Interesting Robotic Wagering discussion started by Frank Angst

[Kasept]

Interesting piece by Frank Angst at Bloodhorse about Computer Robototic Wagering and whether it could fall under the lurching restrictions on DFS.. Many fascinating aspects to this whole arena including rebates, tote protocols and more..

http://cs.bloodhorse.com/blogs/wgoh/...ank-angst.aspx

Computer-Robotic Wagering (CRW)
By Frank Angst

Computer-robotic wagering (CRW) has helped prop up pari-mutuel handle in recent years but has made race betting less attractive for every other player by making it difficult for racing to attract new bettors. Now, based on law enforcement officials’ assessment of fantasy sports wagering, CRW may be illegal.

Considering one in five dollars bet into pari-mutuel pools in 2014 came from CRW, a determination that this betting form is illegal would devastate racing’s economic engine short-term. But long-term, eliminating this unfair advantage in which players use computers programmed to find value in pari-mutuel pools and then target wagers into those pools could lead to a healthier business model.

The instantaneous knowledge of probable payouts, ability to wager thousands of different combinations, and the capability to make those bets just a few seconds before the race when odds have settled, provide CRW players unfair advantages over other bettors. Even if that “human” bettor could instantly discover pool value, speech limitations and relatively slow fingertips would prevent punching in thousands of wagers to take full advantage of that insight.

This setup is especially damaging in converting new players into regular players. Even if new players don’t fully understand what is happening when they bet races, they soon recognize there are few instances where the winning payout is better than expected. The chance of finding that type of payout—an overlay—is race wagering’s real attraction, one of its few advantages over casino games. CRW players have helped reduce instances of such overlays, weakening the betting product for everyone else.

When racing has problems competing with casinos for wagering dollars, some experts will tell you it’s because pari-mutuel wagering is too complicated. Don’t buy that for a second. Poker players have invested countless hours reading books and attending camps to improve their games. Fantasy sports players devote days on end to reading stat sheets and news reports looking for ways to top their rivals and win money.

Pari-mutuel wagering on horse racing is the original daily fantasy sport and should be enjoying a renaissance as bettors look for a greater challenge than pushing a slot machine button. Casino executives at this year’s Global Gaming Expo acknowledged some floor space needs to be converted from mindless slot machines to games-of-skill preferred by younger bettors. But racing’s current model that relies on high takeout and a select few CRW players poaching off others has left pari-mutuel wagering ill-equipped to take advantage of these trends.

New horseplayers are no match for CRW players who receive lucrative rebates to encourage increased play. Those rebates are made possible, in part, by the high takeout paid by new players. If this sad cycle sounds criminal, it may be.

When New York State Attorney General Eric Schneiderman this month asked the two leading daily fantasy sites to stop operating in the state, he referenced unfair advantages for some players.

“… both companies consistently use deceptive advertising to lure consumers into an unregulated online gambling operation that, while marketed as a game that anyone can win, in fact distributes the vast majority of winnings to a small subset of experienced, highly sophisticated players,” Schneiderman said.

The New York Times and Wall Street Journal have reported the FBI is investigating similar problems in daily fantasy sports. The Times reported in October, “many of the professional (fantasy players) use automated processes that let them change hundreds, if not thousands, of lineups in seconds, a decided advantage when last-minute changes are made in the lineups in real football, basketball, or baseball teams.”

That description sounds a lot like CRW. It may be only a matter of time before law enforcement takes a look at horse racing’s unfair pari-mutuel model.

Despite the short-term pain, it’s time for racing to shake its unfair model. Perhaps CRW could gradually be phased out before law enforcement takes action. After all, the money pumped into the current system by CRW players has not been enough to offset the lack of new players sticking around, players reducing play, or players completely leaving the sport. Pari-mutuel wagering is declining or flat and CRW players are part of the reason other players are reducing play and new players aren’t sticking.

Eliminating CRW and putting in place optimal takeout rates for all players so that everyone competes equally would restore the attraction fantasy sports bettors currently enjoy—finding the edge. Millions of new players are finding that attraction in fantasy sports; they should be finding it in horse racing.

[Benny]

Here, here; about time

[Kasept]

Originally Posted by Benny

Here, here; about time

Why and for what? What about when these algorithms miss and exactas overpay? The CRW is in it for the rebate and are only looking to stay close to breakeven. Who says their money in the pools isn't a net gain?

The takeout question is a separate issue within this discussion as is the access the CRW has to interject at the last second. The unknown is what is the potential positive impact of the CRW money out of the pools might be and would their exit (exclusion) help trim takeout rates to bring bettors to the table?

There's a LOT of moving parts to this.

[Dunbar]

Originally Posted by Kasept

Why and for what? What about when these algorithms miss and exactas overpay?

What do you mean by "overpay". If you mean that the exacta pays more than the last projected payout, okay. But if you mean the bettor is getting some kind of windfall, then no. Just because the CRW "misses" doesn't mean the payout becomes an overlay. More likely, the winning bet was still an underlay.

Originally Posted by Kasept

The CRW is in it for the rebate and are only looking to stay close to breakeven. Who says their money in the pools isn't a net gain?

It may or may not be a net gain for the track. But it's definitely a net loss for the average bettor. This quote from Tampa Bay Downs's manager in your Scott Jagow link sums it up accurately:

"Effectively what happens when you have these high-volume outfits and they win more than the norm, you're effectively raising your takeout on the rest of the players in your pools. And that's just not good business.”

If takeout is 18% and some group or groups with a bet-timing advantage can breakeven while making up 1/3 of the pool, then the rest of the bettors will be losing at a combined rate of 27%.

Originally Posted by Kasept

The takeout question is a separate issue within this discussion as is the access the CRW has to interject at the last second. The unknown is what is the potential positive impact of the CRW money out of the pools might be and would their exit (exclusion) help trim takeout rates to bring bettors to the table?

There's a LOT of moving parts to this.

Agree. But as far as computer access to the pool info and the ability to bet after all non-computer bettors, a CRW seems to have patently unfair advantages.

[odbaxter]

Can someone please post a couple of links to these types of CRW's?
I'm not necessarily going to give up the experience that picking my own winners brings, but not looking into this seems to be sticking my head in the sand.

[Kasept]

Originally Posted by odbaxter

Can someone please post a couple of links to these types of CRW's? I'm not necessarily going to give up the experience that picking my own winners brings, but not looking into this seems to be sticking my head in the sand.

Haven't seen where the CRW's have ever made that public offering opportunity suggested in Jagow's piece.. Having a discussion with Maury Wolff this afternoon to get the best understanding possible of the topic in its' entirety. Will find out.

[Kasept]

Originally Posted by Dunbar

What do you mean by "overpay". If you mean that the exacta pays more than the last projected payout, okay. But if you mean the bettor is getting some kind of windfall, then no. Just because the CRW "misses" doesn't mean the payout becomes an overlay. More likely, the winning bet was still an underlay.

Just as the exactas pay less than the late probable displayed, they 'should' pay more when unidentified exacta targets come in. But the robotics cover an inordinate amount of potential results and widely depress the prices, yes..

It may or may not be a net gain for the track. But it's definitely a net loss for the average bettor. This quote from Tampa Bay Downs's manager in your Scott Jagow link sums it up accurately:

"Effectively what happens when you have these high-volume outfits and they win more than the norm, you're effectively raising your takeout on the rest of the players in your pools. And that's just not good business.”

If takeout is 18% and some group or groups with a bet-timing advantage can breakeven while making up 1/3 of the pool, then the rest of the bettors will be losing at a combined rate of 27%.

Right. That's the whole question though the working number is 15-20% of total handle and the WPS/exacta pools are their area. And you don't necessarily know where the target is at any particular time. You can gravitate toward tracks like OP and TAM to avoid them or play in pools where their programming is neutered (tris, super, multis).

Agree. But as far as computer access to the pool info and the ability to bet after all non-computer bettors, a CRW seems to have patently unfair advantages.

I'm not an economics guy, but aren't most equity markets less than a level playing field? Isn't there a patently unfair advantage for any whale getting rebates or with players that can punch $1,500 P6 tix over a typical $96 or $244 play? Where is the line between unbalanced and unacceptable? They certainly have an advantage and make it hard(er) for everyone else, but is that reason to exclude them from the pools?

The whole thing is a fascinating debate and emblematic of how complex the game and its' vexing issues are..

[Dunbar]

Originally Posted by Kasept

Just as the exactas pay less than the late probable displayed, they 'should' pay more when unidentified exacta targets come in. But the robotics cover an inordinate amount of potential results and widely depress the prices, yes..

The bottom line remains that the rest of the bettors do worse than they would do if the breakeven late bettors weren't in the pool.

Originally Posted by Kasept

Right. That's the whole question though the working number is 15-20% of total handle and the WPS/exacta pools are their area. And you don't necessarily know where the target is at any particular time. You can gravitate toward tracks like OP and TAM to avoid them or play in pools where their programming is neutered (tris, super, multis).

I don't get why CRW's wouldn't be betting trifectas. I thought they did, in fact. Maybe you can ask Maury Wolff.

Originally Posted by Kasept

I'm not an economics guy, but aren't most equity markets less than a level playing field? Isn't there a patently unfair advantage for any whale getting rebates or with players that can punch $1,500 P6 tix over a typical $96 or $244 play? Where is the line between unbalanced and unacceptable? They certainly have an advantage and make it hard(er) for everyone else, but is that reason to exclude them from the pools?

Those are good points about the rebate and bankroll advantages. But IMO no one should be given an information advantage that is not available to other bettors. That's where the line should be drawn. The CRWs get two types of information advantage. First, they get a digital form of the complete pool breakdown, allowing them to instantly spot overlays. Second, they can bet at the last second, giving them payoff information that is better than what other bettors can know.

I can't think of any other (legal) gambling situation where one group is preferentially allowed to bet last. Who would ever sit down at a Hold'em game if he or she had to always bet from early position? Only a total rube.

Originally Posted by kasept

The whole thing is a fascinating debate and emblematic of how complex the game and its' vexing issues are..

Absolutely!

[philcski]

Originally Posted by Kasept

Just as the exactas pay less than the late probable displayed, they 'should' pay more when unidentified exacta targets come in. But the robotics cover an inordinate amount of potential results and widely depress the prices, yes..



Right. That's the whole question though the working number is 15-20% of total handle and the WPS/exacta pools are their area. And you don't necessarily know where the target is at any particular time. You can gravitate toward tracks like OP and TAM to avoid them or play in pools where their programming is neutered (tris, super, multis).



I'm not an economics guy, but aren't most equity markets less than a level playing field? Isn't there a patently unfair advantage for any whale getting rebates or with players that can punch $1,500 P6 tix over a typical $96 or $244 play? Where is the line between unbalanced and unacceptable? They certainly have an advantage and make it hard(er) for everyone else, but is that reason to exclude them from the pools?

The whole thing is a fascinating debate and emblematic of how complex the game and its' vexing issues are..

I think you are viewing this example from a "one trial" perspective rather than "in the long run". This matters more in a Pick 6 or other low probability wager because the volatility in success is much higher with a lower bankroll, but the net ROI over many trials for an equal player doesn't matter given the bankroll (assuming they are betting the same percentage of their bankroll to equalize the risk of ruin). Example:

Player (a) has a bankroll of $9,600 for the year and bets 100 pick 6's, all for the same $96. He hits 1 out of 100 (1%) for $10,000, for an ROI of +4.167%.

Player (b) has a bankroll of $150,000 for the year and bets 100 pick 6's, all for the same $1500. Because his "skill" is equal to player (a), he hits at the exact proportion to the amount wagered by player (a), so 15.625 of his pick 6's for the year for $10,000 each. His return is 4.167% as well.

The issue is volatility, which on a low probability wager is extremely high even with a large bankroll. Because player (a) wins at a much less frequent rate, it feels like he's a net loser to player (b), but in the long run their percentage returns would converge.

Similar situation: One player bets only even money shots. The other only bets 10-1 shots. Assuming the likelihood that the even money shot is exactly 10x more likely to win any given bet, the only difference in their long term returns would be how effective their selections are. But the paths to that final return would be very different.

[Calzone Lord]

Originally Posted by Kasept

Why and for what? What about when these algorithms miss and exactas overpay? The CRW is in it for the rebate and are only looking to stay close to breakeven. Who says their money in the pools isn't a net gain?

If they can achieve the ability to stay close to breakeven while betting at a high volume and moving odds late -- they're terrible competition to have from a bettors prospective.

I'd obviously much rather bet against hunch players or people who overbet popular jockeys. That said, most decent handicappers/horseplayers are not close to being long-term breakeven players. When I was 18-years-old and worked primarily as a teller but also occasionally customer service for Players Card members, I was able to monitor all the Off-track Wagering patrons who bet on a card. I knew the people and I could (and did) view and print out all their action over long stretches.

I don't play Daily Fantasy Sports, but I would also guess that similar people with sophisticated algorithms do well in relation to the average competent player.

That said, an algorithm will probably be more effective in DFS than it would assessing value on a horse race -- as I think more human skill and judgement can be applied to the latter.

[Kasept]

Originally Posted by Calzone Lord

That said, an algorithm will probably be more effective in DFS than it would assessing value on a horse race -- as I think more human skill and judgement can be applied to the latter.

What's interesting about that is the 'human skill and judgement' factor is equally what the CRW creators/operators count on as their grind-it-out edge..

[Kasept]

Here's Scott Jagow's extensive CRW piece from a couple years ago for added background..

http://www.paulickreport.com/news/ra...race-wagering/

[Benny]

http://www.bloodhorse.com/horse-raci...#disqus_thread


The industry's embrace of computer-robotic wagering through the high rebates required to sustain these players has resulted in a less attractive game for all other horseplayers, according to an expert presenting March 25 at the Association of Racing Commissioners International conference in New Orleans

[Kasept]

Dick Powell happens to be on show today to talk about DWC and the RCI meeting.. Merging this with the good discussion generated by Frank Angst's piece from November.

[odbaxter]

I haven't had a chance to listen to the podcast; any interesting additions to this discussion?