Gramm-McKinney Study Shows Late CAW Activity in NY Pools is Growing

[Wandering Eyes]

Editor's note, by Dan Ross:

Despite the New York Racing Association's efforts in recent years to limit Computer Assisted Wagering (CAW) participation in the Win, Late Pick Five, and Pick Six pools, the influence these teams exert in terms of last-cycle betting impacts appears to be growing in several other key pools, according to research conducted by economics professors Marshall Gramm and Nick McKinney, both of Rhodes College in Tennessee.

Indeed, they found that since 2022, the percentage share from CAW players of monies wagered last-minute into the Exacta, Trifecta, Superfecta and Early Pick 5 pools (among others), has grown significantly. Why is this a problem for trainers, owners and other industry stakeholders?

With worsening CAW encroachment within these betting pools has come growing attrition from racing's average punters, turned away from the sport by the competitive imbalance. And with the nation's purses fueled primarily by wagering monies, this is placing the economic future of the sport into the hands of fewer and fewer deep-pocketed and influential gamblers. The Gramm-McKinney report appears below.

 

Measuring CAW Participation and Growth Through NYRA Pool Restrictions

By Marshall Gramm and Nick McKinney

Large betting syndicates employing Computer Assisted Wagering (CAW) have become a powerful force in the horse racing industry, contributing significant handle but also disrupting betting pools with large, last-second wagers. These last-cycle bets cause sharp odds fluctuations that frustrate traditional horseplayers and undermine confidence in the game's integrity. Despite growing concern, little public data exists on the scale or growth of CAW activity. Using data from the past four Aqueduct Winter meets, we have estimated the size of CAW participation and how it has grown. Our findings suggest CAW teams now account for a much larger share of wagering than just a few years ago.

Armed with sophisticated wagering models, CAW teams place thousands of individual bets in the final seconds before post time. They use real-time odds to fine-tune their positions while keeping their activity hidden from the broader betting public. Their ability to flood the pools with targeted bets at scale, combined with rebates that often return half the takeout, gives them a significant edge over conventional horseplayers.

A recreational bettor might play dime superfectas at Saratoga using the box or key functions, spreading across dozens or even hundreds of combinations into a 24% takeout with limited efficiency. CAW teams, on the other hand, submit massive volumes of single-combination, optimally crafted superfecta bets executed with precision and boosted by substantial rebates.

Last-cycle wagering provides a measure of CAW activity, as anecdotal evidence and industry experience suggest that a significant portion of their money enters the pools in the final moments before post-time. We can gain insight into where and how CAW teams participate by examining the share of total handle wagered during this period, particularly by comparing pools where their access is restricted to those where they are free to participate.

The New York Racing Association (NYRA) restricted CAW access to its Win pool to combat sharp odds fluctuations. Additionally, it barred the teams from participating in the Late Pick Five and Pick Six, thereby offering a more level playing field for traditional horseplayers. These restrictions offer a rare opportunity to estimate both the size and growth of CAW wagering activity.

NYRA's Pick Five pools are particularly valuable for this type of analysis because they impose different restrictions within the same race card. NYRA racetracks typically offer at least two Pick Fives per card: an Early Pick Five (races 1-5) and a Late Pick Five (last five races). All Pick Fives have a 50-cent minimum and a 15% takeout. Importantly, CAW teams are excluded from the Late Pick Five but may participate freely in the Early Pick Five. Despite often featuring higher-quality races, the Late Pick Five consistently draws lower handle, likely reflecting CAW exclusion.

The table below shows the average share of each pool wagered during the final betting cycle, just before the race begins and the pools close. Over the past four Aqueduct winter meets, the share of last cycle handle in the Early Pick Five rose sharply, from 8.4% in 2022 to 32.0% in 2024 before falling slightly to 28.8% in 2025. The Late Pick Five, by contrast, showed only modest growth over the same period, increasing from 2.2% in 2022 to 9.4% in 2025. The gap between the two pools, those with and without CAW access, serves as an estimate of the share of wagering driven by CAW teams. This gap grew from 6.2 percentage points in 2022 to a peak of 23.0 points in 2024, highlighting the rapid expansion of their participation. While the roughly 20 percentage point difference in recent years supports the view that CAW teams make up around 20% of the Early Pick Five pool (a figure consistent with industry estimates), the pace of growth is more striking than the level itself.

 

Estimating CAW participation in other pools is less precise, as their betting activity likely varies by pool depending on model strength and the perceived inefficiency of each market. In general, CAW betting volume increases with wager complexity, with heavier involvement in exotics, such as trifectas and superfectas, than in straight bets, like win or place. Still, last-cycle activity offers a useful proxy.

In 2022, 11.0% of Win pool handle was wagered in the final cycle, compared to 17.5% for Exactas and 20.9% for Trifectas. By 2025, those figures had grown to 16.8%, 44.3%, and 47.6%, respectively. The gaps between Win and Exacta or Trifecta pools more than tripled, from 6.5 and 10.9 percentage points in 2022 to 27.5 and 30.8 in 2025. In the most recent meet, more than half of the Superfecta handle was bet in the final cycle. While less definitive than the Pick Five comparisons, these figures suggest that CAW teams now account for at least 30-35% of the Trifecta and Superfecta, up from around 10% just three years ago.

This analysis does not intend to single out NYRA. On the contrary, other racetracks should adopt the same CAW restrictions that NYRA has applied to the Win pool. Significant late odds shifts undermine bettors' confidence in the integrity of the pools and alienate new players accustomed to fixed-odds sports wagering. While exotic pools at other tracks also exhibit significant last-cycle movement, none have experienced as sharp an increase as those at NYRA.

It's important to remember that last-cycle wagering is only a proxy for estimating CAW involvement. The true scale of CAW participation remains opaque, known only to their ADWs and the racetracks. Not all last cycle money comes from CAW teams, just as some CAW wagers likely enter the pools earlier.

The use of advanced data modeling is a natural progression in the evolution of horse betting, and CAW teams are not inherently problematic in their pursuit of an edge. In a less-than-zero-sum game, once-profitable strategies inevitably become obsolete as better methods and sharper competition increase. The concern lies not with their sophistication but with the speed and concentration of their wagering activity. The flooding of pools in the final seconds destabilizes prices and undermines the parimutuel system.

Pierre Oller's invention was brilliant in its simplicity: the house takes a fixed cut, and winning bettors are paid from the pool rather than from the operator's pocket. Unlike fixed-odds sports betting, a winning horseplayer can continue to wager freely, often increasing handle over time. But the parimutuel system was never designed to accommodate large volumes of money flooding the pools in the final seconds. In the pre-simulcast era, such strategies were logistically impossible when all bets were placed at the window.

Technology has dramatically expanded the capabilities of bettors, but it also demands new safeguards. Restrictions on batch wagering should be extended to all visible pools, including Win, Place, Show, Exacta, and Double, to limit last-second price distortion. Just as critically, the rebate gap between CAW teams and ordinary bettors must be narrowed. Ideally, this would be achieved through across-the-board takeout reductions, improving fairness and sustainability across the entire wagering ecosystem.

Unfortunately, NYRA's last-cycle betting volume shows that the problem isn't stabilizing; it's worsening.

The post Gramm-McKinney Study Shows Late CAW Activity in NY Pools is Growing appeared first on TDN | Thoroughbred Daily News | Horse Racing News, Results and Video | Thoroughbred Breeding and Auctions.

View the full article