Del Mar Thoroughbred Club will adopt a policy they believe will limit betting in their win pool from computer-assisted wagerers (CAWs), beginning with its 2023 summer meet which opens in July. The policy change was confirmed to the Thoroughbred Idea Foundation by Del Mar president Josh Rubinstein, in response to various questions from TIF about CAW play at Del Mar and in California, overall.

“We get that late odds fluctuation is frustrating to players,” Rubinstein said.

“Beginning with our upcoming summer meet, [we] will be implementing new protocols for CAW players that are designed to limit participation in the win pools at two minutes to post. Based on what we’ve monitored and analyzed from other tracks, we believe these protocols will reduce late odds changes.”

Rubinstein added that the change will create financial incentives for the CAWs to make their win pool bets with more than two minutes to post, while disincentivizing them from making bets closer to the start of the race.

According to data from the California Horse Racing Board (CHRB) and TIF research thereof, the top 17 CAW accounts, betting through Elite Turf Club (Elite) and Racing and Gaming Services (RGS), represented just more than 14% of Del Mar’s win pool in 2022, the lowest total percentage of their play at Del Mar, by pool.

The total percentages are listed below.

1 - Elite and RGS Groups Percentage by Pool.png
*Del Mar hosted the Breeders’ Cup in 2017 and 2021.

While Rubinstein declined to provide specific detail, citing “proprietary commercial agreements,” TIF has learned through various industry sources, which asked not to be named, that CAW players receive a menu of potential rebates on win pool bets under such a setup – the highest rebate if win pool bets come more than three minutes before post, a smaller rebate if their bets are made between two minutes and three minutes out, and the smallest rebate inside of two minutes.

The takeout rate on the win pool in California, set by statute, is 15.43%, the rate which faces all non-rebated customers. Discussions with industry sources suggest the rebate received inside of two minutes to post is roughly half the rebate offered for win bets placed between two and three minutes to post.

Del Mar’s decision to implement this guardrail on CAW play does not go as far as the New York Racing Association, whose subsidiary is a minority owner of Elite, which banned CAWs from betting in the win pool inside of two minutes for the last two years while also barring them from using their CAW platform to bet the pick six and the late pick five across Aqueduct, Belmont and Saratoga.

An April 2023 story from the TDN’s Dan Ross revealed several modifications to CAW play at Santa Anita in a fashion similar to what is suggested by Rubinstein for Del Mar, though results on their impact are only anecdotal.

“CAW players must pay a surcharge of around 3.5% on top of their normal rate if they want to bet to the close of the win-pool, said [TOC president and CEO Bill] Nader.”

In that story, Ross reported Nader suggested that “around 25%” should be a limit on CAW play. CHRB records indicate the total of CAW betting from only Elite and RGS players, which notably does not represent the entirety of CAW play, was over 27% in 2022.


Limiting CAW guardrails to merely a rebate change in the win pool, while potentially helpful in reducing some late odds changes, amounts to little more than window dressing for mainstream customers.

CAW betting in Del Mar’s exotic pari-mutuel pools, particularly the trifecta, superfecta, pick three, four, five and six pools has been substantial. Play from Elite and RGS customers exceeded 32% of total handle in each of those bet types last year and was 38% of pick five handle.

CHRB data does not delineate how much is staked on the early pick five, with its friendly 14% takeout to mainstream bettors, and how much is on the high takeout late pick five, with a 23.68% takeout rate. It is reasonable to expect the percentage of CAW play reflects their enormous pricing advantage over the public on the late pick five.

Total handle figures do not tell the most accurate story. In nearly every pool offered by Del Mar, all non-Elite and RGS handle is declining.

According to the CHRB data, Del Mar ran 454 races in 2018 over 52 race days and 408 races from 44 race days in 2022. Meanwhile, inflation from the end of the Del Mar meet in 2018 has risen by 18%.

TIF has broken out handle figures, per-pool, per-race (or in the case of bets offered only once or twice per race day, like the pick four, five and six, a per-opportunity basis), and adjusted the 2018 handle for inflation. This normalizes the comparison between handle wagered in 2022 and 2018.

The reality is staggering.

Of the 11 main pool types offered by Del Mar. real (inflation-adjusted) handle from the Elite and RGS players has grown by at least 23% in nine of 11 pools. Elite and RGS betting is up 49% in trifectas, 46% in superfectas, 78% in the pick five, 128% in show pools and 403% in the pick six pool.

2 - 2022 v 2018 Real Handle by Segment.png

The story for all customers other than Elite and RGS CAWs looks rather different.

Real per-race or per-opportunity handle from everyone else wagering on Del Mar’s races is down in nine of 11 pools, with declines in the show and pick four pools the most substantial, at 15.2% and 19.2%, respectively.

Real win pool growth is only 1.6% over the period.

The most substantial growth has come in Del Mar’s pick six pool, a single-ticket jackpot offering, now with a low minimum investment since 2021 ($0.20), and which carries an effective daily  takeout rate of more than 46%.

"It's pretty much all bad news on this front," said TIF executive director Patrick Cummings.

“With the exception of the pick six pool, which carries the highest effective takeout rate and the lowest chance of actually winning, real handle in nearly every pool from all customers that are not playing through Elite and RGS is down.”

“If the track was hoping to increase revenue from wagering to purses, the last place you’d want growth to be isolated was the one pool where customers were least likely to win, the pool with effectively the lowest pari-mutuel churn. But that’s what we see here.

“The data from the CHRB, which has recently changed for the worse [see the note at the conclusion of this story], has enabled this analysis and provides a remarkable window into the state of pari-mutuel wagering in California and how it’s shifted in just the last few years. Unfortunately, we estimate that many other tracks around North America are likely very similar to what we see from Del Mar. This is most likely not an isolated case.”


California, which relies on wagering as its only source of purse money, has struggled to compete against jurisdictions with purse supplements from gaming in states such as New York and Kentucky.

Wagering on racing is essentially the only source of purses for California horsemen.

“Thanks to the handle we generate, we’re planning to pay $870,000 per day in purses this summer. That’s a significant figure for a venue without supplemental purse income,” Rubinstein said.

As the chart above showed, the main source of handle growth is contributing the lowest percentage for purses. This presents a serious, long-term concern for California and its horsemen.

Increases in CAW betting do not make-up for the decline from all other bettors.

According to CHRB figures, purses generated per-race from wagering on Del Mar’s live races were $37,406 in 2018 and $37,617 in 2022, a minor 0.5% increase. Adjusted for inflation, real purses generated from wagering per race in 2018 would be the equivalent of $44,185, a figure which represents a 14.9% decline in real per-race purses generated from wagering in the last five years.

CAW betting has increased, mass market play and real purses generated per-race from wagering have declined.

This impacts racing in California far more than any other major jurisdiction in North America.

Del Mar Paddock - Evers Photo.png
Horses circle the paddock at Del Mar


CAW betting, on its own, is not bad.

"North American racing's betting infrastructure lacks guardrails which could more sustainably support CAW play while not dissuading, frustrating or outright crushing ordinary horseplayers," said Cummings.

“It is wholly unrealistic to expect a future where customers are using less technology to assist in wager creation, analysis and bet placement,” said Cummings. “We cannot act in a way that suggests we are allergic to algorithms, that’s just anti-progress.

“What we should be doing is embracing a future where track operators and betting services providers seek a wholesale improvement in the wagering experience for mass market customers.”

Here are five steps for industry stakeholders, particularly track and betting operators, to consider.

1. Until more meaningful actions can be taken, for now, all North American tracks should ban CAW play in the win pool inside of two minutes to post.

Designating one major pool as free of CAW play is a token sign to mainstream horseplayers that track operators care about attracting their business.

All North American tracks should follow NYRA’s lead and ban CAW play from the win pool inside of two minutes to post. While this would likely stop all CAW play in that pool altogether, such an act would likely have the least impact on the CAWs since it traditionally garners a smaller portion of their overall handle.

Win betting has grown more challenging for, and less marketable to, all other players who do not have access to submit bets instantaneously as the last horse walks into the gate. The last customers into the pools, often the biggest, are setting the prices for all others already committed.

Doing the same in place and show pools would be reasonable too.

Tracks should also designate at least one multi-race, exotic pool as CAW-free, similar to NYRA’s elimination of CAW play in their late pick five.

If NYRA, which owns a minority share of Elite Turf Club, can take these steps, it reasons that all others could follow suit without disruption.

2. Pursue policies to decrease takeout for mainstream players in all bet types.

The growth of CAW play proves an undeniable reality which the industry long ignored to its detriment – pricing matters. CAW players are surely the most price sensitive customers in the market, but it matters to almost all bettors.  

One possible combination that should be considered is a future with 12% takeout in the win, place and show pools, 14% takeout in two-horse/race exotics and no higher than 20% takeout on bets with three or more horses/races.

Blended takeout, likely near 21% at present, should come down by at least one-quarter, to roughly 16%, with efforts to continue pushing it lower. Breakage across the continent should shift to the penny, emulating Kentucky’s efforts, amended to that level in 2022. Some tracks have made adjustments to select takeout rates of late, notably Hawthorne offering 12% on all win, place and show betting and Emerald Downs dropping their double and pick three takeout rates to 10%.

As high-frequency trading grew in equity markets through the 2000s, transaction costs for all investors dropped precipitously, even for 401k account holders and mutual fund investors which did not trade frequently. Racing’s failure to reduce takeout as most wagering has shifted online, and CAW participation ballooned, is inexcusable. 

The evolution of host fees, state-wide and even municipality-based source market fees make changing takeout a challenging endeavor. In some states, like California, statutory changes are needed. It is not an easy lift, which is why most pricing changes (for the CAWs) were driven by businesses operating outside the regulatory and legislative process. Changes like the ones suggested above could drive more than half a billion dollars annually back to players, subsequently promoting increased churn and renewing the takeout cycle more often on money that has, until now, been pulled and retained.

Handle should rise, long-term. Takeout rates have long bene too high to generate maximum revenue.

In many states, purse generation from wagering is no longer a primary concern. Gaming is the driver. While that could have underwritten a model which produced lower takeout rates for all players, the industry instead sharply discounted prices to a very small percentage of customers which has allowed CAWs to run amok, driving many long-standing horseplayers from the sport, reducing total handle from others and serving as a bearish indicator to anyone critically analyzing the data.

Horseplayer retention and development is a necessity, particularly as racing’s social license to operate faces ever-increasing threats. Lowering takeout, across-the-board, would be a positive, industry-sustaining act.

3. Tracks must stop offering jackpot bets and shift wagering marketing tactics which have “successfully” driven bettors to low-churn bet types.  

Jackpot bets have to go.

The very existence of the jackpot bet – one in which the accumulated jackpot is paid only when a single winning ticket exists – takes money out of circulation, defying the goal of the pari-mutuel wagering operator to keep gathering more commissions from bettors.

TIF acquired settlement data from the state of Florida which lists the total bet and returned to all bet-takers at Gulfstream Park on a given day. While the data was not divided by pool, it is still useful to garner a sense of how different customer segments perform on the day when a long-accruing jackpot bet, like Gulfstream’s Rainbow 6, is paid out to anyone with a winning ticket. These “mandatory payout days” are meant to attract players to come and get a share of the accumulated jackpot.

As TIF learned in the process, mandatory payout days are often major wins for CAWs at the expense of mass market customers.

Below is a sample of mandatory payout days from 2015 to 2021 at Gulfstream Park. The columns on the left reflect the total CAW play as a percentage of total handle and total winnings paid on those days, plus the return on investment for the CAWs. On the right, in the yellow-headed column, the data reflected are the combined returns of mass-market players through TVG, TwinSpires, Xpressbet and NYRA Bets.

Blended takeout rates are currently around 20% to 21%. On a mandatory payout day, where an accumulated jackpot is paid out, it is reasonable to expect overall blended takeout will be smaller than normal because the carryover is added to the total payout. Instead, mass-market ADW players get back less than normal, and lose well in excess of blended takeout.

3 - GP Mandatory Payout Days Info.png

Del Mar’s jackpot pick six operates with a daily effective takeout of 46%, while some other tracks’ offerings have daily effective rates rise into the 70% range. Retaining customers’ bets while paying out a small portion each day, and using the accumulated jackpot to attract more play is a long-term losing strategy.

Yet, for years, North American racetracks have relentlessly marketed their jackpot bets while also seeking to drive customers to other multi-race bets, reducing the ability of customers to churn winnings.

Just last weekend, when Santa Anita offered a mandatory payout on its jackpot pick six play, its parent company’s ADW, Xpressbet, sought to attract play by offering up to $20 back as a bonus if players staked $100 on the bet, both in the two days before the mandatory payout and on the day itself. The parent company of both Santa Anita and Xpressbet, 1/ST Racing, is the majority owner of Elite Turf Club.

The marketing has “worked,” as has the reduction of minimum bet amounts on such plays. Total handle for the pick six on their last mandatory payout day, June 18, was over $3.7 million. TIF observations of the accumulations of the pool suggest well more than 40% of that came from CAW play, chasing a carryover of just more than $346,000. These figures are similar to other mandatory payout day observations with smaller jackpots. When the carryovers balloon, so too does CAW handle.

While tracks and regulators generally do not identify what amount of mandatory winnings went to mass market players and CAWs, the Gulfstream data suggests these sorts of bets serve mostly as handouts to CAWs and as dumpster fires for mass market players.

Fortunately, a few tracks in recent years have recognized the jackpot disaster. NYRA and Keeneland abandoned their jackpot provisions in recent years, while Horseshoe Indianapolis eliminated its jackpot bet in 2023.

4. Track operators and tote companies must invest in a substantial overhaul of pari-mutuel processing and betting integrity oversight measures.

TIF’s April 2021 series Wagering Insecurity outlined the many challenges unaddressed for years, decades even, amongst pari-mutuel bet processing in North America. Early 2023 announcements from the Hong Kong Jockey Club, the leader in commingled pool arrangements for major international race days, suggested the Club was seeking to branch out on its own with new tote developments, away from the antiquated, existing protocols used by all American-based tote providers.

It should be a question of when, not if, the investment is delivered by racing and betting operators to bring much-needed modernization to pari-mutuel betting.

CAWs betting into North American racing have been given push-button access to dump thousands of bets into pools, essentially, instantaneously. Smaller players wishing to upload bets must use programs that are often metered by their bet-taker – the process can take minutes to get accepted what it takes a formal CAW a fraction of a second to accomplish.

As long as antiquated tote technology remains operational, mid-level customers seeking to grow may be structurally impeded from doing so thanks to North American racing’s failure to invest.

Access to the single pari-mutuel pools available for a track’s races remains unequal. This is one of many issues inhibiting industry wagering growth.

5. Innovate (or imitate) ways to reflect pricing changes (and anticipated final prices) across pools.

Track and betting operators should take steps to publish the imputed win odds from previously closed pools and reflect that across all platforms too.

If the closed daily double and pick three pool shows that #6 is favored, and the horse was bet as if it will be a 2-1 favorite, but its win odds are currently 7-1, the displayed live odds in the win pool are not an accurate reflection of the likely final price of the horse. Why should anyone tolerate misleading information for the public when it is very likely the same bettors who bet the horse into favoritism in previously closed multi-race pools who will eventually show up in the win pool – many of them in the final flash?

Why accept the heartache of a late odds drop when the drop was completely predictable a half-hour earlier thanks to a previously closed pool?

In the absence of widespread adoption of the first recommendation above – eliminating CAW play in the win pool – this recommendation becomes even more important. Mitigating the pain felt by horseplayers thanks to late odds drops could be simply accomplished if those same players had it made clear to them, earlier, the imputed win odds of horses based on other pool data.

It should be basic customer service, but not a single track in North America has done it.

Additionally, any display a horseplayer might reference to follow changes in the market for a race – a track feed, TV broadcast or ADW display – should implement graphical enhancements that are far more detailed and engaging.

Hong Kong has done this well for years.

CAW teams cascade their money into Hong Kong’s vast betting pools in the final minutes before a race (not just as the last horse enters the gate, itself a step that should eventually be required of CAWs in North America). With each frequent update over the final minutes, not seconds, of betting, the HKJC’s odds displays – on track, broadcasts and across its ADW site – light-up with various colors to identify how quinella, quinella place (omni/swinger), win and place odds are changing. Instead of just a solid grid, the public is visibly directed to different shades depicting what percentage the prices are dropping.

This comes at a cost to the CAWs, surely. An invisible tax, so to speak, allowing the public to free ride on the action of the CAWs, whose bets are not enabled to reach the pools all in the final flash. The CAWs have tolerated it for years. It is literally the price they pay to participate in the big pools, and it continues without controversy.


That such a vast amount of racing wagering is generated from such a few CAW players is a monumental risk facing the sport and its horsemen. There is no suggestion within this publication that CAW play should be banned outright, nor gradually eliminated, but it must be managed better, more fairly, and in such a way that other emerging tech-driven players can compete.

The negative effects of CAW play are now more noticeable than ever. At racetracks around North America in 2023, TIF has recorded dozens of examples of exotic pools where more than 50% of all money staked is reflected publicly only after the last few horses take their spot in the gate, and most of it does not come until the race is already running.

Across three tracks in one week in May 2023, TIF found 56% of an exacta pool, 64% of a trifecta pool at a second track and 71% of yet another’s superfecta pool reflected publicly in those last seconds. When asked about the movements, racing executives, particularly those at smaller tracks, have expressed a feeling of helplessness to do anything about it.

“I can’t be the one to stop it,” said one who asked not to be named, “and I sure can’t be the one to turn them away.”

But some are at least rejecting the advances of one player, believed to be the biggest in North American racing (and is the biggest player in California, based on 2021 CHRB records), whose representatives are allegedly offering tracks a deal one executive compared to a “pay day loan for a struggling track.”

TIF heard from multiple track executives which confirmed the deal offered to them involved the CAW player pre-paying a massive portion of the host fee the CAW player paid the track the previous year.

If accepted, the track would cut the host fee paid by one player on all betting over the pre-paid amount, essentially amounting to a massive rebate increase on new handle.

TIF was unable to independently confirm if any track accepted the deal in recent years, but observations of mind-blowing late moves in some pools suggests not all rejected the advance. Not only would acceptance of such a deal crush mainstream horseplayers, but it also has significantly negative consequences on other CAW players too.

The apparent absence of regulators in this is worth noting.


CAW betting is engrained in North American racing wagering. There will be more of it in the future. Managing its growth in a more sustainable fashion is the key.

CAW participation is undoubtedly driving revenue to horsemen, but as the Del Mar data suggests, their low-value growth is not keeping pace with declines in mass market wagering.

While much of the data reflected earlier is based on wagering from Del Mar, it is reasonable to believe that nearly all tracks in North America have a similar problem - a growing imbalance between their customer bases. For most of those tracks, the impact has been less visible. Purse supplements from other forms of gaming have served as a handout for many in the sport – both track operators and horsemen – enabling their indifference to retaining and building a mass-market racing audience.

Many steps are needed to improve the present-day realities of racing wagering in North America. The recommendations above are hardly exhaustive, but would improve the sport’s mainstream wagering popularity – something which benefits all racing stakeholders.  


TIF, and several journalists, contacted the CHRB inquiring as to why the reporting style of wagering data, on record since 2008, was changed for the first time in 2022.

CHRB spokesperson Mike Marten provided the following explanation about the change:

“This change came about when we learned through an intermediary that some customers had complained about the unfairness of being publicly identified, whereas thousands of other bettors are not so identified. CHRB Executive Director Scott Chaney carefully considered the matter and concluded that it was in fact not fair to identify some bettors (i.e., wagering groups) but not others. We then asked our statistical department to identify those groups and eliminate the individual listings by grouping them into the site report…

“The practice of listing individuals/groups was started by some technical person many years ago apparently without thought to transparency, public input, or reporting. The CHRB is not bound to a report-generating program of unknown origin. Mr. Chaney only recently became aware of the practice and decided that reports should treat everyone the same.

“The goal is to eliminate identification of individuals or wagering groups. The objective is to treat everyone the same.”

The CHRB’s decision came only after the data was publicized and garnered significant attention in mainstream press, most notably Oliver Roeder’s exposition for the Financial Times

“Once questions were asked and requests filed, and some of this information became public, something seems to have changed from the regulatory side,” said Cummings.

After nearly a dozen submissions for wagering information from Florida tracks was acquired by TIF at a cost from the Department of Business and Professional Regulation, and some of the findings publicized, subsequent requests have gone unfulfilled.

“We have several requests still pending with Florida, some open for several months now,” said Cummings. “This is highly unusual behavior considering the speed with which such requests were regularly processed in the past after their invoices were paid. Our follow-up inquiries about them have gone unanswered.”


Popular Posts