Fast Track Simulcapping

April 2, 2005.

Class, Performance and Readiness:

Guidelines for qualifying favorites and near favorites.

Introduction

Prior to the 2004 Breeders' Cup races, the concept of Qualified Favorites (the "QFs") was re-presented here. Since then, a further study and its analysis has led to a better understanding of qualifying or eliminating favorites and near favorites through the use of sound handicapping considerations of class, performance ability and readiness to race. Handicappers can use this process as a handicapping tool to help them to make an assessment of favorites and near favorites.

Fabricand's "Principle of Maximum Confusion"

If you read the earlier study on QFs, you will recall that it began as an analysis of Fabricand's study from the late 1960s. The truths he noted regarding the uncanny ability of the Betting Public to predict over the long haul the probability of all horses' chances to win, and narrow in on the most likely winners, remain true today. However, his model suggested that the favorite selected by the Betting Public is essentially to be viewed as a single selection. He considered that the Betting Public could be confused or fooled by horses that had similar records and attributes, and looked for a way to find favorites offered at better odds that had been underbet by his crowd persona. In practice he ended up with a complicated system that few of us can master.

Another View of the Betting Public

Fabricand was correct in his appreciation of the wisdom of the decisions reached by the Betting Public through the democracy of the tote board, but that wisdom is not confined to the first choice as a single selection. Rather than considering the Betting Public as the equivalent of a person, and the favorite as a single selection of such imagined person, it is more accurate to view the selection processes of Betting Public favorites and near favorites in a race as a distillation of many different approaches to handicapping, and a consideration of innumerable factors that are impossible for any one person to take into account in handicapping a race. The result of the Betting Public's collective handicapping processes is an assessment that is translated into odds and probabilities through the tote board.

The Challenge

Collectively the Betting Public does a superior job of handicapping, race after race, better than any single expert. So why not take advantage of all that hard work and collective wisdom? Why not start handicapping at the point where that hard work leads us, and apply a little fine tuning to improve our selection decisions?

Class, Performance and Readiness

Assessing the Betting Public's handicapping prowess against the rigor of basic, sound handicapping principles is likely to improve selections results, and lead to the elimination of some unsound choices. The converse is also true in that selections based upon consideration of those handicapping elements are likely to be strengthened by the addition of the rigor of comparison to the Betting Public. When the two decision making processes coincide, coming up with the same selection for a race, the chances of success is likely to be stronger. When the Betting Public preferences and sound handicapping principles fail to reach agreement this should set off alarms, and there is reason for concern.

There are 3 time-tested primary handicapping essentials to be taken into account by any handicapping method: — class, performance and readiness. The first two are basically different approaches to measuring ability while the second indicates whether or not a horse is ready to run on the day of the race.

1. Class — the AER ("average earnings ratio")

   Class among race horses is a hard thing to quantify. There are many ways to look at it and for it. Examining the racing conditions is a common approach, but in recent years this has become a more difficult task with the proliferation of different categories of allowance races and combination allowance and claiming conditions within the same races. Another way is to look at the average earnings per race. The level of competition and the consistency of success combine as component factors, measured in earnings. It is a time tested approach, and has the added advantage of being straightforward to calculate, especially with the aid of computers. In Quirin's 1979 study of 1,803 horses with the highest earnings average in the race, he found an average return of $1.82 without considering any other handicapping.

   The Fast Track Handicapping AER rating is similar to Quirin's calculation, although it makes use of the average earnings over two years, with more emphasis on the current year. However to achieve more useful and readily comparable numbers we make a further adjustment to the raw scores. After calculating the average earnings figures, each is compared as a percentage of the highest raw score. The highest AER per race is thus 100 (unless there is a scratch of that horse), and a horse that had average earnings half as high will receive a score of 50.

2. Performance − the B2 and FTS speed ratings

   A horse's racing ability can be measured in terms of its past performances in prior races. It is generally acknowledged that it is best not to rely on just one race in determining ability, such as the most recent or the best race. One race may easily be uncharacteristic of a horse's ability.

   To measure ability most experts now rely upon speed ratings rather than considering just the raw time values of a race. Speed ratings attempt to make adjustments between tracks, distances and take into account the differing track conditions from day to day. Generally the assign an arbitrary value of 100 against which performances are compared with the higher numbers (or "figs", short for "figures") generally indicating better performances. There are many different brands of speed ratings out there. The most famous, and the one published by the Daily Racing Form (the DRF), are the Beyer speed figures. Because the Beyer figures are so readily accessible they have a significant influence on the odds, and many experts recommend creating your own or using other speed figures to obtain an edge.

   The B2, LR and FTS figures published by Fast Track Simulcapping each rating measure performance ability a little differently. The B2 is the average of the best two speed ratings in its current racing record of up to 10 races. The FTS measures an average of the best of the two most recent races and the best of the remaining 3 in the last 4 races. The LR is simply the speed rating for the last race.

3. Readiness − recent work (the DLR & the TF scores)

Once questions of class and performance have been considered attention should be given to whether the horse will be ready to run to its class and performance levels in the current race. How much recent activity has a horse had in terms of furlongs raced or in workouts leading up to this race? No matter what a horse's inherent class, prior performances or native ability, unless it is in condition to run it will not be able to perform to that level. The DLR ("days since last race") and the TF ("total furlongs raced or worked in last 21 days) are useful indicators of how ready a horse is likely to be. Unless the horse meets a minimum criteria for work in preparation for today's race, it should be eliminated from consideration as a potential selection.

Testing The Theory

Depending on the size of the field, horse races are generally won by one of the first three favorites about 70% of the time. In this study, a decision was made to analyze the top three Betting Public preferences in each race, and to collect our data on rankings and measurements contained within the Fast Track Simulcapping daily published information.

The raw data for the study was the ratings and calculations for 151 races from randomly selected race-cards at tracks from different regions across the United States. The tracks included Aqueduct, Calder, Churchill Downs and Hollywood Park. Races including two year olds were excluded, as were races which had numerous scratches and races at distances of less than 6 furlongs.

The total number of horses analyzed was 453 (i.e. 3 per race). As before, in our other studies, the effectiveness or score of the handicapping tool was measured in terms of win percentage and the theoretical average return on each two dollar wager to win. All horses evaluated were one of the three preferences established by the Betting Public at post time.

The tables that follow indicate how various class, performance and readiness measurements, calculated and provided by Fast Tack Simulcapping's computers, interact with those top three selections of the Betting Public without other consideration or handicapping analysis. For more background on what the various ratings are, and how they are calculated, please view our web page at http://www.simulcapper.com/howto.htm.

In many cases more than one horse may have a similar score or rating within the same race. This will naturally tend to skew the scores negatively since at best only one of the two could win. The useful B2 rating is a good example. 278 horses were within the top four B2 ratings in their races. Since there are 151 races this means there are almost 2 horses per race (1.84) with one of the top 4 B2 ratings in the field. Despite this, 27% of those horses won, for an average return of $2.15.

The Results

The results of the study will focus first on the general success of the top 3 Betting Public preferences. Then attention will be focused to each of the three handicapping elements when combined with those Betting Public preferences − Class, Performance and Readiness − followed by an examination of combined results.

BETTING PUBLIC PREFERENCES

Table 1 shows the average winning results and percentages of the Betting Public’s top three favorites. These results are as expected with 73% of all races won by one of the first three Betting Public choices. In our sample the Betting Public's first choice wins more often than the second choice, and that choice wins more often than the third choice. If all first choices horses were played the win percentage would be 31.8%, but because some horses at low odds are overbet the return is about $1.66 per $2.00 wager.

When a minimum 3/2 odds requirement was introduced, this reduced the number Betting Public first choices considered to 87 horses. Although the win percentage was slightly lower, there was a higher rate of return at $1.91 thanks to the better odds.

Table 1

The minimum 3/2 requirement was subsequently followed as a standard in the review of the remainder of data. Where no reference is made to odds, assume that the odds must be 3/2 or better.

CLASS − AER (Avg. Earnings Ratio)

The strongest single factor when combined with the first Betting Public preferences appears to be owning the top AER rating (hereafter "AER(100)"). Combining this finding with other factors is likely to be a good starting point to build an objective handicapping tool.

Table 2 — The AER

A little more than half of the races (78 of 151) had a top rated horse with an AER(100) among the top 3 Betting Public choices, and those horses won at a 33.3% rate in this sample with a solid average return at $2.68, even though no other factors or handicapping principles were applied. Dropping down to AER(80 to 99) is a smaller sized group at 59, again with no other handicapping, producing a 30.5% win rate with an average return at $2.23. It is to be expected that these results will improve if other criteria are added.

PERFORMANCE

In considering the relative value of the B2, LR and FTS performance ability ratings among the first three crowd choices, the ratings were ranked within the field of horses in each race. The top rated horse would have a ranking or "1", the second highest rated horse would have a ranking of "2", and so on.

From the results of this study, it appears the LR (last race final time) is not going to prove as productive as the B2 or FTS. Still this is good and useful knowledge. Many bettors rely on the final time for the last race as a primary indicator of a horse's ability and wager accordingly. The LR could only muster an average return of $1.84 and win rates below one in 4. Of interest is that the LR did better with a ranking of more than 4 when combined with being one of the top 3 betting choices. This suggests that when the reasons are not related to a strong last race, but the horse is still highly regarded, the horse tends to be underbet.

Both the B2 (best two races) and the FTS (our general and recent ability measures) outperformed the LR. The B2 seems to have a slight edge over FTS, although in practice there is little to choose between them. In terms of negatives, when the B2 ranking was more than 4 the results are quite poor ($1.33). This suggests requiring a B2 ranking of 4 or better may be useful.

In many races there will be more than one horse meeting the requirement for a good B2 or FTS ranking. They will often compete against each other, and water down the positive results to a degree. For that reason it may be appropriate to consider using the performance rankings as a supporting requirement when another factor points to a contender. It would appear useful to require either or both a B2 or FTS rating to be within the top 3 in the race for races with fields of 7 or less, and require it be within the top 4 for fields of 8 or larger.

Table 3 Performance (B2, LR and FTS rankings)

READINESS

There were two measurements for readiness which will be considered separately, and then together.

1. The DLR (days since last race)

Readiness can be measured in terms of the amount of time since a horse's last race. When the horse is fit it makes economic sense to race relatively soon after its last race, before its form declines. If it is unsound or injured there will usually be a relatively long period between races, as the stable rests and then gets the horse ready for its next racing campaign. Table 4 sets out our findings from the sample, with no other factors considered. Although the DLR is measured in days, it is also useful to think of this measurement in terms of weeks.

Table 4 — The DLR Score

The DLR results indicate that a horse returning to race within 14 days is a positive sign. Thereafter it appears to tail off. The results for layoffs of more than 3 weeks through 8 weeks runs well below the break even point of $2.00. After that the average return drops significantly. However, when the winners of those races were closely examined, it was found that there was also a higher TF score, on average of 12 or higher, and that may be the key to selecting horses returning from a long layoff.

B. The TF Score (total furlongs raced or worked in last 3 weeks)

Our study indicates that a TF score of 11 or higher is a positive indicator, and will usually ensure significant work within the last 21 days. If the horse has raced during that period, this means it will very likely also have had a workout. That is a positive sign. If it has been off from the races for longer, it will have worked an equivalent distance during the three weeks which is also a sign of its readiness, and its stable's positive intentions. There were no TF values of 1 or 2 in the sample.

Table 5 — The TF Score

This table suggests that TF values of 11 or more, as a single measurement, may be combined as a positive indicator with one of the three favorites. The values for TF of 10 are surprising low compared to 11 and higher, and there is no clear explanation, but it provides a convenient dividing line for our purposes.

Other than the TF value of zero, the other values did not do as well. The explanation for a positive result for TF(0) is twofold. Some trainers may not wish to have a recorded workout that might tip off the competition or the Betting Public prior to returning a horse that has been freshened. Also some zero scores occur with a DLR score close to the 21 day deadline, and a recent race may have been enough to keep them sharp. Where the DLR is "0" the horse's past performance record needs to be considered carefully to determine stable intentions.

C. Combining the TF and DLR

The next question to consider is how do these two readiness measurements work together? Taking a glance at Table 6 shows us some very interesting results. When a horse is returning within 28 days it appears there is little difference in whether a TF score is higher or lower. The average return for all is just under break even. Horses returning within 29 to 59 days don't seem to fare as well no matter what the TF scores are. However, there is a marked improvement for horses returning after 60 days if they have a TF score of 10 or higher (with an average return of $3.52 compared to $1.31 for the others).

Table 6 — The TF and DLR scores together

CLASS AND READINESS COMBINED

Looking closer at the top category of AER(100) horses, it was found that considering readiness to race appeared to be a significant factor within that group. Some of these top AER horses are ready to run, and some are being raced into shape with an eye to the next race. Horses returning in less than 3 to 4 weeks did exceptionally well. Those horses returning within 29 to 59 days did poorly. However those returning after more than 60 days performed better provided there was a high TF rating, averaging 12. In the latter group, the FTS proved to be the most influential of the ability/speed ratings. Consider the findings in Table 7.

Table 7 — AER(100) combined with DLR ranges

AER scores combined with TF(11+)

These findings suggest that a contender's readiness based upon the recency of its last race being within no more than about 24 days of the current race appears to be an important factor to take into account when the top AER(100) horse is being considered as one of the top 3 Betting Public preferences. It was noted that 8 of those horses returning within 24 days had a B2 rating of more than 4, and there was only one winner among them. This reinforces our earlier finding of the importance of a B2 ranking within the top 4.

The second portion of the Table indicates that top AER(100) types are less inclined to be fit simply because they have a good TF score, returning on average $1.84. For those it looks better to rely on the DLR. On the other hand, for contenders with AER in the 50 to 99 range, a good TF score of 11 or better is a good indicator with a $2.92 average return. Those with an AER of less than 50 did very poorly despite a strong TF score.

COMBINING PERFORMANCE AND READINESS

As shown in Table 8, combining the B2 or FTS performance-ability ratings and TF values can lead to positive results. There is little difference between the significance of the B2 and the FTS results. Requiring a performance standard of the B2 or FTS rating to within the top 3 or 4, and combined with the TF(11+), apparently tends to yields satisfactory result. The results for B2(≤3) were somewhat stronger in fields of 7 or less apparently, while the results for FTS(≤4) were superior in fields of 8 or more.

Table 8 — The B2 and FTS Rankings with the TF Score

How recent the last race was did not seem to work as well with the B2 or FTS ranking, as shown in bottom of Table 8. When combined with the strong DLR(≤21) the average return results only a little over the $2.00 standard.

COMBINING CLASS, PERFORMANCE AND READINESS

Having found some positive indications in Tables 3 and 7, is there any way to improve the results? Since the AER score did as well on its own, it is logical to consider combining it with the other positive factors. AER scores above and below 50 were evaluated in combination with the B2 (FTS can be substituted for B2) and TF. Where the top AER(100) horse is scratched, the next highest AER rating can simply be divided in half to derive the equivalent value of the AER(50) rating. For example, with a scratch of the AER(100) horse, if the next highest AER score is 88, then consider a contender with an AER of 44 or higher.

Consider Table 9. The results from our sample derived from combining the AER of 50 or better, the B2 ranked in the top 3 or 4, and TF score of 11 or better is quite positive.

Table 9 — Combining the AER with B2/FTS and TF

In the study there was virtually no difference in the results between B2(≤3) and B2(≤4), in terms of win percentage or average return, but there are a few more plays for the B2(≤4) group. When compared to the FTS ratings, the B2 rating performed slightly better, although, for those who wish to use the FTS as an alternative, it still did well with a 37.2% win rate, and an average return of over $3.00 in our sample.

The table also shows the very poor win percentage and average return where the AER is below 50. Where the TF score is ignored entirely there is still a positive return just combing AER(50+) with B2(≤3) despite this including more horses than there were races. Likewise, ignoring the B2, and combining just the AER(50+) with the TF(11+) produced an almost identical return, but with just 97 such horses in the 151 races. Combining the B2(≤4) with TF(11+) also yields a positive result.

Looking back at Table 4 you will recall that Betting Public preferred horses returning in 2 weeks or less did well generally. Table 10 shows the results when that is combined with a good B2, LR or FTS score and a requirement of an AER of 50 or better. This proves to be one area where the LR showed a positive result. A horse with a good last race should be returning quickly to race again. If the horse is held back for whatever, the LR loses its predictive value. The B2 still performed better than the LR in this group however, partly because of the better average mutuel price. The FTS lagged a little behind in this grouping.

Table 10 — Layoffs of less than than 2 weeks

Layoffs of more than 4 weeks.

Findings for layoffs of more than 4 weeks are set out in Table 11.

Table 11 — Layoffs of more than 4 weeks

The above findings included all horses, regardless of the length of time from their last race. Earlier it was noted that horses returning after more than 4 weeks from a race tended on average to do relatively poorly. For those results refer back to Table 4. Our study did find a group of horses that presented a winning profile for that category. This included having an AER of 50 or more, a TF of 11 or more and a FTS or B2 of 4 or less. The FTS outperformed the B2 by a small margin in that category.

These results show that the Class, Performance and Readiness considerations work well for horses regardless of the length of layoff, since the numbers for the over 4 week group actually outperforms the larger group including all horses regardless of layoff.

HANDICAPPING GUIDELINES

Based on the above findings from this study, a systematic process for qualifying favorites and near favorites can be developed using the Class, Performance and Readiness data from the Fast Track Simulcapping scores and ratings as a tool. Essentially, this data can be used to confirm or eliminate from among the top three contenders preferred by the Betting Public. The result will hopefully a more reliable favorite or near favorite. Alternatively, where all of the Betting Public top preferences are unacceptable, it may be possible to look for value elsewhere.

The following suggested approach draws upon the strongest findings among the Class, Performance and Readiness factors determined in our sample, combined with the wisdom of the Betting Public.

Positive Indicators

• AER score of 100 with DLR of less than 21 (or 24, or 28).

• AER of more than 50 with TF of 11 or better.

• DLR less than 14 days.

Negative Indicators

• AER of less than 50.

• B2 ranked worse than 4.

• FTS ranked worse than 4.

• DLR of 4 to 7 weeks.

Eliminations

The results from the study suggest three strong negative findings to be used in eliminating contenders.

1. Class and consistency: horses with an AER rating of less than 50 seldom win and yield a poor average return (16.0% and $1.40, Table 2).

2. Performance: horses with a B2 score ranked worse than fourth are unlikely to win and also yield a poor average return (16.4% and $1.33, in Table 3).

3. Horses returning to race within the range of 22 to 59 days of their last race also underperform ( see Table 6). Exceptions tend to be horses with TF(11+) and good performance rankings.

Prime CPR Selection

The logical foundation of our qualifying process is selecting the top AER(100) horse in the race where it combines its strong class and consistency with strong performance ratings and acceptable readiness criteria. Depending on how conservatively the criteria are set limits the number of possible selections, but hopefully increase the reliability and return. The following criteria produce about one selection for each race card in our sample.

The top AER(100) horse

• Must be one of the Betting Public top 3 choices.

• Must be offered at odds of at least 3/2.

• Must have the top AER in the field (which will be 100 unless that one is scratched).

• Must have had its last race within 28 days (better if less than 24 or 21 days) of today's race to confirm readiness.

• And its performance ability must also be confirmed by having its B2 ranking within the top 3 or 4, depending on whether the field size is 7 or less, or 8 or more.

Based upon the races in our 151 race study, taking these prime selection criteria into account would have resulted in 34 selections of which 18 won (52.9%) at an average price of $7.92, and an average return of $4.20 for a $2.00 win ticket.

For those inclined to a slightly more conservative approach it is recommended that the minimum odds required by raised to 2/1 rather than 3/2, and that the minimum DLR be set at 24 rather than 28. Based upon the races in our 151 race study, taking those alternate prime selection criteria into account would have resulted in 24 selections of which 14 won (58.3%) at an average price of $8.12, and an average return of $4.74 for a $2.00 win ticket.

Secondary CPR Selections

When there is no horse that qualifies under the Top AER(100) evaluation process, consider the following alternative or secondary selection criteria. These horses must meet the minimum AER standards, show strong performance ratings and demonstrate readiness with a strong TF score.

The secondary selection horse

• Must be one of the Betting Public top choices.

• Must be offered at odds of at least 3/2.

• Must have an AER of 50 to 99.

• Must have a TF score of 11 or more.

• And its performance ability must be confirmed by having both its B2 and FTS ratings ranked within the top 4 of the field.

These secondary CPR selection criteria would have produced 53 selections from our sample of 151 races. Of those 21 won (39.6%) at an average price of $8.09 for an average return of $3.21 per $2.00 win wager.

Morning Line and Multiple Race Wagering

The above findings will not necessarily translate to using the morning line ratings in place of the crowd selections. The morning lines are generally produced by only one person at each track. Their accuracy varies as a result. It is best to wait until close to race time and watch the tote board, either at the track, or online before making a selection. Making a selection otherwise will not actually conform to accurate use of the CPR handicapping tool. Many times it will be obvious that a horse is going to be a QF long before race time.

Final Comments

This study is an exploration of results derived from a set of positive findings for a specific sample of randomly selected races. It may or may not repeat in some other sample of races. It should not be blindly followed, but is intended only as a tool to assist in your handicapping. You are encouraged to observe how these CPR factors interact with the top 3 choices of the Betting Public at your favorite track or tracks. Please get back to us with any comments or findings.