Monday, June 27, 2011

Why is Optimal Takeout So Difficult to Understand?

If you peruse the Interwebs and the question of takeout crops up, there is the inevitable chatter about the cost to put on the show, or my favorite line: 'we only raised takeout because we had to'. It seems, to this camper, the concept of optimal takeout is something that alludes many.

A poster on summed up optimal takeout once, and I thought it was pretty dang good. I share it here.

  • Optimal takeout means maximizing revenue. There is only one takeout rate that will maximize revenue for a given "demand for X" over the long term. Zero will never be an optimal rate, it returns zero is officially the least optimal takeout rate (it shares that honor with 100% takeout rate.)
  • Optimal takeout doesn't exist in anyone's mind, it's not an opinion. It exists, as an actual number. Nobody knows what it is, exactly...but the truth is out there. It's the takeout rate where you get maximum revenue.
  • The reason the cost of the show doesn't matter is because it doesn't...optimal pricing is going to be optimal pricing for any given show, no matter how much you pay the star. You can't do better than maximum revenue. Optimal takeout rate = maximum revenue for the tracks and horsemen.
  • Whether maximum revenue pays for the show or not is a separate issue -- because you can't get any more money from your customers. If you raise prices your revenue will go down because they will reduce their purchases at a higher rate than the increase in price makes you.If you lower prices your revenue will go down because they won't buy more of it a rate high enough to make up for the decrease in per unit cost. You end up a net loser again. You can't bring in any more money, ever, than you'd get with the optimum rate. There is no more money for your show.
  • If there isn't enough to pay for the show at that point, then your show is too expensive to put on, and you should either consider putting on a different show, or just stop putting on shows -- because you are going to go out of business. Not all shows are blockbusters, many result in bankruptcy.
  • It is believed by many that racing is currently charging more than the optimal rate, and is therefore leaving money on the table that could otherwise go to pay for the show.

It really is that simple. You don't want to charge 0% takeout because purses would be zero, you don't want to charge 100% because purses would be zero, too. There is a number, somewhere in between, that makes the tracks the most cash, and funds the purses to their highest possible level. Our current takeout rates were never decided upon by trial and error, or an econometric model (like for example, how a casino does it; Vegas slot machines charged 25% in 1970, but 5% today because they make more cash at 5%). They were decided upon by governments, horsemen and tracks raising them each time they felt they wanted money, with near impunity, because racing held a monopoly*. We're not a monopoly any longer, so the old rules simply do not apply.

Every study since 1970 has said that the optimum level is lower than the current takeout rate for horse racing, and if they're right, tracks are costing themselves money in the long term, when they raise takeout.

It's not about horseplayer greed, it is about answering a simple question that millions of successful businesses have answered before, and continue to answer each day. I am at a loss as to why this easy to understand concept is such a foreign one. The next time you hear "our show is expensive to put on so we have to have high takeout" I hope the above explains why many of us pull our hair out.

* In the early 1900's, the initial pari-mutuel takeout rate in the US was 5%. Source: Colin's Ghost.


Anonymous said...

The nuns at your local church bingo understand.

Anonymous said...

Very good summary. Four caveats.

1. Could be that zero revenue for racing kicks in at a much lower % rake than 100%. I'd bet it would.

2. Could be that there are regions of the demand curve which are relatively flat - more than one optimal takeout point/not necessarily single peaked revenue curve is a possibility.

3. Since every wager is a differentiated product, there is some optimal takeout rate for each. Maybe some are factually the same, but in principle setting the same rate for W, P and show may be completely sub-optimal.

4. All of this also assumes that taxing total handle is an optimal revenue generating mechanism in and of itself. It needn't be. You could tax winnings/returns, not stakes, and the entire system of generating revenue might be more conducive for incentives to bet, and permit more efficient taxation.

However, nobody in racing charged with setting takeout currently knows what the optimal takeout or price is on any wager, as far as I can make out, since there has been no systematic, credible study determining that. And from the recent dialogue there is a frightening percentage of those setting takeout that have absolutely no clue what a revenue curve for a product taxed at a proportional rate on total sales might even look like.

Mike D said...

Probably the most concise summary of the problem - and solution - I have read. I'd note that the optimal takeout is most likely not the same for all bets, which is why it's so hard to discover.

Cangamble said...

Yes, there is a problem with determining optimal takeouts. For instance, the once a year player can play against a takeout of 50% for what it is worth (how much handle is really generated from these players anyway?).
For those who go a little more frequently a lower takeout is important, because the more given back, the increased chance that the once a month or so player will eventually play more.
What horse racing needs to do is determine how much handle is generated from players who wager at least once a week. If it upwards of 90%, then really no differentiation should be made (just like there isn't when it comes to slots).
Of course, optimum takeouts are different depending on the bet type too. The higher the real average payout, the less likely all the money returned will be churned. Horse racing actually understands, or at least understood that as supers and tris had the highest takeout rates.
But now with 50 cent tris and 10 cent supers, that game has changed a bit too. Not only that but horse racing is ignoring this concept today as lately the pick 5's have been put on with the lowest takeout....makes no sense from a churn or optimal takeout perspective.

Anonymous said...


.... another reason to read blogs in lieu of industry websites-- you actually learn something, without racetrack bought-and-paid-for spin.


That Blog Guy said...

It's been years (gosh I feel old) since I took my economics class, but there must be a way to calculate the optimal takeout. Can't the industry find an economist to do the calculation?

Eric P said...

Ontario's harness tracks offer a tremendous opportunity to experiment with takeout. A large number of the tracks operate with a pittance in handle, so in essence they would be risking very little revenue by reducing take out.
I wonder why none of the tracks have tried experimenting with takeouts, especially since the CPMA has allowed each track to set their own takeout for each of their wagers??

Cangamble said...

Eric, that would be great for on track bettors, but who would take their signal in Canada?

Eric P said...

CG...fair point. My answer is, those who are interested in determining what the optimal takeout percentage is.

Anonymous said...

It continues to amuse the vast majority of the racing world when people who admittedly do not know what they're talking about

""Nobody knows what it is""

continue to harp on mindlessly while pointing the finger at others while being unable to see that they themselves are the core problem.

You don't know what it is because people like yourself, and Cangrumble, are the core problem in racing today.

There are indisputably too many 'sharks' and people who would aspire or consider themselves to be 'sharks' in today's mutuel pools.

That you are too dense to identify that central problem is further indicative that you don't know of what you speak.

The correct means toward solving this flawed equation is from the ground up.

It is only when racing as a whole does more for the novices all around every track in the land, that it can overcome you problems.

You simply can not be any part of the solution when in fact you are the problem.

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...