 712, 707, 699, 591, 590, and 577. Yes, this is a huge range of averages, which was of course one of the problems. (I was on the worst team, of course.) The 233 bowler was on the team with the 712 average. So we have two questions to answer:

1. If every team bowled basically their average in this league, what were each team’s chances of a first-place finish with a handicap formula of 90% of 220 and negative handicaps not allowed?

2. What would those chances be if the handicap formula were changed to 90% of 240 to accommodate the 233-average bowler?

You’d think (or at least I would) that someone somewhere would have had software that could make such a computation. Not that I can find. So I made my own software, which calculates the chances of each team coming in each possible place of finish. The approximate probabilities of each team coming in first place, using the original handicap, are as follows:

769: 23.9%, 719: 12.2%, 712: 47.4%, 707: 10.2%, 699: 9.0%, 591: 1.3%, 590: 1.3%, 577: 1.0%

Results total a little over 100% because ties credit both teams with the higher place of finish. Notice that the third-best team can be expected to come in first place almost one season out of two. And guess what? That team ran away with first place that season.

Adjusting the handicap to 90% of 240 (or just allowing negative handicaps) changes the results to this:

769: 35.1%, 719: 18.5%, 712: 16.8%, 707: 15.7%, 699: 14.1%, 591: 2.3%, 590: 2.3%, 577: 1.8%

The 233 bowler no longer had an unfair advantage, so at least now the results have a correlation between team average and expected chances of success. BUT…this league is still horribly unfair, with the three poorest teams expected to come in first place about one season for every 15 or 20 seasons the best team will.

Adjusting the handicap formula to 95% of 240 puts more emphasis on bowling well relative to ability and less emphasis on high scoring. The results now:

769: 24.2%, 719: 17.2%, 712: 16.3%, 707: 15.9%, 699: 14.9%, 591: 6.5%, 590: 6.4%, 577: 5.9%

You can see how much fairer this league would be. You can also do what I call “what-if” analyses to determine what kinds of performances a team like mine would need to have the kinds of chances the best teams would have.

There is much more to explain, including how things like league length and splitting seasons affects teams’ chances, but this note is far too long already. I just want to make it clear what my software does and – most importantly – that way too many leagues are simply not fair because there have never been any tools to measure fairness. Unfair leagues are killing league bowling and I don’t want that to happen!

I understand your skepticism about my intentions here, but please please believe me that I’m not just here to sell something.

[bowl1820]

Taking a quick look through the manual, it's similar to a Sabermetrics program.

It's basically taking statistical data on player performance and using that to track trends and compute the probability of who will win and then allowing you to adjust the parameters to get the out come you want. So you could set up a league based on those new parameters.

The program looks good, I'll have to read the manual more in depth (the idea of dual handicaps in a split season looked interesting.) and try the free trial.

[Aslan]

I've been in 80% 240, 90% 220, scratch, and 100% 210.

As far as how fair they "Seemed"...

4) Scratch- Because the absent bowler got a 160 average in a sport league...so if you don't average 160...you're always hurting your team.
3) 100% of any number is just silly. My team was relatively BAD...3 guys that have rarely bowled and one guy that hasn't bowled in decades...and we finished 4th.
2) 80% 240 seemed stupid because nobody in the league even averaged over 220...and guys with high averages were still getting handicap...making it even harder on below average teams to be competitive.
1) 90% 220 seemed fair. The better teams were near the top of the standings...but so were some lesser teams.

There's no perfect system. The teams that generally win...have a formula that works well in handicap leagues. 2 above average bowlers, a good female bowler, and 2 low-level bowlers that will get better. The good guys sandbag in the first few weeks...the low level guys establish horrible averages...and then the better bowlers start performing while the beginners learn the game and rapidly improve.

My team right now is completely non-competitive. We have no chance of ever winning. Because we have a 100 average girl bowler with no interest in bowling "properly" and we have an elderly guy that isn't going to get any better. Since they rarely have a really good game (much less series)....their handicap doesn't come into play. If I was putting together a team to dominate a handicap mixed house league...I'd find an elite male anchor, an elite/good female, and then I'd look for two young men that are new but taking lessons, practicing, etc... Then the next season, I'd replace the two new guys with two other new guys that practice, take lessons, etc... You'll win every year.

[bltuneup]

Taking a quick look through the manual, it's similar to a Sabermetrics program.

It's basically taking statistical data on player performance and using that to track trends and compute the probability of who will win and then allowing you to adjust the parameters to get the out come you want. So you could set up a league based on those new parameters.

The program looks good, I'll have to read the manual more in depth (the idea of dual handicaps in a split season looked interesting.) and try the free trial.

Yes, that’s basically it. This tool allows a league manager to set a handicap formula that gives appropriate weights to the two main factors that influence a team’s success in a handicap league: average and performance relative to average. Once that’s been done for a league, the teams joining the league can be assured it has been configured to give every team a realistic chance to succeed. I don’t know of any other tool that can give league bowlers that assurance.

Again, I am trying to ascertain how much of a problem league fairness has been considered to be by league bowlers. I must admit that for my first 20 or 30 years of league bowling, I just assumed that the standard handicap formula my bowling centers have used for all their leagues must work just fine. Now I know that is not the case. I would love to hear from forum participants about whether you perceive fairness (assuring every team has a fighting chance to succeed) to be a potential problem in your own handicap leagues. If this is a problem that is not perceived to even exist, that will obviously be a pretty big impediment to my pitching a solution!

bowl1820, I really appreciate your interest and would greatly value any feedback you may have!

[bowl1820]

Again, I am trying to ascertain how much of a problem league fairness has been considered to be by league bowlers.

Given if you go by just what has been said on the bowling forums, the perceived problem with the current handicap system is that it gives low average bowlers/teams too much of advantage. Which causes the secondary problem being that of sandbagging.

For the years that handicap horse has been beat to death, the main conclusion which everyone seems to have reached.

Is that it's not so much the handicap system, it's the fact that easy conditions and ball advancements allow lower average bowlers to bowl more over their average, more often than higher average bowlers are able to.

There is also the thought that the average system currently used is part of the problem, in that it doesn't represent the bowlers true ability well enough. Thus skewing the amount of handicap a bowler receives.

This connected with the high handicaps that low average bowler/teams have, is perceived as making it unfair to high average bowlers/Teams. (The high average bowlers always over look the long term statistics that show they win more often than not.)

As for sandbagging again given if you go by just what has been said on the bowling forums. Then everyone is a sandbagger except for "Me and Thee".

[bltuneup]

I've been in 80% 240, 90% 220, scratch, and 100% 210.

As far as how fair they "Seemed"...

4) Scratch- Because the absent bowler got a 160 average in a sport league...so if you don't average 160...you're always hurting your team.
3) 100% of any number is just silly. My team was relatively BAD...3 guys that have rarely bowled and one guy that hasn't bowled in decades...and we finished 4th.
2) 80% 240 seemed stupid because nobody in the league even averaged over 220...and guys with high averages were still getting handicap...making it even harder on below average teams to be competitive.
1) 90% 220 seemed fair. The better teams were near the top of the standings...but so were some lesser teams.

There's no perfect system. The teams that generally win...have a formula that works well in handicap leagues. 2 above average bowlers, a good female bowler, and 2 low-level bowlers that will get better. The good guys sandbag in the first few weeks...the low level guys establish horrible averages...and then the better bowlers start performing while the beginners learn the game and rapidly improve.

My team right now is completely non-competitive. We have no chance of ever winning. Because we have a 100 average girl bowler with no interest in bowling "properly" and we have an elderly guy that isn't going to get any better. Since they rarely have a really good game (much less series)....their handicap doesn't come into play. If I was putting together a team to dominate a handicap mixed house league...I'd find an elite male anchor, an elite/good female, and then I'd look for two young men that are new but taking lessons, practicing, etc... Then the next season, I'd replace the two new guys with two other new guys that practice, take lessons, etc... You'll win every year.

Aslan, you’ve put your finger on a key to success in a handicap league: in-season improvement. Lower handicap percentages put more emphasis on high scoring; higher handicap percentages put more emphasis on improvement…specifically, on bowling better each week than whatever average is being used to set your handicap. This is, of course, why sandbagging is such a huge problem.

Fortunately, as my app shows, the use of book averages for some number of weeks to start the season is highly effective in preventing sandbaggers from cheating the system. I think all handicap leagues should use book averages for this reason.

You said the 240 base number of the handicap formula was “stupid” because no one in the league was averaging anywhere near that. The base number is in the formula solely to turn high averages into low handicaps and low averages into high handicaps via subtraction. It is not there to influence competitiveness; that’s what the percentage is for. Whether that number is 240 or 260 or even 300, the results of your league would be exactly the same!

Suppose the league’s best bowler had an average of 220 and you had an average of 180. If the formula were 90% of 220, your handicaps would be 0 and 36, respectively. In other words, you’d get 36 more pins added to each game you bowled than he would…almost, but not quite, making up for the 40-pin difference in your averages. If the formula were 90% of 240, your handicaps would be 18 and 54. You’d still get 36 more pins added to each game than he would. So he gets no advantage from that handicap. Even if the formula were 90% of 300, there’d still be a 36-pin difference between your handicaps. High numbers have no effect on a below-average team’s ability to be competitive!

Low base numbers, on the other hand, can take a fair league and destroy it. If the handicap formula were 90% of 200, the 220 bowler would have a handicap of 0, while you’d have a handicap of 18. Before, if you both bowled your averages, you’d lose by 4 pins; now you’d lose by 22 pins. You’ve essentially said to the 220 bowler that you’ll let him add 18 extra pins to every game he bowls before determining the winner. You’d never overcome that kind of disadvantage in the long run.

The only problem with using high base numbers is that handicap scores tend not to look like bowling scores anymore. It looks stupid to beat someone 348 to 323. If that bothers people in your league and you therefore want to keep the base number low, it is imperative to allow negative handicaps. If the handicap formula were 90% of 200, your handicap would still be 18, but the 220 bowler would have a handicap of -18. In other words, he’d have 18 pins SUBTRACTED from his score each game. Again, people think this punishes the good bowler…but it doesn’t! You’d still get 36 more pins added to each game than the 220 bowler would and he'd still win by 4 pins if you both bowled your averages. If all leagues used negative handicaps, it would put an end to all the discussions about what base number to use and all the problems about what happens when a bowler gets his or her average over that number! Negative handicaps should be mandatory for all leagues because there is no downside to using them. But it appears psychology keeps triumphing over math, all to the detriment of league fairness.

As for your evaluations of the success of various handicap formulas, please keep in mind that these are your perceptions of what worked for those specific leagues. There is no one-size-fits-all handicap formula that works for all leagues. The length of the league, the difference in skill levels (averages) between the best and worst teams, and several other factors all have profound influences on what handicap formula should be used to maintain a league’s proper competitive balance. And you’d really need hundreds or thousands of seasons of results to derive a mathematically sound conclusion about what formula works well and what doesn’t. No one can do that, which is why I wrote my app.

Regarding your result in the 100% league, keep in mind that while a scratch league rewards only high scoring, with your performance relative to your average being irrelevant, 100% leagues are exactly the opposite. Your average is irrelevant, and only your performance relative to your average matters. The problem is, only the average is published, so you have no measurement of your performance relative to your average. My app computes this latter value (I call it the bowler’s trend), helping to explain handicap league results that simply can’t be explained when the only published value for each bowler is his or her average. I’m guessing your team achieved a very good trend value that season, explaining your 4th-place finish.

[bowl1820]

Fortunately, as my app shows, the use of book averages for some number of weeks to start the season is highly effective in preventing sandbaggers from cheating the system. I think all handicap leagues should use book averages for this reason.

The majority of leagues have been doing that already. (using book averages and holding them for several weeks at the start of the season just for that reason. And going a step farther with the use of a 10 pin drop rule.)

[bltuneup]

Given if you go by just what has been said on the bowling forums, the perceived problem with the current handicap system is that it gives low average bowlers/teams too much of advantage. Which causes the secondary problem being that of sandbagging.

For the years that handicap horse has been beat to death, the main conclusion which everyone seems to have reached.

Is that it's not so much the handicap system, it's the fact that easy conditions and ball advancements allow lower average bowlers to bowl more over their average, more often than higher average bowlers are able to.

There is also the thought that the average system currently used is part of the problem, in that it doesn't represent the bowlers true ability well enough. Thus skewing the amount of handicap a bowler receives.

This connected with the high handicaps that low average bowler/teams have, is perceived as making it unfair to high average bowlers/Teams. (The high average bowlers always over look the long term statistics that show they win more often than not.)

As for sandbagging again given if you go by just what has been said on the bowling forums. Then everyone is a sandbagger except for "Me and Thee".

I must respectfully disagree that current lane and ball conditions somehow benefit lower-average bowlers more than higher-average bowlers. An equally persuasive argument could be made that only very good bowlers are able to take advantage of specific oil patterns and the monster hooks that today’s balls facilitate, giving poorer bowlers an even bigger disadvantage than they had before.

In any case, saying any factor can allow a bowler to bowl over his or her average more often doesn’t make much sense to me. Even if specific conditions do raise a bowler’s average, that has nothing to do with being able to bowl above your average…which is presumably even harder to do after you’ve raised your average. At the end of the season, each bowler will have approximately the same number of games above and below whatever average he or she achieves. That’s how averages work.

I completely agree with you about the PERCEPTION that poorer teams are given too much of an advantage in handicap leagues. I’ve argued (oops…I mean I’ve had level-headed conversations) about this in my own leagues constantly. It’s always the teams that always finish at the top of the standings that ***** about low-average teams being given too much of an advantage…despite the fact they keep finishing at the top of the standings. We’d need to consult a psychologist to explain this phenomenon, I think.

The perception boils down to the notion that any time a low-average team finishes high in the standings or (heaven forbid!) in first place, the whole idea of handicapping is fatally flawed, as it must be giving poor teams an unfair advantage. The best bowlers and teams tend to believe they should win every season…because they’re better!!! They also tend not to be able to articulate an answer to the question of why a lower-average team would ever want to join that league. In my app’s user manual, I call this “scratch-league mentality.”

I believe the problem with perception is that when a poorer team does well in a handicap league, there’s often no way to know WHY it did well. So everyone becomes a conspiracy theorist and concludes the league must be giving an unfair advantage to poorer teams.

Scratch leagues reward high scoring. Period. 100% handicap leagues reward bowling well in relation to your established average. Period. Leagues with a handicap under 100% reward a combination of the two. The lower the handicap percentage, the more high scoring is rewarded; the higher the handicap percentage, the more personal success is rewarded.

The problem? Everyone knows everyone’s average, but no one knows how everyone has done in relation to their averages. Without that second number, you can’t understand why a poorer team ended up so high in the standings. My app computes this second number, which I call a trend. Simply put, it’s the average number of pins that you’ve bowled above or below the average being used to set your handicap, computed over the entire season.

In the user manual, I give an example of a real league in which a very poor team finished in second place. I show that its trend was far better than any other team’s. So the league wasn’t set up poorly…it rewarded the team that did better in relation to its established level of performance than any other team. Which is what handicap leagues are supposed to do.