Does Pace Matter?

At the end of 2010, I wrote up an article for the DPCD Sprint Championship.  I noticed that in a field of 11, there was only 1 pace-setter (Talkin’ Baseball), and only 2 more that would stay even remotely close to the pace.  This convinced me that the race would easily be won by the pace-setter, and that the stalkers would finish near the front as well.  I based this almost solely on the pace scenario of the race.  My belief?  That pace mattered in the sim.

I was right.

Talkin’ Baseball took home the win, and while one of the 7 closers did manage to get up for 2^nd (and I had predicted Pinchingdale 3^rd), the two other stalkers finished 3^rd and 4^th.

My prediction wasn’t solely based on pace.  Talkin’ Baseball was good enough to win the race, and even though there was no real quick pace, I thought that Pinchingdale was good enough to catch up, even if the other closers couldn’t.  I thought it was pretty clear that even a perfect pace scenario would not allow an inferior horse to overcome his competition.  And that proved to be the case, for the most part.  This all led me to wonder – How much does pace really matter in the sim?

It is not an easy question to answer, as there are many variables that come into play.  From the instructions: “A lot of factors go into determining the outcome of a race: pace, distance, surface, and post position are some of the race factors. Other things include the horse’s condition, his form, and how well it takes to the trainer.” (Emphasis added).  But I decided to take a look at the results based on pace.

——————–

Data:  Results from all Saturday races in 2010 were collected, with 77,376 races in total.  The data was further broken down based on 5 categories: Surface, Distance, Number of Pace-Setters, Start Speed of the pace-setter, and Running Style of the winner.

                 Surface: Races were split into Dirt Races (43,636) and Turf Races (33,740).

                Distance: Races were split into 3 Distance categories: Sprints (Less than 8f – 34,310), Classics (8f-10.5f – 35,452), and Longs (11f or longer – 7,614).

                Number of Pace-Setters:  In order to be classified as a “pace-setter,” a horse must be within 1 length of the lead at both the first and second call of the race.  Races are classified as either having 1 (46,221), 2 (22,189), 3 (6,737), or 4+ (2,229) pace-setters.

                Start Speed:  Start Speed is a number, between 0 and 100, which indicates how fast (how close to the lead) the jockey took the horse at the start of the race.  A 100 means that the jockey pushed the horse hard out of the gate and went for the lead (typically a front runner), and a 0 means that the jockey let the horse walk out and drop back quickly (typically a deep closer).  Stalkers are typically in the 70 range, Midpack is in the 50 range, and Closers are in the 20 range.  Theoretically, the lower the start speed of the pace-setter, the slower the pace is.  Races were split into 4 categories: Where the pace-setter in the race had a start speed of 100 (11,037), 90-99 (24,843), 80-89 (20,534), and 79 or less (20,962).

                Running Style: Races were finally divided into 5 categories, based on the horse’s running style in the race.  These categories were based on my Speed Study Series Presents: Track Bias article.  The 5 categories are: Front-Runner (no more than ½ length back at any point – 25,156), Stalker (½ – 2 ½ lengths back – 16,212), Midpack (2 ½ – 5 ½ lengths back – 18,847), Closer (5 ½ – 8 ½ lengths back – 10,848), and Deep Closer (more than 8 ½ lengths back at any point – 6,313).

Note that the totals provided above represent the portion of the whole for each category.  The data is further broken down so that any combination of any of the 5 categories is known.

———–

Results:

There are a lot of different numbers, and a lot of different ways of considering them.  In addition, some information, such as odds, were not included in this, meaning that some important information, that could sway the data, wasn’t included.  Furthermore, there are some unexpected results that require some additional explanation and investigation.  Plus, as I’m even farther removed from my last statistics class than I was with my last study (The Retired Mare Series), I remember how to do a proper statistical analysis even less.  So instead, I will attack the Pace Analysis in a different format:

Mythbusters Style.

For those who don’t know, the Mythbusters take a common belief and test it to see if it is actually true.  So that’s what I’ll be doing here.  I will take some beliefs that I had about pace in the sim, and use the 2010 data to “confirm” or “bust” that belief. So without further ado…

 

Myth #1: You can determine your horse’s best distance by using start speeds.

Plausible.    This myth is only tangentially related to pace, but it allows me to further explain what a Start Speed is.  This theory was posited, at least to me, by sim veteran barndog a number of years ago.  He found that the longer a horse ran, the higher the start speed would be.  If the horse hit a 100 start speed, Barndog believed (or at least I understood from him) that the horse was running too hard out of the gate, and that the distance was too far for the horse.

Essentially, the only way that this makes sense is if start speed is a “comparative” number.  Logically, a horse that runs an opening quarter-mile in 23-1/5 in a 6f race will not end up running a 21-1/5 quarter-mile in a 12f race.  But it could be that the 23-1/5 would be run in both.  So whereas that might translate into an 85 start speed in the 6f race, it might translate to a 100 start speed in the 12f race, as faster speeds in a longer race requires the horse to be pushed harder out of the gate.

So is it possible to determine the horse’s best distance by using start speeds?  Maybe.  The results of this analysis do not directly address this.  But it does show the percentage of races in which the pace-setter had a 100 start speed.  Out of 34,310 sprints, 1,065 had a pace-setter going at a 100 start speed (3.1%).  Out of 7,614 Long races, 3,411 had a pace-setter going at a 100 start speed (44.8%).  Classic races were 6,561 out of 35,452 (18.5%).

The differences here are significant.  Even ignoring the percentages, long races have 3 times more 100 start speeds as sprints, but have only 1/5 of the number of races.  This is about as clear a result as you could possibly ask for.  This data blatantly indicates that the longer the race, the more likely the pace-setter will be going “all-out” at the start.  Of course, this doesn’t really indicate what a horse’s best distance is.  It merely proves the first part of this, that start speeds likely increase as the races get longer.

So keeping how start speeds work in mind (where a 100 SP means the horse was going all out at the start, while a 70 start speed means the horse was holding back), let’s take a look at some other pace beliefs.

Myth #2: Races with more pace-setters cause the pace to be faster than races with less pace-setters.

Busted.  This myth seems to make sense in theory.  If you have 4 horses going for the lead in a race, each of them will be forced to push themselves harder in order to take the lead.  But the data does not back this up.

Of the 46,221 races that had only 1 pace-setter, 7,796 (16.9%) had a start speed of 100, and 10,691 (23.1%) had a start speed of 79 or less.  Meanwhile, of the 2,229 races with 4 or more horses near the leader, only 102 (4.6%) had the pace-setter with start speed of 100, but 927 (41.6%) had the pace-setter with a start speed of 79 or less.  The best way to understand this is to look at a chart showing how the percentage of races with the various start speeds changes as there are more horses “pushing” the pace.

As you can see from the graph, when there is one pace setter (the Blue line), the likelihood of the start speed being 79 or less is lowest, and the likelihood of the start speed being 100 is the highest.  As more horses press the pace, however, the starts speeds seem more likely to drop.  In other words, when there are 4 or more horses “pressing” the pace, it is far more likely that the pace-setter’s start speed is a 79 or lower.

What does this mean?  Well, it means that if there are 4 horses all within 1 length of the leader, then the leader is probably not running all out.  This doesn’t mesh with the myth that more front-runners cause a faster pace.  But if you think about it, the result does make sense.  If there are 4 horses all able to keep within 1 length of each other, odds are that the pace that is being set is fairly slow.  If one horse really wants the lead all to himself, he can go all out (the 100 start speed) and grab the lead, and he will probably be the only one within a length of the lead.

As for the myth, it’s busted based on the data above.  If there are more horses within 1 length of the leader, rather than causing the pace to be faster, it is more likely a sign that the pace is not all that fast.  Note that this does not mean that multiple front-runners do not push the pace to be fast.  This data does not consider previous running styles or the actual speed of the horses, since we do not have fractions (yet) in the sim.  The data merely shows that if there is a grouping in the front, odds are that the pace is slow.

Myth #3:  Closers are more likely to win long races than sprints.

Confirmed.  The theory behind this one is pretty simple.  The longer the race, the more of a chance a closer has of gaining the lead by the finish.  There is a competing argument to this one, however.  Some believe that horses who can sustain their speed over the long distance (likely one that’s at or near the front) will be better off than one who has one run in him (such as closers), because that one run will be less effective over the long distance (the horse will be more tired).  I have no idea which is truer in real life.  In fact, both seem to have their merits.  But in the sim, it’s pretty clear that closers are more likely to win long races than sprints.

There might be a second reason for this, however, and it goes back to start speeds and the definition of “closers.”  Taking the example from Myth 1 again, let’s say that a horse in a sprint runs an opening ¼ mile in 23-1/5.  Horses may stay relatively close to this pace, and the pace may not be quick enough for closers to get up for the win.  But now imagine that the same pace-setter runs in a long race.  What would have been an 80 start speed before is now a 100, as the horse again gets the opening ¼ mile in 23-1/5.  Now, since the race is longer, the rest of the horses relax more, thus dropping further back.  On top of that, the pace-setter tires out quickly, and easily gets passed by the rest of the field.  This leads to more closers winning races.

Maybe this seems obvious, but it’s worth looking at the numbers anyway (if for no other reason than I have them).  Of the 34,310 sprint races, 4,428 of them (12.9%) were won by either Closers or Deep Closers.  Meanwhile, in 7,614 long races, 2,766 (36.3%) were won by Closers or Deep Closers.  The classic races are in between, with 9,967 of 35,452 races (28.1%) being won by either Closers or Deep Closers.  To further break the numbers down by surface as well, take a look at the chart below.

The light red/green shows that closers’ winning percentages rise as the distance of the race rises.  It’s interesting to look at this and note that on the dirt, front-runners are slightly more likely to win than on the turf, and closers are slightly more likely to win on the turf.  But clearly this graph shows an advantage for closers in longer races, to the point that closers are actually more likely to win long turf races than pace-setters are.

Now with that, let’s look at the 2 myths that directly address pace and winning races.

Myth #4: The more horses at the front of the pack, the less likely the pace-setter wins/more likely a closer wins.

Busted.  This one goes back to what we established in Myth #2.  It turns out that based on the data, the more horses there are at the front of the pack, the slower the pace probably is.  And if the pace is slower, it is actually less likely that a closer will come on to win.

This chart shows the percentages that each running style wins a race.  As the number of pace-setters increases, the chances almost universally drop.  Clearly, you can see that whatever chances are diminished from each of the other four groups are picked up by stalkers, who win races with 4 or more pace-setters at a 39.79% clip.  Front runners seem to be fairly consistent throughout all four types (32.70% in 1 pace-setter races, 31.18% in 4+ pace-setter races).  There is also no discernable difference between surfaces, but there is between distances.  Sprint and Classic races are fairly similar to this breakdown (Sprints have a higher Stalker win % in 1 pace-setter races, and Classics see a greater drop-off of Front Runners in 4+ pace-setter races), but the Long distance races have a very different appearance.

The first interesting thing to note in this graph is that Deep Closers are actually the 2^nd most likely winners in 1 pace-setter races (22.32%).  This graph also seems to indicate that the more pace-setters there are in long races, the more wide-open the race is, as every style has between a 14.29% chance (Closers) and a 23.81% chance (both Stalkers and Midpack) of winning the race.  But this result is mostly due to sample size.  There are only 21 races out of the 77,376 in 2010 (0.027%) that are long races and have 4 or more pace-setters.  The trend from 1-3 pace-setters, however, seems to indicate the same thing as the other graphs – Stalkers are more likely to win at 4+ pace-setters, and everyone else drops.

But with this data, it’s pretty clear that an increased number of pace-setters does not set things up for closers, or limit front runners in any way, which makes this Myth busted.  But that’s not the last thing to test, because as we saw from earlier evidence, number of pace-setters could be the exact opposite of what you’d expect in terms of pace.  Instead, the better question is how start speed relates to winning running styles.

Myth #5:  The higher the start speed of the pace setter, the less likely a front runner wins/more likely a closer wins.

Confirmed.  Based on the data at hand, it seems that start speed is actually the best way to determine pace (especially since, as we discovered in Myths #2 and 4, number of pace-setters seems to be inversely related to pace).  And that’s why this Myth is probably the most important one to look at.  When the pace-setter runs faster out of the gate, does it set up the race for closers? 

The chart above is fairly dramatic in confirming this myth.  As the start speed got higher (meaning as the pace got faster), front-runners and stalkers were less likely to win, and closers and deep closers were more likely to win.  When running at a start speed of 79 or less, pace-setters had a 44.12% chance of winning, and deep closers only had a 1.84% chance, out of 20,962 races.  This is pretty convincing that if you have a deep closer, you want to make sure that there’s someone in the race that can push the pace.  Meanwhile, if the pace-setter is going fast to the point where he’s at a 100 start speed (11,037 races), deep closers have a 19.81% chance of winning (over a 1000% increase!) while front runners drop to a 23.15% chance of winning (nearly a 50% decrease).  The only running style that seems relatively unaffected is Mid-pack horses, who maintain a 21-27% winning percentage regardless of the start speed.

 

Here is an (admittedly compressed) look at the three different types of distances that come together to form the main chart above.  The Dirt/Turf charts, meanwhile, look almost identical to the main chart, so I won’t show those.  The moral of these charts, as well as the main one above, is that start speed has a strong correlation with the winning running style.

Now let’s take a look at this another way, when combining some of the stats. 

These two charts show the differences for Front Runners and Deep Closers based on both number of pace-setters and start speed of the pace-setter.  You can see that a front runner’s chances of winning are fairly even regardless of how many pace-setters there are, but the likelihood of winning drops as start speeds rise.  For deep closers, meanwhile, the likelihood of winning rises as start speed rises, but unlike front runners, there seems to be a more pronounced decline in winning percentage as the number of pace-setters rises.

The closer chart resembles the Deep Closer chart (as should be expected), though the decline by number of pace-setters is slightly more gradual.  You can also see that the separation based on start speed is smaller for Closers than for Deep Closers.  Midpack horses, meanwhile, seem to be the least affected by the pace.  There is no common trend in how midpack horses do with different numbers of pace-setters, and the difference in winning percentage based on start speed appears fairly small.

I wanted to single out the stalkers here, because the graph is so different from any of the others.  This is the one example where the more pace-setters there are, the more likely a horse is to win.  There also does not seem to be a clear trend between start speeds.  With less pace-setters, the slower the start speed, the higher the chance of winning.  But as you get to 3-4+ pace-setters, it no longer seems to matter as to what the start speed is.  The 100 start speed, 4+ pace-setters mark is most likely a result of small sample size, as there are only 102 races that fit this classification. 

——————

So, what is the takeaway from this brief look into pace scenarios?  First of all, start speeds, and not number of horses at the front of the field, is probably the best way we currently have to determine the actual pace of the race.  But more importantly, the data above shows that pace does, in fact, seem to make a difference in the race.  If the pace is slow (a start speed of 79 or less), front runners have a significant advantage (44.12%) and deep closers are at a severe disadvantage (1.84%).  However, if the pace is very fast (start speed of 100), deep closers have their best possible chance of winning (19.81%) and front runners have their worst chance (23.15%).  We also learned that closers are more likely to win long races in the sim than sprints.  There also seems to be a slight advantage for closers in turf races compared to closers in dirt races, though the difference is fairly small.

Of course, there are some limitations to this study.  As I mentioned earlier, information such as odds were not considered in this.  A slow pace would probably not prevent a 1/5 closer from winning a race.  Another thing that was not considered was the track.  5 years ago, I did a study on Track Bias to look at how different tracks played for the various running styles, and the data showed that there was at least some amount of track bias in the sim.  That could definitely play a role in how pace affects a race.  Unfortunately, that information is a lot more difficult to reconcile with this set of information.  Hopefully I will revisit the track bias study somewhere down the line, and then I will be able to combine both pace and track bias, but until that time, we will just have to assume that all else is equal.

Still, this is an interesting look at the effect of pace in the sim, and it strengthens the belief that we should know the fractions of a race after they’re run.  It also stresses the importance of start speeds in best placing your horses.  For those of you who don’t know where to find start speeds, they are provided by the Super Sim Searcher (SSS) or from Yoss’s website.  So for people who do a lot of research into the fields they are entering before they enter, keep this pace issue in mind.

———–

This was a brief look at the impact of pace in the sim, and I did not use all of the data that I have.  I also only tackled 5 myths.  You may have questions about pace that were not answered here.  If you do, feel free to send me a PM on the message board, or drop me an e-mail at Bryan.Arsham@gmail.com.  I’m happy to tackle other pace myths, assuming I have data on them, and I can also provide the 2010 pace data, broken down as explained earlier, for anyone who wants it.

Hopefully you all enjoyed this look at how pace affects races in the sim.