Jai-Alai Handicapping

How to analyze and handicap jai-alai games

      The first question we must ask is why would you waste your time doing this? The sport's popularity has dropped so much from its peak in the 1980s that there really are just a handful of people who bet the games on a regular basis. The mutuel pools have shrunk to microscopic levels, making it nearly impossible to make a serious profit.

      For the purposes of this article, we'll assume you know the basics of jai-alai. Just looking at the Win-Place-Show Stats tells you pretty much how important post position is. If all the players had equal ability, it would just be a numbers game. But all players are not equal. So how do we accurately rank the players on a fronton's roster? That task will be the focus of this article. Read on, porcupine.

      The best way to "rate" players is simple: keep track of every point played! Unfortunately, that takes a lot of time. You have to watch every performance and spend additional time adding all the numbers up for each player. Unless you work at the fronton or you're, dare I say it, an insane gambler, this is not something that's worthwhile. It's way, way, way, way, way too much time to spend to gain a small edge.

       The fact is, if you watch the action at a fronton a few times per week for a couple of months, you should be able to rank the players fairly accurately. Over that span of time, the players will have had enough games to make their win-place-show records meaningful. Fortunately, there's a relatively easy way to use those WPS records to come up with some good ratings.

       If wins, places, and shows had a certain value, rating the players would be easy. When the frontons run tournaments for the players, a common system used is to give 5 points for wins, 3 for places, and 1 for shows. How accurate is that system? Is a win worth five more times than a show? Inquiring minds want to know! Even if the 5-3-1 system was accurate, the "rating" it produces has no real value. An average player would earn 9 points per every 8 games, an average of 1.125 per game. It's okay, but the number itself has no particular value. We want to come up with a number that has some meaning. Hmmmm ... how would one go about doing that?

       What we need are some stats. Not just some ordinary stats, but some serious in-depth ones. [IG: Ask and you shall receive -- voila! Here are some stats based on 1,072 random 7-point games played at Dania and Miami in 2009.] Don't worry that these stats are several years old. Numbers don't age like players do.

1 168 214 160 15.7 35.6 50.6
2 156 212 180 14.6 34.3 51.1
3 177 169 139 16.5 32.3 45.2
4 162 131 133 15.1 27.3 39.7
5 94 119 144 8.8 19.9 33.3
6 115 87 119 10.7 18.8 29.9
7 92 63 117 8.6 14.5 25.4
8 108 77 80 10.1 17.3 24.7

       You'll notice than some of the totals are a little out of whack compared to what the computer simulations predict. That's not surprising as 1,072 games is a drop in the bucket compared to the million games of results the computer can crunch out. Also, as the next table shows, each post didn't win exactly 50% of its points. Again, 1,072 games sounds like a lot, but it still allows for a substantial deviation from expected results. The table shows the number of points each post won and lost in the 1,072 games. All points counted as one, no matter which round of play they were in.

1 4643 2323 2320 50.0
2 4699 2356 2343 50.1
3 4451 2244 2207 50.4
4 4212 2158 2054 51.2
5 3848 1889 1959 49.1
6 3535 1797 1738 50.8
7 3176 1566 1610 49.3
8 2802 1349 1453 48.2

       This next table [IG: Another table? This page is starting to get a little crowded. PUTP jai-alai guy!] Okay, okay. This next table breaks down the amount of points played and won per game by each post when it wins, places, shows, or finishes out of the money. The bottom "ALL" line represents the total points won and played per game for all the post positions combined in the 1,072-game sample. In other words, the 1,072 winning posts played an average of 5.555 points per game, winning 4.634 of them.

1 5.89 4.70 4.96 3.14 5.07 2.59 3.35 0.84
2 5.82 4.62 5.12 3.27 4.97 2.53 3.46 0.93
3 5.86 4.76 4.72 3.14 4.81 2.54 3.32 0.88
4 5.70 4.79 4.97 3.35 4.75 2.70 3.10 0.90
5 5.57 4.75 5.13 3.65 4.56 2.81 2.88 0.85
6 5.25 4.67 5.04 3.54 4.51 2.99 2.61 0.80
7 5.27 4.58 4.97 3.63 4.73 3.13 2.28 0.69
8 4.49 4.00 4.20 2.96 4.02 2.56 2.07 0.60
ALL 5.555 4.634 4.927 3.296 4.734 2.718 2.808 0.796

       Now let's take a gander at the records of some fictional players and see what can be done to analyze them. Let's start with Normalnaga. As his name suggests, he is an average player. He gets exactly one win, place, and show for every eight games. If the numbers in this formula are correct, he should be winning 50 percent of his points. To figure out about how many points he played in his 400 games, multiply his wins (50) times 5.555, places (50) times 4.927, shows (50) times 4.734, and the out-of-the money finishes (250) times 2.808. Add the four numbers and you get 1,462.8.

      To compute the points he won, multiply his wins (50) times 4.634, places (50) times 3.296, shows (50) times 2.718, and out-of-the-money finishes (250) by 0.796. Add the four numbers and you get 731.4. Divide 731.4 by 1,462.8 and you get exactly 50 percent! Hey, it works! I know this might sound complicated at first glance, but once you understand the concept, it's pretty easy.

      Now we'll try to compare Torreaga to Elbum. Torreaga has many more wins but Elbum has more places and shows. Who is better? If we do the math, Torreaga appears to have won 50.9 percent of his points, while Elbum checks in at 47.5. His places and shows couldn't make up for his anemic win percentage. Looking at the other two players, Solotegui has an excellent rating of 52.9, while Carrera is at a lofty 56.3.

So there you have it, a simple way to rate Jai-Alai players! [IG: Warning! It's not that easy. There are all sorts of flaws in this system.]

Ummm, yeah, I was getting to those.

PLAYER  G   W   P   S   OTM W% WP% WPS%
NORMALNAGA 400  50  50  50  250 12.5 25.0 37.5
TORREAGA 401  66  44  33  269 16.5 27.4 35.7
SOLOTEGUI 372  60  40  64  208 16.1 26.9 44.1
CARRERA 250  53  39  29  129 21.2 36.8 48.4
ELBUM 379  25  54  60  240   6.6 18.0 36.7

      Okay, so there a lot of things that can go wrong with this system. First off, as those financial brokerage ads say, "Past performances are not necessarily indicative of future results." Or something like that. If we're halfway into a season and a player has won 53 percent of his points, there's no guarantee he'll repeat that the rest of the year. Baseball is similar to this in that just because a guy has a .340 batting average at the All-Star break, that doesn't mean he'll end up at .340. But if he's a good player who has a career average over .300, we know he should continue doing fairly well.

      Another fault in this system is that it underrates the good players and overrates the bad ones. How so? It stands to reason that the best players will perform better than average in all of the WPS and OTM situations. For example, the system credits a player who finishes out of the money with approximately 0.8 points won out of 2.8 points played. The good players are more likely to have better numbers in that situation. Just how much is tough to tell. The opposite is true for the poor players. They're more likely to get shut out in a game. How much of a difference does all this make? Not sure, and in fact, it really doesn't matter all that much. If the system rates a player at .525 and in reality he's .535, that's no big deal ITIGHO.

      The important thing is to have all the players ranked so that when handicapping a game, we can easily see who's best and who's worst. Now if jai-alai had huge mutuel pools and we had access to a supercomputer that could spit out winners based on the numbers we input, we'd want to be exact as possible. But if we're just scribbling the ratings on the program and merely want to have a general idea as to who the favorites should be, it's no big deal.

      It's extremely important to know in what games the player normally plays in. Most frontons put the weaker players in the early games and the stronger players in the late games. A player who wins a lot in the early games will not have the same success in the late games. Let's say Jose, who plays in the early games, has a rating of .523. But he's listed in Post 1 in a later game. It looks good on paper. Heck, it looks great on paper. But it's similar to a minor league baseball player getting called up to the majors. He's batting .345 in the Pacific Coast League. BFD. Until he proves he can compete with the big boys, you just have to take a wait and see attitude. The bottom line is this: the best players in the early games are still the worst players in the late games. GIGIG.

      More to come...this is a work in progress.

      IG: Hey, How long has it taken you to write this article? Six months? A year?? Five years??? Right now it's January 17, 2016. See if you can get this done by the end of the year. PUTP or we'll find somebody else to finish it. GIGIG


IG takes pride in having accurate and updated information. If you see something that needs to be corrected, added or deleted, please e-mail us at: error@insanegambler.com

Unleaded IG

IG = Always On Tilt