Thursday, August 10, 2017

If Racing Wants to Make Big Wagering Decisions, At Least Study the Numbers

For a long while, we've often wrung our hands at how the business of horse racing (publicly at least) studies their facts and figures when it comes to wagering.

What should be a deep dive - something that looks at things with a microscope with statistical significance attached - is often fishing in a pond; a pond stocked with whatever fish you're looking for to boost your argument.

I read this line in a few places, regarding the Keeneland takeout increase, and it's been passed around some as a justification for raising rates.

"Churchill Downs boosted its rates to the same thresholds in 2014, and records indicate that despite horseplayer backlash handle has continued to increase slightly over the past several years, from $501.3 million in 2014 to $511.8 million in 2015 and $516.9 million last year."

There are many problems with this 'fact', and this is very common in the horse racing business when racetracks offer out new policy.

  • The numbers include Derby week, which is an outlier. If it rained a few years, turf cards cancelled and handle tanked, it had little to do with a takeout rate. Derby week is, arguably, one of the least price sensitive week in all of North American racing.
  • The numbers do not include 2013, which was the last year at the old (lower) takeout rates
  • Field size and number of races have an effect - the elasticities for both are, according to a 1998 study, about -0.6. 
  • The business never includes real numbers, i.e. real as in including inflation rate. $501M in 2014 is worth $510M in 2016 for example. If the price of hay goes up, handle needs to go up too.
  • We'd want to look at underlying subsets for comparison and to gauge market share.  If Wal Mart sales are flat, but Dollarama up 15%, Wal Mart is going to want to know it. 
  • Factors - like the organized Churchill Downs boycott in 2014 - can be important and need to be examined
So what can we learn about Churchill's takeout hike, if anything?

In 2013 (omitted from the above statement), outside Derby week, at the 16%-19% takeout rates, handle was $322.6M.

In 2014 at the takeout rates Keeneland just hiked to, this handle was down to $248.4M, or off about $74 million.

2015: $252.6M

2016: $260.8M

Since 2013, total handle outside the Derby is down about 20%, or $62 million.

When looking at handle per entry, which includes field size, number of races and controls a little for weather and other outlying issues, the handle is down 13% since they changed the takeout rates.

When looked at the underlying handle numbers, which are down about 2% since that time (2.29% with inflation), Churchill has done appreciably worse.

Another phenomenon we've noticed is that large signals at big tracks are doing better since 2013. Last year, for example, tracks with over $200M in meet handle grew at 12% in gross. Churchill is one of these tracks and has not followed suit - they've been down (although in 2017 they seem to be catching up after a decent couple of months).

On the surface handle is down 20%, per entry handle down 13% and $62 million less is bet than it was a few years ago.

Those who are pro-takeout hike, or carry water for industry decisions often pivot their argument by moving the goalposts at this point, and talk about gross revenue. That, in my view, is a complete rabbit hole that does no one any good at all. As I've often noted, if any track wants to increase gross revenue for a few years just double takeout rates and that will happen. It's what happens afterwards that's the problem. When Italy doubled trifecta takeout to 41% revenue was up for awhile, but now Italian racing is Blockbuster Video.

So, it's the time I should probably conclude that the takeout rake hike was absolutely horrible - the worst thing ever! After all, $322M in handle down to $260M with other big signal tracks holding their own or growing since that time, dates down, races down, and horses entered down is not good news. The supply and demand ecosystem have both struggled.

But if I have a lick of intellectual honesty, I can't even do that - with rebating, signal fees, upwards of handle coming from large players, and other "noise", we can't really conclude anything with 100% certainty. At best we can say the takeout increase at Churchill probably didn't do any good - other than the increased "EBITDA" Derby week, which they could've achieved by hiking takeout for Derby week alone.

What I offer as an argument is simple: What makes for analysis in this business when people argue big, important items, is poisonous to the future of horse racing. Without honest, detailed, statistically significant modeled data, the sport will never get anywhere. And we'll likely see more moves like Keeneland did Monday, buttressed with low hanging fruit data, that says whatever they want it to say.


kyle said...

Two good posts on the Keeneland thing. Think you got them both right.

Anonymous said...

A well done analysis, PTP.

By the way, Keeneland has not announced the increases in their news site yet. Maybe never?

Keeneland has fallen out of overall first-rank status along with a noticeably smaller average field size since they dropped Polytrack. I don’t see a touted purse boost increasing field size. Instead I see them just paying the same group of horses more money, IF there is more money for them, as has happened elsewhere. There’s too many other racetracks of similar caliber running at the same time on most of Keeneland’s dates.

A 3% increase in takeout insures more early tapouts, so less churn, no increase and a possible decrease in total handle. I think by the end of the meet purses might be down over last year, despite Fall Stars Weekend on the first 3 days.

You can’t get blood out of a stone, as is said, but Keeneland will be the next in line to try.

-- Eudaemon

Most Trafficked, Last 12 Months


Carryovers Provide Big Reach and an Immediate Return

Sinking marketing money directly into the horseplayer by seeding pools is effective, in both theory and practice In Ontario and elsewher...