The Kentucky Derby is such a unique race. Three-year olds at a mile and a quarter at a time of year when no three-year old has been permitted to go a mile and a quarter. The handicapping of this race, like any other, starts with questions about what it takes to win. The difference is that our 89,000 race database doesn't contain too many "races like this one" to answer those questions.

We start by building a query asking for three-year old graded stakes for males at 1 1/8 miles or longer before May 1st of each year.When this query was run against our database, we found 23 such races. To that we added the last 3 Kentucky Derbies bringing the total to 26 races.

Next, we move to the WINNING PROFILES section of The Handicapper's Notebook...

... and press "C" for CONTENDER.

We have sorted the factors based upon their ability to get the most winners in the top 4 ranking positions in their respective fields inlcuding ties. The best factor was "APV," which stands for "Average Purse Value." This factor reflects the levels at which the animal has been racing. Thus, in the 26 races, the top 4-ranked APV horses in each field have combined for 25 of the 26 wins! (Notice that there were 121 starts in the 26 races, which indicates 17 horses were tied for fourth.) In addition, the $2.89 $net indicates that for every $2.00 bet to win on these horses, you would have received back $2.89, a profit of 44.5%!

The second contender factor is our "CLASS" factor ("Clss" above), which had 24 of the 26 winners. CLASS, another earnings-box based rating, takes into account APV, earnings-per-start, and consistency.

The third factor, "sSC09," stands for "speed rating, stretch-call, average of the last 3." In other words, we have created a stretch call speed rating for each horse and the best way to look at these ratings is to average the last 3 such ratings. 23 of the 26 winners had this factor in common.

Notice how many factors produced 23 of the 26 winners. As you study the names of those nine factors in the left-hand column, notice how many are "SC" (stretch call factors). This shows that positioning at the stretch call has a tremendous impact on who wins the race. Bear in mind that the "stretch call" is not at the top of the turn as they head for home. Rather, it is always one-furlong from the wire.

Here is a code list to help you understand the "s" (speed rating factors).

F1   First Fraction (rating to the 1st call) 01   last race EP   Early Pace (rating to the 2nd call)   02   best of last 2 races SC   Stretch Call (rating to the 3rd call)   03   best of last 3 FT   Final Time (entire race rating)   04   best of last 4 SR   Stretch Run (from 2nd call to wire)   05   best 2 of last 3 SP   Sustained Pace (average of EP and SR)   06   best 2 of last 4 FW   Factor W (EP + SR, weighted strongly towards EP)   07   best 3 of last 4 Pw   Power (favors horses that can run good final times without fast EP)   08   average of last 2 LP   Late Pace (1st call to wire; i.e. ignores 1st fraction)   09   average of last 3         10   average of last 4         11   best ever         12   2nd best ever         13   best 2 ever

Zooming in to the previous screen shot, we see that Final Time ("FT") also plays an important part. In looking at the two-digit codes (01 thru 13) notice what is represented.

Perhaps you may deduce something else, but here is what I get from this:

1. The animal must have no more than one mediocre speed rating in its last 3 or 4 races (05's and 07's indicate this).
2. Have a high average over the last 3 races (09's indicate this)
3. The horse's must have a competitive 2nd best speed rating (12's indicate this).
4. The last race is not that important (the lack of 01's indicate this). The last race is often a prep race.

Let's summarize what we know about this race thus far...

Note that the APV and Clss factors both carry with them a 10% margin for error. In other words, any horse within 10% of the 4th-ranked horse qualifies as a contender.

Before betting the farm on Indian Charlie, let me point out a few things. Going 8 factors deep out of a 26 race sample is excessive. It results in what we call a "retrofit." That is, we are no longer predicting the future, but rather imitating the past. Let's be pleased that we can get the field down to four or even three horses (Favorite Trick might be considered marginal) at this point.

Second, these selections are made before final scratches. Scratches could alter the selection results and, with 23 entries, there will likely be some scratches.

Let's revisit our contender process once again, this time sorting by win% in the top 3 rankings...

This differs from last time in two ways. We are now only going 3 ranks deep instead of 4 and we have sorted not on "total winners" but on "win percentage" instead. The top four factors now contain a "Factor W" (an early pace factor) as well as two final time factors. Let's consider the impact these would have on our selection process...

Halory Hunter and Indian Charlie come up again. Confirmation is a good thing.

Let's do a "top 2" query and check again...

"Earn" references simply the total money earned in the last two calendar years. Notice that we now have F1's (First Fraction) represented. Can you see that as we become more restrictive in the rankings "Early" parts of the race (EP and 1st fraction) become more important? They are serving as tie-breakers.

Permit me to present one last set of tables then we shall summarize the summaries, so to speak.

This time we have asked the question, "What factors produce the highest win% in the top ranked position?" Suddenly, the importance of the positioning in the first fraction becomes apparent! This is a long race, with a lot of horses. A horse that doesn't have enough tactical speed to get a good position early, could lose the edge needed in that last furlong of the race. In other words, he may be a contender to the stretch call, but without good first fraction ability, he will be worn out in the stretch drive.

Notice that the big F1 horse has not even been mentioned as a contender before! Here is the summary...

To make things easier, I have shown the rankings rather than the raw ratings.

I believe that a horse should be a contender all the way down to rank 2 (4-3-2). However, when we have contradictions as we do here a different approach may be called for.

Let's try to ask the right questions here.

"Of the top 4 ranking contenders, who is the most likely winner?"
4-Halory Hunter, Cape Town, Indian Charlie
3-Halory Hunter, Indian Charlie
2-none, but edge to Indian Charlie

"Of the top 3 ranking contenders, who is the most likely winner?"
3-Halory Hunter, Indian Charlie, Event of the Year
2-none, but edge to Indian Charlie

"Of the top 2 ranking contenders, who is the most likely winner?"
2-Favorite Trick, Cape Town
1-none, but edge to Cape Town

"Of the top 1 ranking contenders, who is the most likely winner?"
1-Yarrow Brae

Let's look at a few other characteristics before closing our discussion and moving on to the tote board.

A recent workout is very important. Horses that have not had a workout or a race in the last 7 days have combined for 2 wins in 43 starts. However, before counting them out, consider that the number of "pool expected winners" ("PExW"; based upon the way the public bet the money) is only 3.4. The sample is only 1.4 winners shy of where it should be. Before we consider ruling anyone out on this basis, consider looking at the trainer stats for the individual horses.

What if we asked the question, "How many bad races has the horse had since its last good race?" A "good race" is a race in which the horse finished 1st, 2nd or 3rd, or within 2 lengths (3 in a route race) of the winner." Thus, the horses in the "0" category (above) ran a good race in their last start. Not much to recommend here. If there was, it would be to prefer a horse with a bad (not good) last race!

Number of good races in last three starts does very little for us.

Horses that have won three times in their last 6 starts are to be preferred over their rivals, especially those that have won less than twice.

The table above addresses the difference between today's distance and the "Average Competitive Distance" (ACD) of the horse. Not that a horse in the 1.01 and above category may be either stetching out or shortening up by a furlong or more. Of course in the case of the Derby, no horse will be shortening up.

The point to this table is that the distance a horse is "wrong" by does not matter. A horse whose ACD was 9.5 furlongs (only 1/2 furlong off of today's) has historically shown to be at a disadvantage to a horse whose ACD was only 8.5 furlongs! Distance is obviously a factor that can be ignored.

Followers of the Quirin Early Speed Points, (8-point horses almost always go to the early lead; 0-point horses almost never do) will see that horses with less than 1 point do very poorly and can probably be eliminated. In today's race the only horse that fits that criterion is Hanuman Highway.

When we look at the tote board in terms of public choice rankings, we find that favorites (including co-favorites) won only 19.4% of their races and returned a dismal 78 cents for each $2 wagered. In other words, they lost their backers 61% of their investment! This, of course, fits with the Derby pattern, since no favorite has won since 1980.

Interestingly, horses below the top 5 choices won only 2 of the 26 races (131 starts). In other words, the public is consistently wrong in picking its favorite, but just isn't that far wrong in terms of its top choices.

When we look at the actual odds themselves rather than the rankings, certain odds ranges clearly have peformed better than others. For example, in win betting it is easy to see that the range from 5-2 to 7-1 has done very well by looking at the $net. If you total them up you would find:

Does this mean that one shouldn't bet horses that aren't in these ranges? Certainly not. It simply means that one should consider structuring tickets centering around the better bets.Linking up a bunch of 20-1 shots in a trifecta reminds me of an old Warren Nelson (owner of the Cal-Neva in Reno) quote: When asked what he thought of card counters, he replied, "When the lamb goes to the butcher, the lamb might kill the butcher, but we like to bet on the butcher."

Our Quinella Analysis screen says a lot about the potential for exactas. The legend, in the upper right corner, shows we have broken the horses into four odds groups. The first piece of notable information can be found in the lower left corner. Combinations with the lowest odds horse being a "C" (i.e. 10-1 to 24-1) combined with another "C" or a "D" produced only one of the 26 quinellas. And not for lack of trying. There were a combined total of 367 "ways" on all those quinella tickets!

So, if that group won only a single race, then the "others" (i.e. those that contained an "A" or "B") went 25-for-26. According to the above table, the A-B's are the best bet, showing a $2.04 $net.

When you structure your exacta and trifecta tickets in this race consider that the winner will likely be in the top 5 for public choice and the most profitable of those will fall into the range of 5-2 to 7-1 (from our Odds Analysis earlier). The most profitable place and show wagers will be from the 5-1 to 9-1 range. Try to organize tickets that key the horses you like with other horses in these ranges.

Have a good time at the races.