QUINIELAS, PERFECTAS AND TRIFECTAS

INTRODUCTION

After 20 years of betting on horses, I had become fixed on the idea of playing just one horse to win. When I first approached jai-alai, I made straight win bets only. I would wait until one minute to post time, select the team or teams where the scoreboard odds gave me the greatest margin over what I felt the true odds were, and rush to the $10 Win window.I would walk back to my seat, only to find that my 10-1 shot had closed at 4-1. After this happened many times, I began to realize that others were directly or indirectly doing the same thing. Collectively, we were breaking the odds at the last minute.

My next step was to turn to Quiniela betting. But frankly, my experience with horse racing made me dread the idea of "worrying" about multiple entries in the same game. Watching two horses in a horse race, rooting for them to come in 1-2, is pure agony. On the other hand, making combination bets in jai-alai adds to your enjoyment rather than reducing it. In fact I think you will find that some form of boxing of three teams is best from the point of view of maximizing your entertainment value.

If you bet to win only, and your team loses the first time up, it often means a long game for you to watch. On the other hand, if you have three teams, one of them is usually on the court or about to come on, so you have a positive rooting interest in every point. If your top team is cold, one of the others may get hot, running up some points until your favorite can get up again.

Three-team Quiniela Boxes are fine, but Quiniela payoffs are generally low. If you are betting the stronger teams it is difficult to maintain an average payoff as high as $40. There are 28 possible Quiniela combinations, and the fans concentrate their money on the ones that come in most frequently. For example, almost three out of every five Quiniela payoffs involve Teams 1 or 2, and over 75% involve Teams 1, 2 or 3. The fans bet these Quinielas heavily, reducing the payoffs to unsatisfactory levels. If the smart bettor tries to go contra to this by playing Quinielas not involving 1-2-3, he finds that he gets very long losing streaks, even though the payoffs are much higher (e.g., on June 23, 1976 a 6-7 Quiniela at the Hartford Fronton paid $177.80).

Also, as mentioned earlier, the two ways to win with a Quiniela are not equally likely. Suppose you bet a 3-6 Quiniela. You will find that in over 70% of the times you win it will be with Team 6 coming in first and Team 3 in second place. This same phenomenon occurs for other combinations of numbers as well, although the proportions vary.

It is difficult to provide a simple explanation of why this should be so. It has to do with the way that the players come up in rotational order and what sequence of wins and losses must occur for any two teams to come in first and second in that order. Don't worry about it. It happens in the real world and our computers have verified that it is no temporary fluke.

The point is that when you bet a $2 Quiniela it is really like betting two $1 Perfectas, except that the payoff is the same no matter which way the two teams finish. I soon found that if I just bet Perfectas directly I could concentrate my bets on the higher probability combinations, ignoring the ones that hit infrequently.

Actually, it is not quite that simple. The 2-7 Perfecta is just as likely to come in as the 7-2. You cannot follow a simple rule. You have to know the percentages (we will come to them in a moment). But the critical thing to realize is that it is easier to get better odds relative to your chances of winning when you bet Perfectas than when you bet Quinielas. This is because there are 56 Perfecta combinations versus 28 Quinielas you can bet. The more alternatives that you give a betting crowd, the less well they will focus their money. This gives the smart bettor a better chance to find bargains.

BETTING QUINIELAS AND PERFECTAS

Because I find that Perfectas offer a better reward-to-risk ratio, I look to Perfecta opportunities first. In order for me to bet a Quiniela, both ways (e.g., 2-6 and 6-2) have to have about the same frequency of hitting, and theaverageodds on the two corresponding Perfectas must not be more than double the Quiniela odds. Don't let that complicated sentence frighten you.Suppose, for example, that you like Teams 2 and 7. I have already said that they are as likely to come in 2-7 as 7-2. Suppose that the Perfecta odds on 2-7 are 40 to 1, and on 7-2 they are 50 to 1. The average of these two is 45 to 1. Half of 45 is 22.5. Therefore, if the 2-7 Quiniela is 22-1 or less, I will bet the two Perfectas. If the 2-7 Quiniela is 23-1 or more, I will bet the Quiniela.

To understand this a little better, let's assume that the 2-7 Quiniela is only 20-1. If I bet $6 on the 2-7 Quiniela and it comes in (paying $42 for $2), my $6 bet will return $126. On the other hand, if I had bet Perfectas instead ($3 each, for the same $6 total bet), my return would have been $123 if they came out 2-7 and $138 if they came out 7-2. In other words, I could do no worse than the Quiniela payoff and have a 50-50 chance to do about 10% better. It is the little things like this that add up to make you a winner. In actual practice, I find my returns (per dollar invested) average over 10% higher betting Perfectas as compared with Quinielas.

The important things about betting in either pool are that:

(1) your own wagering does not have much effect on the odds, and

(2) the odds are relatively stable so that the odds when you bet are likely to be fairly close to what the final odds will be.Whether you bet Quinielas or Perfectas is up to you. Perfectas involve more work; the odds are not displayed for as long as the odds on the more popular Quiniela; and there are fewer Perfecta betting windows. If you are willing to make the extra effort, you will be rewarded. If not, you can still have fun and come out ahead.

MAKING YOUR SELECTIONS

No matter which way you bet, your first decision should be the selection of your top team. Which team(s) to combine them with comes second.I begin by rating each team. This involves five steps.

1. Write down the skill ratings of the frontcourt man and backcourt man on each team.

2. Add or subtract points for any player who is clearly playing above or below his normal skill level thus far in today's program.

3. Add or subtract points from each team to adjust for post position (see below).

4. Total Items 1 through 3. This is the basic team rating.

5. Add up to 3 bonus points for any pair where you know that their abilityas a teamis greater than would be expected from just looking at their individual skill ratings. (Caution: This is generally more applicable to the more moderate players than the superstars who should play well together.)The team with the highest score is your most desirable team to bet.

It is important to note that for purposes of betting the Perfectas and Quinielas, I do not simply use the team post position adjustments described in Chapter 6. Instead, I use

one-halfof those values, with a maximum of 3 points.

Post

PositionAdjustment to Team RatingsSpectacular 7

ScoringSingle Point

Scoring1 0 0 2 + 0.5 + 0.5 3 - 1 - 0.5 4 - 2 - 1 5 - 3 - 1.5 6 - 3 - 2 7 - 3 - 2.5 8 - 2 - 3 Examples of these calculations will be given in Chapter 8.

The reason that I do not use the full post position adjustments worked out by the computer is that they force you to choose Teams 1 and 2 in a disproportionate number of games. But the odds on Perfectas involving the first two teams are often too low, resulting in the need to choose an alternate.

I have found that use of the less stringent adjustments spreads out my selections more evenly over different post positions, but still produces about one winner out of six games, and generates acceptable average payoffs.

At this point you have a set of ratings, adjusted for post position, and you have a top-rated team. What do you do next?

PERFECTA FREQUENCIES

Not all of the 56 Perfecta combinations come in with the same frequency. If they did, each would win about 1.8% of the time. The actual frequencies with which you can expect to see each combination win are shown below. While these vary slightly from fronton to fronton, they may be taken as a reliable indicator of what to expect over a typical season.

PERFECTA FREQUENCIESPlace Post PositionWin Post

Position1 2 3 4 5 6 7 8 Total

Win %1 3% 2% 2% 2% 1.5% 1.5% 1.5% 13.5% 2 3% 2.5 2 2 2 2 2 15.5 3 3 3 1.5 1 1 1.5 1.5 12.5 4 3 3 2 1 0.5 1 1 11.5 5 2.5 2.5 2.5 1.5 0.5 1 1 11.5 6 2 2.5 2.5 2.5 1 0.5 0.5 11.5 7 1.5 2 2 2.5 2 0.5 0.5 11 8 1.5 2 2.5 2.5 2.5 1.5 0.5 13 Total

Place %16.5% 18% 16% 14.5% 11.5% 7.5% 8% 8% 100% Note that 25 of the 56 combinations have below average (1.8%) winning frequencies. I do not normally bet on these 25 combinations, pinpointing my bets instead among the 31 combinations with above-average percentages.

To be specific, depending on my selection of the team I like most, I usually couple it in Perfectas only with the numbers listed below.

Top Rated

TeamAcceptable Second-Place

Teams in Perfecta1, 7 or 8 2, 3, 4, 5 2 All 3 1, 2 4 or 5 1, 2, 3 6 1, 2, 3, 4 These are the ones listed in the table of Perfecta Frequencies as having a 2% or higher frequency. If the odds are sufficiently high, I will occasionally take a chance on the combinations with a 1.5% frequency when I think the second team is particularly strong.

Do I bet all of the acceptable combinations? No. My favorite approach is to bet only the two highest rated.

What do I do about splitting teams with identical ratings? I use the team with the highest rating without regard to the post position adjustment.

Do I bet regardless of the odds? Absolutely not! I want minimum odds of 25-30 to 1, with just two exceptions. If my top team is #6, the minimum acceptable odds go up to 30-35 to 1, while if the top team is #3, I'll accept 20-25 to 1 (these were all worked out with the aid of the computer).

What happens if the odds are not acceptable? Surprisingly, getting your minimum odds is much easier in Perfecta betting than when betting straight (i.e., to Win only) or in the Quiniela pool. In those situations where the odds are less than I require, and the Quiniela odds are equally unattractive, sometimes I switch to my second-rated team (if they are close to the top-rated team); sometimes I pass the game and go get a hot dog; and sometimes I rationalize and bet my top team anyway. Look, I never said I wasn't human, did I?

Since I've admitted that much, I might as well confess that I also have other weaknesses. I don't go to the fronton and bet as mechanically as I have indicated above. Oh, I do all the rating arithmetic for the entire program. In fact I try to have it done before I even get there, if at all possible. But I have biases and prejudices like everyone else. Sometimes I vary from the rules, but not too much.

For example, one team may not be a clear-cut favorite in my ratings. Or just two teams may stand out from all the rest. I often bet Perfectas involving both of my two top-rated teams, coupled with one or two of the acceptable teams shown in the last table. But my overall control is always how much I bet per game.

Ideally, I bet two Perfecta combinations per game. This costs $6. But I bet an average of $10 per game. Therefore, I have several choices. I can "save" $4 for a later game. I can bet two Perfectas on each choice," over-betting" by $2 on the game. Or, I can bet two Perfectas on my "most preferred" combination, and just one on the other combination, for a total of $9. If the Quiniela odds are more attractive, I may elect to spread my money that way. I'm flexible.

Commonly, I make additional bets involving my top-rated team in

secondplace, with another high-rated team on top on the Perfecta ticket. This provides some coverage against an "almost" win by my favored team. Of course I only do this in a combination which has an above average chance of scoring in the Perfecta pool.My objective is to

cashan average of more than one Perfecta ticket per 12-game program. With one Perfecta per program I usually break even. If I can average more than one, I'm a winner.The general approach I have described picks the winning team for me an average of two times per program, once I have the player ratings established fairly satisfactorily. My approach to selecting Perfectas gets me the place team 40%-50% of the time when I've selected the winning team.

Therefore, I expect to

averagebetween 0.8 and 1.0 Perfectas per 12 games. That is based on buying two Perfecta tickets per game or 24 per program. However, I usually double up such that I actually buy 40 tickets per program. This brings my expected number of tickets cashed up to between 1.3 and 1.7 per program.**2 winners x 40% place x $40/$24 = 1.3

2 winners x 50% place x $40/$24 = 1.7

QUINIELA FREQUENCIES

If you don't want to play Perfectas, and you want to play Quinielas without worrying about the Perfecta odds, there are two approaches I recommend. Both involve boxing or partial boxing of three teams.Before going into them, let's first take a look at the expected frequencies associated with each of the 28 Quiniela combinations. Again it is obvious that not all come in the average of 3.6%. Usually I avoid the 8 combinations which come in with less than 3.5% frequency. These are 4-5, 4-6, 4-7, 5-6, 5-7, 6-7, 6-8 and 7-8.

QUINIELA FREQUENCIESPost PositionPost

Position1 2 3 4 5 6 7 8 1 5.5% 5% 4.5% 4.5% 3.5% 3.5% 3.5% 2 5.5 5 5 4 4 4 3 3.5 3.5 3.5 3.5 4 4 2 3 3 3.5 5 2 2.5 3.5 6 1.5 2 7 1.5 8 (Minor inconsistencies between this table and the

Perfecta Frequency Table are due to rounding)

Under both approaches I use, the first step is to come up with adjusted team ratings as described in the section "Making Your Selections."

In the first method, I box the highest-rated team among post positions 4, 5, 6, 7 and 8 with the

twohighest-rated teams among post positions 1, 2 and 3. The rationale behind this approach is that you just can't afford to ignore the first three teams in Quinielas. So you couple the two strongest among them with the best of the teams in the worst five post positions. When using this approach, I require that thesumof the odds on the three Quinielas in the Box be at least 55.For example, suppose the three teams selected for the Box are 2-3-5. The Quiniela odds on the three tickets might be as follows:

2-3 Quiniela ..... 17 to 1

2-5 Quiniela ..... 20 to 1

3-5 Quiniela .....21to 1

Total58Since 58 exceeds my minimum acceptable "total odds" number of 55, I would make my bet. If not, I would substitute the second highest-rated team among post positions 4 through 8 and check the total odds on a Box involving that team plus the two highest-rated among posts 1 through 3. However, since I am settling for second best, I now increase my requirement for the total odds from 55 to 60. If this test fails also, I don't bet a Quiniela Box this game.

The second method I use simply boxes the three highest-rated teams, with two exceptions. At least one of the teams must be the best of post positions 1, 2 or 3. And I will not play any of the eight low-frequency combinations mentioned above. For example, if my three teams were 2-7-8, I would play the 2-7 and 2-8 Quinielas but not the 7-8. When this occurs I either bet only two Quinielas that game, or I also couple the two higher-numbered posts with another high-rated team among posts 1, 2 and 3, making four Quiniela bets that game.

While either of these methods produces reasonable results, I find that the Perfecta-betting approach described earlier produces a much more satisfactory return on investment over the long term.

TRIFECTA WAGERING

Since the odds in the Trifecta pool are unknown, it should come as no surprise to you that I don't favor it as a way of betting. But, beyond that, Trifecta wagering is such a longshot proposition (the most frequent combination to hit, 5-3-2, comes in less than 1% of the time) that it takes on some of the characteristics of the Lottery.Nevertheless, no book on jai-alai betting can ignore the Trifecta, because the fans love it. I get a big kick out of watching some guy with his face taut, screaming an exhortation at two teams playing off a tie for Show, needing that point to cash his Trifecta ticket. The roars and yells that you hear when a several thousand dollar payoff is posted are all part of the color at the Fronton.

There have been occasions when only one ticket has been sold on the winning combination. When this occurs, that person gets the entire net pool being distributed (usually over $10,000) and an invitation to fill out some forms that the Internal Revenue Service enjoys reading.

Many people pick three numbers and buy an $18 Trifecta Box. If you are the type of person who plays your birthdates numbers, or the numbers on your license plate, or your street address, I guess boxing those numbers is as good a way to bet as any.

But if you want to bet Trifectas in some fashion other than randomly, you should know something about the frequency with which each of the 336 Trifecta combinations pays off. Based on statistics for over 12,000 Trifectas, I have constructed a table of frequencies similar to the ones we looked at for Perfectas and Quinielas. Rather than blur your vision with a giant grid of numbers, however, I simply indicate in the table below whether the various combinations have relatively high, medium or low payoff frequencies.

Keep in mind, however, that any Trifecta combination only hits infrequently. The ones I have labeled

High,for example, only come in a little more often than once in 200 games.

TRIFECTA FREQUENCIESThird Place TeamWin-Place1 2 3 4 5 6 7 8 1-2 LowMedHighHighHighLow1-3 High*MedMedMedMed1-4 HighLow*LowMedMed1-5 HighMedLow*LowMed1-6 MedMedMedLow*Low1-7 MedMedMedMedLow*1-8 *MedMedMedMed*1 2 3 4 5 6 7 8 2-1 MedHighHighHighHighLow2-3 HighLowMedMedHighHigh2-4 HighLow*LowMedMed2-5 HighHighLow*LowLow2-6 HighHighMedLow*Low2-7 MedHighHighMedLow*2-8 *MedHighHighMed*1 2 3 4 5 6 7 8 3-1 HighHighHighMedMedMed3-2 HighHighHighMedMedMed3-4 MedMed**LowLow3-5 MedMed***Low3-6 MedMed****3-7 MedMedMedMed**3-8 LowMedMedMedLow*1 2 3 4 5 6 7 8 4-1 HighHighMedMedMedMed4-2 HighHighMedMedMedMed4-3 HighHighLowLowLowMed4-5 MedMedLow ***4-6 LowLowLow***4-7 LowLowMed***4-8 LowMedMed***1 2 3 4 5 6 7 8 5-1 MedMedHighMedLowMed5-2 MedHighHighMedLowMed5-3 HighHighMedLowLowMed5-4 MedHighMed*LowLow5-6 LowLowLowLow**5-7 LowLowLowLow**5-8 **LowLow**1 2 3 4 5 6 7 8 6-1 MedMedHighHighLowLow6-2 LowMedHighHighLowLow6-3 HighHighMedHighLowLow6-4 HighHighMedLowLowLow6-5 LowMedMedLow**6-7 **LowLowLow*6-8 **LowLow**1 2 3 4 5 6 7 8 7-1 *LowMedHighMed*7-2 LowLowMedHighMedLow7-3 MedMedLowHighMedLow7-4 MedHighMedLowMedLow7-5 MedMedHighMed**7-6 LowLowLowLow**7-8 ***Low**1 2 3 4 5 6 7 8 8-1 **MedMedHighMed8-2 **LowMedHighMed8-3 LowMed*MedHighHigh8-4 MedHighMedLowMedMed8-5 HighHighHighMedLowMed8-6 LowMedMedMedMed*8-7 *LowLowLowLow**- Less than 0.1%What does all this tell us? First of all, there are only 62 "high frequency" combinations. 89 are low frequency, and another 74 are of such infrequency that I don't even dignify them with the descriptive adjective "low." Thus we see that practically half of the Trifecta combinations can be pretty well ignored as they collectively account for less than 20% of all the winning Trifectas. At the high end, on the other hand, less than 20% of the combinations account for about 40% of the winners.

How then would I suggest betting Trifectas? As I said, I don't bet Trifectas, but if I did I would go at it in one of two ways. Either I'd pick one team to win (based on the skill rating system I use for Perfectas) and play it in the high frequency Trifecta combinations; or I would select a

Perfectacombination that I liked and couple it with a third team(s) in a high frequency Trifecta combination.To be more specific, the next table shows the best Trifecta combinations with each of the possible winning teams.

Winning

Post

PositionHighest Frequency Trifectas1 1-5-2 1-2-5 1-2-6 1-3-2 2 2-3-1 2-5-1 2-4-1 2-3-7 3 3-1-2 3-1-5 3-2-5 3-2-1 4 4-2-3 4-3-2 4-1-3 4-2-1 5 5-3-2 5-1-4 5-2-4 5-2-3 6 6-3-2 6-4-2 6-2-5 6-1-5

6-4-17 7-4-2 7-2-5 7-5-3 7-3-5 8 8-5-3 8-3-6 8-5-2 8-4-2 Whether or not the payoffs would be high enough to justify the bets required is difficult to say, given the extreme variability of the payoffs. If you can pick two winning

teamsper program (which you can with the methods explained in this book) and you bet 4 Trifectas ($12) per game, you should hit one about every two or three days. This would involve $288-$432 of wagers for each winner. This seems like a reasonable amount which should be recovered, plus a profit, when you score. But I think you would have to be a several-times-a-week regular before you would have the Law of Averages working for you. Betting Trifectas can produce incredibly long losing streaks.My second approach would be to pick one or two Perfectas, as previously explained, and add one or more third place teams, using the Frequency Table to concentrate on the combinations with the higher chances of coming in. For example, if I liked a 6-2 Perfecta, this approach would bet it in the two Trifectas (6-2-4 and 6-2-5) which are shown as

Highfrequency in the table. The basic idea is similar to the first one: The rifle approach is more effective than the shotgun approach when you have something to aim at (i.e., you have a good fix on the strongest team or the best Perfecta combination for the game at hand).What about boxing and wheeling Trifectas? I'm against it. The reason is that you are making some of your bets on low probability events when you buy a Box or wheel the Trifecta. And they require you to bet a lot. Betting heavily on low probability events is the sure road to the poorhouse.

The most popular Trifecta Box involves Teams 1, 2 and 3. While four of the six permutations of those three numbers are shown as having

Highfrequency in the table, there is oneMediumfrequency combination (2-1-3) and one with aLowexpectancy (1-2-3).There are some Boxes such as 1-2-5 which are all "

Highs"and"Mediums,"but that is not the point. If you are going to bet Trifectas, don't bet unless you know what each combination's chances are. Don't bet the low frequency ones unless you have some reason to believe that its chances or expected payoff are unusually high.My advice is to leave Trifectas to the people who are there to speculate on numbers. You will have a lot more fun (and profits) handicapping the players to come in first and second. That should be challenge enough for anyone.

Probably the best advice I can offer on Trifectas comes from my stockbroker. He went to jai-alai a few times, bet on Trifectas, and lost. Then one night he hit one for several thousand dollars, recouping his losses and showing a huge profit. He hasn't gone back since, arguing: "How could I hope to improve my return on investment?" His logic is sound, but he's been missing a lot of fun.

People who play Trifectas love to talk about their wins. Few talk about their net profits. Few have net profits to talk about. If you hit one or several Trifectas early in the season, particularly if they are big ones, you may come out ahead if you conserve your gains. But anyone who plays Trifectas regularly and heavily has to be awfully lucky to end up a winner on the year. Think about it. If you play an $18 Trifecta Box in every game you are betting $216 per program. If you go six nights, you'll be betting almost $1,300 a week. That means you need two "average" Trifectas per week just to stay even.

SYSTEMS BETTORS

Having said how betting Trifectas is strictly for amateurs, I now will digress to talk about the so-called professional systems bettors allegedly making fortunes betting Trifectas.On October 7, 1977,

The Hartford Courantbroke a front-page story of an investigation by the Connecticut Commission on Special Revenue into "systems bettors" who were reported to have "won more than $1 million in Connecticut during the last year" ... "wagering only on Trifectas." As I was writing this chapter at the time, it naturally came as somewhat of a shockI am not an investigative reporter, but

The Courant'smethod of presenting the story as it unfolded over the next few months struck me as being rather slanted, and it led many uninformed readers to believe that jai-alai is crooked. Among the impressions given were these.1. That fronton employees were in cahoots with certain gamblers;

2. That gamblers had illegal access to computer statistics which improved their chances of winning;

3. That the public was being ripped-off;

4. That the Fronton managements were extending special, and possibly illegal, privileges to professional gamblers; and

5. That some players were involved with the gamblers.As the investigation developed, the story turned out to be quite different. The "systems bettors" were several individuals who bet very large amounts every night, exclusively in the Trifecta pools, and had cashed tickets amounting to hundreds of thousands of dollars over several months. There is, of course, nothing illegal in either betting heavily, or cashing many tickets, or winning.

There is also a significant difference between cashing lots of tickets and showing a profit. Statements relative to their net profits indicated that profits, while significant, were but a small fraction of the amounts trumpeted as "winnings." Nevertheless, they apparently were winning and winning significant amounts.

Jai-alai is not like casino games where the bettors play against the House, trying to "break the bank." The Fronton Management has no interest in who wins among its patrons, or how often, as long as no one is winning dishonestly at the expense of others. On the other hand, since the fronton's revenues are a percentage of the total amount bet, Management does have an interest in people betting heavily. A big bettor is essentially a good customer. Some systems bettors bet $10,000 per night, over 100 times what the average patron wagers.

If the Frontons give them free parking in preferred areas, free admission, and even a special betting area, is this any more than reasonable treatment for high volume customers?

*If security escorts are provided to the parking lots, is this anything but reasonable business courtesy which would be extended to any customers carrying large amounts of cash? The newspapers indicated that this did not appear to be a case of underworld characters trying to launder money, so that was not a consideration.*Treatment varies from location to location. World Jai-alai claims it tries to avoid giving any customers preferential treatment which might be resented by other patrons. Bridgeport Jai-alai, on the other hand, has a special betting window for high volume bettors ($400 or more per game), reasoning that the $3 bettors would rather have these people handled outside normal channels, instead of tying up a regular Trifecta window.

But what about the illegal help?

Big gamblers are traditionally big tippers. It makes good sense. When you are wagering $10,000 a night you can't afford to have mispunched tickets. Systems bettors bet 200-300 tickets

per game. To get these punched correctly requires the absolute attention of the ticket puncher.Ticket sellers are human. If some guy shows up and asks for 300 tickets, they may get flustered, irritated or confused. Three hundred Trifecta tickets involve punching 900 numbers, no easy task. Fingers and mind can become numb. To help relieve numbness, a systems bettor finds a sharp ticket seller and gives him or her a sizable tip. The newspapers mentioned $200/week. This is not illegal. Many people tip both sellers and cashiers, although obviously not in amounts of that size.

If the bettor's profits can cover it, what has been done wrong? As long as other bettors are not being denied the opportunity to bet, who has been hurt? It is no different than tipping the Telewager Clerks as was discussed in Chapter 2. The tipping is not for "hot tips" received, but for efficient service rendered---something of critical importance to the systems bettor.

A more serious allegation was that the Frontons were providing illegal information. It turned out that what was being provided was information about how much had been bet on each Trifecta combination. This information was furnished after the games, however,

not while the betting was going on. While it was bad judgment on the part of the individual who did it (because of the implications of impropriety which might be drawn), the fact is that this same information is available as public information. I consulted it myself in the Wethersfield offices of the Commission on Special Revenue as part of my research for this book. (Also, at the Bridgeport Fronton, it is now posted in the lobby after each game.)What the systems bettors want to study, of course, is average odds on each of the 336 Trifecta combinations. By comparing these to average win frequencies and average payoff prices, they can find which combinations the public overbets and underbets. In other words, they are looking for bargains. This is not illegal. In the investment field it is called basic research. The data is available to anyone. Analyzing it intelligently is hard work.

It was also stated that some gamblers knew some of the players. There have been allegations about fixed games. A tape recording turned up, claiming a conspiracy. It was apparently part of an extortion attempt. Of course, it is not illegal for players to talk with patrons outside the fronton. I have discussed at length why I believe that the game and the players are honest. You will have to form your own conclusions.

What it all boils down to is that certain bettors appear to be betting and winning consistently. Because they are high volume customers, they are treated well by Management. But other than a technical violation involving someone being in an "unauthorized area," nothing illegal or detrimental to the other fans appears to have transpired. Where laws have been violated (e.g., using third parties to cash tickets-to evade required tax reporting), the Frontons have put a stop to such activities.

OK, so how do they do it?

To tell you the truth, I don't know for sure. I have not spoken to any of the people involved. But I am relatively certain that I can deduce what they do and show you how to do the same thing ... temporarily. If too many people try it simultaneously the systems bettors will lose. In fact one of the things that came out in the investigation was that some of the systems bettors were paying others

not to bet the same way.Here is what I believe they do.

Anyone who bets heavily must be "working an edge." Systems bettors make 200-300 bets per game. Since some of them reportedly have been in operation for years, they must be working on a system that exploits permanent inefficiencies in betting patterns rather than one of the approaches which bets combinations that are "due."

That means that they have constructed some sort of Trifecta frequency table such as the one I have given you. Undoubtedly they are betting the combinations I labeled

HighandMediumfrequency almost exclusively.They can't bet them all, however. There are too many (173 actually). If you bet all 173 each and every game, you would lose over an extended period. (The fact that they are betting over 200 tickets per game does not mean that they are buying all different ones. They are most likely buying two or three of each of about 100 different combinations.)

My guess is that they are using a key team(s) concept. This means that they are selecting one or more teams and playing all of the higher frequency Trifectas involving those teams in either the win, place, or show positions. But which of the eight post positions do they play?

To answer this question, first let's consider the proportion of the time that each post position shows up in

eitherthe win, place or show position. Based on the same 12,000+ game sample as before, here are the percentages.

Percentage of games

in which indicated post position

finishes in top 3Post%Post%1 44% 5 36% 2 49 6 31 3 44 7 28 4 39 8 29 As anyone who goes to jai-alai frequently knows, you can't play Trifectas and ignore post positions 1, 2 and 3. At least one of the three is involved in the Trifecta in about 93% of the games played.

Two "key" teams seems to make sense. If you try to get by with only one, you'll be forced to overconcentrate on posts 1-2-3. If you try more than two, you'll be making too many bets. After some trial and error the right balance seems to be to pick two teams, one from post positions 1-2-3 and one from post positions 4-5-6-7-8.

The percentage of games in which either or both of the two teams comes in first, second or third is shown below:

1 & 4 68% 2 & 4 71% 3 & 4 72% 1 & 5 66 2 & 5 70 3 & 5 68 1 & 6 63 2 & 6 67 3 & 6 64 1 & 7 63 2 & 7 66 3 & 7 62 1 & 8 65 2 & 8 68 3 & 8 63 In other words, Teams 1 and 4, for example, will be involved (either or both) in 68% of all Trifectas. That may be surprising, but it is true. By way of contrast, at least one of Teams 7 and 8 comes in first, second or third only 53% of the time.

If the Matchmaker did a perfect job of handicapping, so that team strength differentials exactly offset relative post position advantages, then any two post positions would turn up (one or the other or both) in the Trifecta 64% of the time.

*The fact that some of the percentages are significantly higher or lower indicates that constructing a program is not an exact science.*For the mathematically inclined, this is derived from the formula:

where ! represents a 'factorial', e.g., 4! = 4 * 3 * 2 * 1

I have said that using the methods of evaluating players described in this book I can select two winning teams per night. Similarly, my top selections in each game ought to come in second about twice per night and third about twice, for an in-the money percentage of 6 out of 12, or 50%.

If my second key team in each game came in first, second or third in only 40% of the games, then the proportion of the games in which

at least oneof my two key teams comes in the money will be 70%.*I would expect system bettors, who go every day and see every game, to have a pretty good working knowledge of all of the players' abilities, enough to do as well as I do. Therefore, let's use 70% as a working figure for the percentage of the games that at least one of their two key teams finishes first, second or third.*50% + 40% - (50%)(40%) = 70%

Does this mean they win the Trifecta in 70% of the games? No. Remember that they only bet

HighandMediumfrequency combinations,notall of the Trifecta combinations involving their two key teams.

Assuming their table looks something like mine (a few minor details may vary, depending on their data), they will be making the following number of bets in each game, depending on the two post positions they are keying off.

1 & 4 121 2 & 4 124 3 & 4 127 1 & 5 106 2 & 5 109 3 & 5 106 1 & 6 110 2 & 6 111 3 & 6 108 1 & 7 109 2 & 7 112 3 & 7 107 1 & 8 117 2 & 8 118 3 & 8 111 The average of these is 113. Allowing for the fact that the strongest team is usually in post 5, 6 or 7, one could project that selecting the two key teams based on strength would probably reduce the average to about 110. From the fact that the systems bettors are studying actual betting patterns as well, I gather that, in addition to eliminating low frequency combinations, they are probably eliminating chronically over-bet combinations which produce low payoffs relative to their frequency of finish. So they likely reduce their average number of combinations bet per game to around 100. This is consistent with the statement that they make 200-300 bets per game. They could be betting each combination two or three times.

But, in order to test things out on a more conservative basis, let's assume 110 bets ($330) per game.

In a test of a sample of 1,800 games, these 110 bets produced a winning Trifecta 80% of the time

whenone or both of the two key teams finished in the top three positions. Since one of the two key teams can be expected to be in the money 70% of the time, this means that the systems bettors are probably hitting the Trifecta 80%x70% = 56% of the time.In the sample, the high frequency Trifectas averaged a payoff of $665.

In 100 games (roughly a week's action) this means if you bet $33,000 you would collect 56 x $665 = $37,240 for a profit of $4,240 or 12.8%.

If the systems bettors are betting $10,000 per night with a profit margin of anything like 12.8% ($1,280 per night, it is no wonder they can afford to tip the ticket sellers $200 a week!

As with most gambling lore, there is a certain amount of exaggeration of amounts allegedly bet. $10,000 a night may be an outside figure, but let's assume an average bet of half that amount. If a systems bettor went to 20 programs per month and bet $5,000 each time, he would bet $100,000 per month or $600,000 in the 6-month time frame mentioned in the newspapers. With a 12.8% profit margin, this would mean that the gross amount of the tickets cashed was $676,800, about in line with the $700,000 figure reported in the papers. A $76,800 profit (before tips and other expenses) is not bad for six-months!

So my detective work turned up a systematic approach that has 3 steps.

1. Select the strongest team from posts 1-2-3

2. Select the strongest team from posts 4-5-6-7-8

3. Bet all Trifectas involving either or both of these teams, but excluding thosenotlisted asHighorMediumfrequency on the Trifecta frequency chart in this chapter.A probable refinement is to eliminate certain combinations which usually have a low payoff, but I have not researched this in great detail.

Does this approach really work? I have not tested it with cash, but I did run a paper workout for the opening month of the 1977-8 season at Bridgeport, just using Teams 2 and 5 as the key teams in every game. Remember that those two teams come in 70% of the time. The profit was so high I didn't believe it. I don't believe it would continue. There were several unusually high payoffs. Over a long period, just betting numbers has got to be a self-defeating process.

The systems bettors are extremely vulnerable to competition. The newspaper articles talked about payoffs to various individuals to discourage their playing the same basic system. It is important to understand why.

The 1,800 game sample was based on actual games in Connecticut during 1976-1977. Presumably, the systems bettors' activities are reflected in the results. But suppose just one other bettor had been playing the same system at the same time. The profits would have been cut roughly in half. If two others had been playing, there would have been no profits. This is why they become frantic when others move in.

The reason has to do with the distribution of the Trifecta Pool. In Chapter 6, I explained the problems of the small size of the Win Pool. While the Trifecta Pool is much larger (usually over $10,000), the number of tickets sold on any combination is relatively small. This is why the payoffs are so large; only a few folks are cutting a large melon. But the addition of even one person to the group holding winning tickets cuts down the average payoff. The systems bettors cash a lot of tickets. They can't afford to see much shrinkage in their average payoff. Let's see how this works.

Suppose the net Trifecta Pool being distributed is $12,000. The table below shows what happens to the payoffs if just one additional person is cut in:

Trifecta

PriceWinning

Tickets Sold

=$12,000/(1)(2) + 1 Reduced

Trifecta Price

=$12,000/(3)Percentage

Shrinkage

=1-(4)/(1)(1)(2)(3)(4)(5)$200 60 61 $ 196 2 % 300 40 41 292 3 400 30 31 387 3 500 24 25 480 4 665 18 19 632 5 750 16 17 706 6 1,000 12 13 923 8 1,200 10 11 1,091 9 1,500 8 9 1,333 11 2,000 6 7 1,714 14 3,000 4 5 2,400 20 6,000 2 3 4,000 33 With a $665 Trifecta price, the table indicates a 5% shrinkage by the addition of just one more winning ticket. Actually, with a $665

averagepayoff, the average shrinkage is closer to 7% because of the extraordinary reductions at some of the higher payoff amounts. Note that with a $2,000 payoff to 6 people, the addition of 1 more winner cuts the pie 7 ways, reducing the payoff to each winner by nearly $300.Successful Trifecta systems necessarily involves capitalizing on the fact that there are so many available combinations (336) that the "good ones" are likely to get less attention than they would if the number of choices available to the public was smaller. If too many people get on to a good thing, then the edge disappears and profits do likewise.

In addition to the competition issue, there are other reasons why the number of systems bettors is not large. It takes a sizable amount of capital to get started. Amounts bet are significant (minimum: about $300 per game), and you can get far behind before the percentages begin to work for you.

You need nerves of steel and tremendous self-confidence, plus the ability to work rapidly and with lots of numbers. Not too many people can work accurately with large amounts of money at stake under heavy pressure. Those who can are usually employed at more traditional occupations at large salaries.

There is a big difference between gambling on paper and gambling with cash. The pressure is intense in high-stakes gambling of this type, both internal pressures and the pressure to execute transactions accurately in a brief time period. No wonder the systems bettors want special rooms and special ticket sellers. Every care must be taken to minimize the chance for error.

Newspaper headlines of "$700,000 in 6 Months" sounds like a bank heist. But it isn't all profit, and it isn't at all easy. And it certainly isn't for everyone! Let's get back to betting for you and me.

THE IRON LAW

Return on investment is a basic concept in business. It refers to how much average profit you make for every dollar you invest. It is an important concept for the serious gambler to understand. How much are your total winnings, as a percentage of the total amount you have wagered? That is your profit margin or return on investment.The formula for winning at jai-alai is fundamental. The percentage of tickets you cash

timesyour average payoff on winning bets must exceed your average bet.If you bet Quinielas, then, depending on your average payoff, the table below will tell you how often you must win (assuming you always bet the same amount) to achieve various profit margins.

Your Average

Quiniela PriceMinimum Frequency of Cashing Tickets 25% Profit 35% Profit 50% Profit $45 1 in 18 1 in 17 1 in 15 40 1 in 16 1 in 15 1 in 13 35 1 in 14 1 in 13 1 in 11 30 1 in 12 1 in 11 1 in 10 If you bet Perfectas, then this table indicates the required win frequencies.

Your Average

Perfecta PriceMinimum Frequency of Cashing Tickets 25% Profit 35% Profit 50% Profit $140 1 in 37 1 in 34 1 in 31 120 1 in 32 1 in 29 1 in 27 100 1 in 26 1 in 24 1 in 22 80 1 in 21 1 in 20 1 in 18 As I have said earlier, I expect to cash an average of between 1.3 and 1.7 Perfecta tickets per program. I bet an average of 40 Perfectas per program, so that means somewhere between 1 in 24 and 1 in 31.

Looking at the Perfecta table above, if I hit 1 in 24 bets at an average payoff of only $100 (which is near my minimum acceptable payoff on Perfectas), then the table shows I'll have a 35% profit. On the other hand, if I only cash 1 in 31 bets, but average $140 per winning ticket, the table shows that my return on investment will be 50%.

Thus, we have the

Iron Law of Gambling: The more you win when you win, the less often you have to win. Conversely, the less you get paid when you cash a ticket, the more frequently you have to cash tickets in order to achieve your profit objective. This is true whether you are trying to "win a bundle" or just stay even.The lesson is this: However you bet, you have to watch both how

oftenyou score and howmuchyou score for. When you bet to Place and Show, you cash a lot of tickets, but the amounts are too small to yield much profit. When you bet Trifectas, the payoffs are large but you don't collect very frequently. To achieve a satisfactory overall return on investment at jai-alai, I find that betting Perfectas, and to a lesser extent Quinielas, offers the best balance between high average payoffs and reasonable frequency of winning.

HOW OFTEN SHOULD YOU EXPECT TO WIN?

This depends on how much you bet, how often you win, and your average payoff. If you are betting on the stronger teams and the more favorable post positions, you cannot expect to get the normal average payoff in the pool in which you are betting.As you know by now, my goal is to hit an average of about 1.5 Perfectas per program. I bet an average of $120 per program and collect an average of $120 when my Perfectas come in. This gives me a 50% return on investment if I can hold to those averages over an entire season.

With that kind of a profit margin, people expect me to win most of the time. Yet it doesn't work out that way. Suppose I go to my local fronton once a week. If I attend 52 programs over a year, I can expect

*to lose or at best break even 28 times, or more than half of the times I go! I can expect to hit two Perfectas on 13 other occasions, and in the remaining 11 programs I will cash 3 or more Perfecta tickets. It is on those 11 days that the year's profit is made.*Based on a close fit of a Poisson density function to my underlying data. Chi-square test significant at 99% level.

Of course all that is so much mathematical theory. As all gamblers know, how you do with cash is often quite different from how you do in theory. So let's examine what really takes place during a typical day at the fronton.