‡
Big Event Race Days and Dubious Statistics
I made this off-hand comment on a recent comment thread that included Mike @RepoleStable :
I would also be willing to bet that the Derby-related success that he has lauded is at least to some extent smoke and mirrors.
To which he responded:
You’re an idiot!!!! 😂
Explain the smoke and mirrors. The years of TV viewership growth, the revenue growth from Derby week, and on and on.
The Derby had 24 million viewers this year.
He supported his assertion primarily with the attached all-source handle chart, which, superficially, shows flat growth for the Preakness, Belmont, and Breeders' Cup Saturday cards, and in contrast, impressive looking growth for the KY Derby. In 2007 the all-source KY Derby day handle was $168m, while this year it was $333m, an apparent 97% increase.
The problem, which I'm not surprised that Mike failed to take into account, is that the figures used to construct the chart were presumably measured in nominal dollars. And when adjusted for inflation, the $168m was, in terms of purchasing power, equivalent to $273m of today's (degraded) dollars. That translates to a more apples-to-apples increase of 22%, rather than 98%, which is not only much less impressive, but reveals what is essentially a 345% overstatement of the growth rate (relative percentage error) from 2007-26, based on nominal dollars.
2007 Derby day
nominal dollars → $168m
inflation-adjusted → $273m
2026 Derby day
→ $333m
▶️ Comparison
nominal dollars → +$165m/+98%
inflation-adjusted → +$60m/+22%
Reasonable debates can be had about the value of a 22% rise over that period of time, but to use figures based on nominal dollars in historical comparisons is dubious at best, and the longer the timeline, the more misleading they tend to become. Having said that, the comparisons with the other three events on the chart do flatter the KY Derby day results, but it is important to note that the contrast is also reflective of relatively poor performance by the Belmont, Preakness and BC.
Beyond that, and as Mike really should know, comparisons between Big Event Race Days (BERD henceforth) over time, even when calculated as accurately and honestly as possible, do not necessarily tell the whole story. Why? These, I would say, are three primary reasons:
➡️ wagering handle for such days does not rise in a vacuum, as at least some of the gains will almost invariably come at the expense of other race days – i.e. cannibalization (think about the Met Mile having been moved to the Belmont day, as an example)
➡️ as the focus of major racetracks on BERD increases, wagering handle gains (perceived or real) may have the effect of masking declines in handle for the everyday, bread-and-butter racing cards, further feeding the cycle on dubious grounds
➡️ the increasing emphasis on BERD may have the effect of essentially conditioning fans and bettors to care less about blue collar racing cards, which in turn feeds associated negative feedback loops that have already damaged the industry (e.g. the increasing dominance of "super trainers", and corresponding decline of "middle-class" stables; decreasing foal crops and stallion options, etc.)
***
Back to the numbers, and for further context, let's take a look at some broader data points based on the annual all-source handle for Churchill Downs over the 20 year period from 2006-25:
2006 annual
nominal dollars → $650m (excluding the Breeders' Cup)
inflation-adjusted → $1b
2025 annual
→ $1b
▶️ Comparison
nominal dollars → +$210m/+26%
inflation-adjusted → no change
So in nominal dollars, there was an increase of 26% over those 20 years. But given a more honest, inflation-adjusted accounting of Churchill's full-year 2025 performance vs. 2006, the numbers were actually flat. Which begs the question of the extent to which the health of wagering on the Churchill Downs racing product has actually improved with the increasing BERD focus, if at all.
To be fair, let's also take a look at the 10 year period from 2016-25, during which much of the growth on the embedded Derby day chart is indicated:
2016 annual
nominal dollars → $529m
inflation-adjusted → $709m
2025 annual
→ $1b
▶️ Comparison
nominal dollars → +$471m/+89%
inflation-adjusted → +$291m/+41%
2016 Derby day
nominal dollars → $192m
inflation-adjusted → $270m
2025 Derby day
→ $349m
▶️ Comparison
nominal dollars → +$157m/+81%
inflation-adjusted → +$79m/+29%
***
The figures from the above subset (2016-25) tell an interesting story. They are, in one fundamental respect, very similar to those of the longer timeline, but at the same time they tell a very different story.
They are fundamentally similar as once again the more accurate inflation-adjusted figures cut through the smoke and mirrors of those based on nominal dollars:
• the 41% increase of the annual handle, rather than 89%, reveals a 117% overstatement of the growth rate based on nominal dollars
• the 29% annual increase of Derby day figures, rather than 81%, reveals a 179% overstatement of the growth rate based on nominal dollars
All of the adjustments made to the figures over both timelines demonstrate that the impressive looking growth trajectory of the KY Derby day handle illustrated in the embedded chart is misleading, though certainly less so for its growth relative to the Belmont, Preakness and BC.
It is important to note that there is also a variable that appears to have made a particularly strong impact on Churchill Downs more recent annual and Derby day wagering handle increases.
While the adjusted figures show that from 2006-25, the annual wagering handle remained flat, it rose 41% from 2016-25. What happened? Well, Churchill Downs began receiving gaming revenues from historical horse racing machines (HHR) in 2018, and its purses began to rise rapidly in 2019. In fact, the purses for its 2019 Spring meet rose an unprecedented 46%. The KY Derby purse rose that year from $2m to $3m; the purse for The Old Forester Turf Classic doubled, from $500k to $1m. Etc.
That massive infusion of "gaming" money was clearly a major driver of the relatively sharp rise of both CD's annual, and KY Derby day handle from 2019-26.
Now, here's where it becomes tricky. Some might argue that the figures, whether inflation-adjusted or not, should be seen through the lens of the downward trajectory of the overall American wagering handle over those times periods. After all, from 2006-25, overall U.S. handle fell by roughly $3.7b in nominal dollars (-26%), and $12b (-85%) when Inflation-adjusted.
So it certainly is reasonable to question how the Churchill Downs numbers above should be viewed in light of those broad declines. The tricky part is that it would be naïve to assume that the numbers should necessarily be thought of as being particularly impressive because they occurred during a period of broader handle declines.
The most obvious reason for that has already been mentioned: the infusion of gaming money that catalyzed sharp purse increases, which would in turn be expected to have attracted more and better quality horses, larger fields, and greater resulting handle.
One of the less obvious reasons is that while the industry has undoubtedly been contracting, as evidenced pungently by the overall handle decline, the effects of the contraction are of course not evenly distributed. Since 2000, it appears that 41 tracks have permanently closed in the U.S., and it should go without saying that the aggregate handle produced by those tracks did not all migrate to other active tracks. And despite the broad contraction, some of the major tracks, and especially those benefitting from gaming subsidies, like CD, were able to increase their shares of the diminishing pie, helping to offset at least some of the fallout from the broader declines.
***
In conclusion, I would say that while Churchill Downs handle has improved over recent years, the above analysis demonstrates that the use of nominal dollars to make historical comparisons is dubious, and can result in misleading claims. It is also no coincidence that charts such as the one attached tend, far more often than not, to mislead in favor of those who produce and utilize them, as they may have incentives to exaggerate performance.
The analysis also raises a number of related questions, including, but certainly not limited to these:
• To what extent has Churchill Downs' increasing emphasis on Big Event Race Days actually improved its overall performance based on wagering handle?
• To what extent are the improved recent handle numbers a reflection of that business model, and to what extent are the slot-fueled purses indirectly responsible?
• What is the significance of the fact that the annual CD handle from 2016-25 rose 41%, while the Derby day handle rose only 29%? Could it be evidence that slot-fueled purses are actually the primary drivers of the recent gains, rather than increased focus on BERD?
• If Belmont Day and Breeders' Cup Saturday are essentially flat over the past 10 years, as suggested in the chart, how far down are they in inflation adjusted dollars over that period? Should NYRA and the BC be worried, and what would notable declines in handle, as could be revealed by more honest accounting, indicate about NYRA's increasing emphasis of BERD?
Questions raised by the attached chart, and above analysis, are potentially very useful, and not only in relation to the performance of Churchill Downs. The increasing emphasis on Big Event Race Days has thus far been met by mixed reactions from fans and bettors around the country, and even if CD has been able to show growth as the experiment has unfolded over recent years, the fact that neither Belmont day nor Saturday Breeders' Cup cards have shown growth even in nominal dollar terms, raises serious questions about the model.
Big Event Race Days and Dubious Statistics
I made this off-hand comment on a recent comment thread that included Mike @RepoleStable :
I would also be willing to bet that the Derby-related success that he has lauded is at least to some extent smoke and mirrors.
To which he responded:
You’re an idiot!!!! 😂
Explain the smoke and mirrors. The years of TV viewership growth, the revenue growth from Derby week, and on and on.
The Derby had 24 million viewers this year.
He supported his assertion primarily with the attached all-source handle chart, which, superficially, shows flat growth for the Preakness, Belmont, and Breeders' Cup Saturday cards, and in contrast, impressive looking growth for the KY Derby. In 2007 the all-source KY Derby day handle was $168m, while this year it was $333m, an apparent 97% increase.
The problem, which I'm not surprised that Mike failed to take into account, is that the figures used to construct the chart were presumably measured in nominal dollars. And when adjusted for inflation, the $168m was, in terms of purchasing power, equivalent to $273m of today's (degraded) dollars. That translates to a more apples-to-apples increase of 22%, rather than 98%, which is not only much less impressive, but reveals what is essentially a 345% overstatement of the growth rate (relative percentage error) from 2007-26, based on nominal dollars.
2007 Derby day
nominal dollars → $168m
inflation-adjusted → $273m
2026 Derby day
→ $333m
▶️ Comparison
nominal dollars → +$165m/+98%
inflation-adjusted → +$60m/+22%
Reasonable debates can be had about the value of a 22% rise over that period of time, but to use figures based on nominal dollars in historical comparisons is dubious at best, and the longer the timeline, the more misleading they tend to become. Having said that, the comparisons with the other three events on the chart do flatter the KY Derby day results, but it is important to note that the contrast is also reflective of relatively poor performance by the Belmont, Preakness and BC.
Beyond that, and as Mike really should know, comparisons between Big Event Race Days (BERD henceforth) over time, even when calculated as accurately and honestly as possible, do not necessarily tell the whole story. Why? These, I would say, are three primary reasons:
➡️ wagering handle for such days does not rise in a vacuum, as at least some of the gains will almost invariably come at the expense of other race days – i.e. cannibalization (think about the Met Mile having been moved to the Belmont day, as an example)
➡️ as the focus of major racetracks on BERD increases, wagering handle gains (perceived or real) may have the effect of masking declines in handle for the everyday, bread-and-butter racing cards, further feeding the cycle on dubious grounds
➡️ the increasing emphasis on BERD may have the effect of essentially conditioning fans and bettors to care less about blue collar racing cards, which in turn feeds associated negative feedback loops that have already damaged the industry (e.g. the increasing dominance of "super trainers", and corresponding decline of "middle-class" stables; decreasing foal crops and stallion options, etc.)
***
Back to the numbers, and for further context, let's take a look at some broader data points based on the annual all-source handle for Churchill Downs over the 20 year period from 2006-25:
2006 annual
nominal dollars → $650m (excluding the Breeders' Cup)
inflation-adjusted → $1b
2025 annual
→ $1b
▶️ Comparison
nominal dollars → +$210m/+26%
inflation-adjusted → no change
So in nominal dollars, there was an increase of 26% over those 20 years. But given a more honest, inflation-adjusted accounting of Churchill's full-year 2025 performance vs. 2006, the numbers were actually flat. Which begs the question of the extent to which the health of wagering on the Churchill Downs racing product has actually improved with the increasing BERD focus, if at all.
To be fair, let's also take a look at the 10 year period from 2016-25, during which much of the growth on the embedded Derby day chart is indicated:
2016 annual
nominal dollars → $529m
inflation-adjusted → $709m
2025 annual
→ $1b
▶️ Comparison
nominal dollars → +$471m/+89%
inflation-adjusted → +$291m/+41%
2016 Derby day
nominal dollars → $192m
inflation-adjusted → $270m
2025 Derby day
→ $349m
▶️ Comparison
nominal dollars → +$157m/+81%
inflation-adjusted → +$79m/+29%
***
The figures from the above subset (2016-25) tell an interesting story. They are, in one fundamental respect, very similar to those of the longer timeline, but at the same time they tell a very different story.
They are fundamentally similar as once again the more accurate inflation-adjusted figures cut through the smoke and mirrors of those based on nominal dollars:
• the 41% increase of the annual handle, rather than 89%, reveals a 117% overstatement of the growth rate based on nominal dollars
• the 29% annual increase of Derby day figures, rather than 81%, reveals a 179% overstatement of the growth rate based on nominal dollars
All of the adjustments made to the figures over both timelines demonstrate that the impressive looking growth trajectory of the KY Derby day handle illustrated in the embedded chart is misleading, though certainly less so for its growth relative to the Belmont, Preakness and BC.
It is important to note that there is also a variable that appears to have made a particularly strong impact on Churchill Downs more recent annual and Derby day wagering handle increases.
While the adjusted figures show that from 2006-25, the annual wagering handle remained flat, it rose 41% from 2016-25. What happened? Well, Churchill Downs began receiving gaming revenues from historical horse racing machines (HHR) in 2018, and its purses began to rise rapidly in 2019. In fact, the purses for its 2019 Spring meet rose an unprecedented 46%. The KY Derby purse rose that year from $2m to $3m; the purse for The Old Forester Turf Classic doubled, from $500k to $1m. Etc.
That massive infusion of "gaming" money was clearly a major driver of the relatively sharp rise of both CD's annual, and KY Derby day handle from 2019-26.
Now, here's where it becomes tricky. Some might argue that the figures, whether inflation-adjusted or not, should be seen through the lens of the downward trajectory of the overall American wagering handle over those times periods. After all, from 2006-25, overall U.S. handle fell by roughly $3.7b in nominal dollars (-26%), and $12b (-85%) when Inflation-adjusted.
So it certainly is reasonable to question how the Churchill Downs numbers above should be viewed in light of those broad declines. The tricky part is that it would be naïve to assume that the numbers should necessarily be thought of as being particularly impressive because they occurred during a period of broader handle declines.
The most obvious reason for that has already been mentioned: the infusion of gaming money that catalyzed sharp purse increases, which would in turn be expected to have attracted more and better quality horses, larger fields, and greater resulting handle.
One of the less obvious reasons is that while the industry has undoubtedly been contracting, as evidenced pungently by the overall handle decline, the effects of the contraction are of course not evenly distributed. Since 2000, it appears that 41 tracks have permanently closed in the U.S., and it should go without saying that the aggregate handle produced by those tracks did not all migrate to other active tracks. And despite the broad contraction, some of the major tracks, and especially those benefitting from gaming subsidies, like CD, were able to increase their shares of the diminishing pie, helping to offset at least some of the fallout from the broader declines.
***
In conclusion, I would say that while Churchill Downs handle has improved over recent years, the above analysis demonstrates that the use of nominal dollars to make historical comparisons is dubious, and can result in misleading claims. It is also no coincidence that charts such as the one attached tend, far more often than not, to mislead in favor of those who produce and utilize them, as they may have incentives to exaggerate performance.
The analysis also raises a number of related questions, including, but certainly not limited to these:
• To what extent has Churchill Downs' increasing emphasis on Big Event Race Days actually improved its overall performance based on wagering handle?
• To what extent are the improved recent handle numbers a reflection of that business model, and to what extent are the slot-fueled purses indirectly responsible?
• What is the significance of the fact that the annual CD handle from 2016-25 rose 41%, while the Derby day handle rose only 29%? Could it be evidence that slot-fueled purses are actually the primary drivers of the recent gains, rather than increased focus on BERD?
• If Belmont Day and Breeders' Cup Saturday are essentially flat over the past 10 years, as suggested in the chart, how far down are they in inflation adjusted dollars over that period? Should NYRA and the BC be worried, and what would notable declines in handle, as could be revealed by more honest accounting, indicate about NYRA's increasing emphasis of BERD?
Questions raised by the attached chart, and above analysis, are potentially very useful, and not only in relation to the performance of Churchill Downs. The increasing emphasis on Big Event Race Days has thus far been met by mixed reactions from fans and bettors around the country, and even if CD has been able to show growth as the experiment has unfolded over recent years, the fact that neither Belmont day nor Saturday Breeders' Cup cards have shown growth even in nominal dollar terms, raises serious questions about the model.