Automatic bettors are hard to find. This search has covered a lot of frequent flier miles, so much that I could probably be blamed for global warming. Occasionally, I find an automatic bettor. Then I wait for a year to revisit him, but he’s gone. Done in. Most of the time, I find no one who walks up to the window and obeys the dictates of a formula. But I have found one player, a guy named Ed Bain. I’ve been tracking Bain for years, and he’s still calmly plunking down automatic bets. And collecting.

Here’s our latest exchange.

Mark: This column is called “In Search of the Automatic Bet.” How long ago did your search begin, and how did it evolve until you finally came up with what we can call an automatic bet? Tell us about your search.

Ed: I bet speed for many years. I had my own speed figures for the Maryland tracks. I tracked every bet I made over a five-year period and found I lost 10 percent. I made a decision to find something that worked, something I could bet and make a living betting or get out of handicapping. In 1992 I decided to track trainer statistics. It took two years of cumulative information for the stats to take shape, to evaluate and bet.

I track the first, second, third and fourth races after a layoff separated by sprint and route. I have the same four-step procedure for claiming trainers as well as debut trainers. As I was recording I found certain trainers had the ability to win off the layoff or on the first race after a claim.

As an example, Ben Feliciano, a trainer in Maryland hits on first after a layoff in a sprint at a 31 percent win rate.

Howard Wolfendale, also in Maryland, hits at a 40 percent win rate on first after a claim in a sprint.

I started to look for trainers who carried at least a 30 percent win rate and had to have at least four wins to qualify as a bet. That is the name of my betting stat, the 4+30. I compiled a list of all the trainers who could hit with those two benchmarks. In 1994 I started betting specific trainers as automatics, and I have been betting almost the same group of trainers for the last 12 years.

Mark: Can there be an automatic bet based on conventional handicapping, or is it absolutely necessary to go against the grain of the usual variables?

Ed: I am not a fan of conventional handicapping like speed, pace, class and form. It never worked for me. As a handicapping tool those four things have failed miserably. The reason is horse racing is financed by recreational players. Fifteen billion is bet on the horses each year. If conventional approaches worked, we would be a nation of horseplayers. You would not be able to get a seat at a track or OTB.

There are two items needed to explain an automatic bet. The first is win rate percentage, the second is return on investment. I have never heard conventional approaches to handicapping explained with those two items.

Mark: When you talk about an automatic, do you mean 100 percent automatic, or is there still something judgmental in the mix?

Ed: To bet any automatic you have to bet it every time it shows up. To be profitable you have to bet it 100 percent of the time. All but three of my automatics are track specific, meaning how the trainer performs specific to today’s track.

As an example, Wayne Catalano at Arlington is an automatic on first after a claim in a sprint or in a route. Catalano has made 162 claims, won with 65 for a 40 percent win rate. His average win mutuel is $6.74 and his $2 ROI is $2.66 (a flat-bet profit of $.66).

The reality of betting automatics is not as neat as explaining its hit rate through a percentage. At Arlington the past two years Catalano has made 34 claims and won with 19 for a 55 percent win rate. Within those 34 claims he had a five-race losing streak. To get the benefit of Catalano’s hit rate you have to bet all his first-after-a-claim sprint or route horses, including a five-race losing streak. A five-race losing streak may not sound like a lot. Catalano's claims from 2005 to the present include about 30 weeks of racing at Arlington Park. With 34 first-after-a-claim races, he is producing 1.1 automatic bets a week, a win about every other week. That is why I have to bet all his automatics.

If human judgment is projected into the automatic bet, a lot of winners will be passed and the reason for the automatic bet is compromised.

Mark: Your automatics do change from time to time? Stock market investors have told me that getting off is always a tougher decision than getting on. What tells you to get off?

Ed: The only way I change the automatic is at the end of the meet. If the trainer cannot maintain his high win rate and profitability, I remove him from next year’s list. I do not take him out or stop betting him if he has gone 0 for 10. I bet all his automatics for the entire meet and give myself the opportunity to be profitable and that trainer the chance to maintain his hit rate and ROI.

Mark: Well, finally someone answers this question with a structured approach. Now, can you tell us of any bizarre or one-of-a-kind trainer bet?

Ed: Many trainers win in rare clusters. I keep a statistical eye out for them. Ralph Martinez is an automatic on opening day at Fairmount Park because in 2005 and 2006 he returned with seven first-after-a-layoff sprint horses and he won with six each year. Fourteen races and 12 wins for two opening days. I was ready for him this year. His wins paid $5.40, $4.40, $5.80, $3.60, $3.40, $5.00.

This is the only way to bet automatics. Perform all the work to find them, make a list, and then bet them all as a mechanical approach.

Mark: And how much information do you need in order to add a bet to the portfolio?

Ed: Every bet I make has to come from the first, second, third and fourth race after a layoff, claim or debut. I separate out their specialty move from those races. In addition to these applicable statistics, I look for subsets from this overall stat.

Every trainer move is a subset. One example so you get an idea. I track how a trainer performs with first time Lasix on third start in a sprint or second time Lasix on first after a layoff in a route. Or, what is a trainer’s win rate with a workout within three days or second after a layoff in a sprint?

Bobby Frankel has returned 227 first-after-a-layoff route horses that won their last race, at all tracks. Over the past 5½ years Frankel won with 80 of them for a 35 percent strike rate. His average win mutuel is $6.87 and his ROI is $2.42.

I look for information from which I can derive a 35 percent win rate or higher and I look for plays in the subsets as well as the applicable statistic. Over the past year and a half, Frankel has had 48 runners on first-after-a-layoff route/won last race. That’s a runner about every 11 days. Frankel won with 16 and that’s a win about every 34 days. Bobby Frankel raced 581 times in 2005 and so far this year he has started 319 races. That’s 900 starts or 1.59 runners per day. Only 5 percent of all Frankel’s runners qualify as an automatic bet. I get a win from him a little more than once a month.

Automatics are about a diverse group of trainers who do not show up every day. Automatics are about trainers who have a special uncommon and noteworthy purpose. That is really what I would label intent. They train determined to produce a win.

Mark: Experts like Barry Meadow insist that there can never be an automatic bet because no sample is large enough to become deterministic and everything will eventually break down. How would you respond?

Ed: Barry Meadow is a great handicapper and a true skeptic who doubts or disbelieves everything he hears about handicapping and making a living betting on the horses. Although I have never heard Meadow explain his approach to handicapping, he must believe something works for him but maybe not for other people.

I don’t like to bet anything subjective. I need to have a stat for everything. For example, 354,267 horses have returned from a layoff of 45 days or longer over the past five years and 37,630 won for a 10 percent win rate. If you bet all 354,267 horses, you lose your shirt. But within this negative stat, there are trainers with a positive stat. The naysayers refer to the negative overall stat but do not see the subsets.

I isolate a small group who perform at a high win rate. There are approximately 10,000 trainers who returned a horse from a layoff, and from these 10,000 I have 76 trainers who qualify as an automatic on first after a layoff. That means I will not place a bet on 9,924 of those trainers. I will bet only the 76. That’s just one of my subset categories.

Mark: What percentage of your bankroll is dedicated to automatic bets? Ed: I do not employ a percentage of bankroll approach. I am a grinder, a win bettor. I bet the same size win bet on every stat. As a win bettor you have to start with a bank that is only used for win betting. No expenses come from the bank, including personal or business expenses. That bank has to be large enough to cover your longest losing streak, in my case, 32 in a row. The longest losing streak betting automatics has been 20 in a row.

Mark: In my own flat-bet profit research, I find that most automatically profitable plays have a relatively low hit rate, between a 12 and 18 percent win rate. But that means that losing streaks will be inevitable. How do you know whether your factor is simply no longer valid or if it’s passing through a normal low-performance period?

Ed: I start with a win rate of 35 percent and higher. This is a long-term hit rate. I end up with a 30 percent win rate. I will have losing streaks. Some trainers will not make next year’s automatic list. There will be some trainers who will be on the automatics list for the first time and some who were taken off the list that may make it back on. A 10-race losing streak is not unusual, and there has been one 20-race losing streak from the automatics in the last 14 months.

If you have a 12 percent or 18 percent win rate you can expect a 30-race losing streak from every 100 bets. That is a stressful approach to handicapping. I could not bet on the horses with a 15 percent win rate. That is too much stress and I would have difficulty in thinking and concentration and eventually become discouraged and quit racing.

Mark: Can automatic bets help avoid burnout. It seems that most good players have their half-life, and then seem to wind down after so much intensity.

Ed: Yes, automatics can and do help avoid burnout. I make three automatic bets a day. I make about 1½ bets a day from the 4+30. There are days I am relieved the automatics are doing all the work.

The reason for the automatics is to reduce the amount of mistakes from over-handicapping or over-evaluating a bet. The longer I go without a win the more mistakes I would make and the more I would agonize over my selections when they lose. Time spent between cashing bets is a problem. It becomes a confidence issue. That is why I do not let any trainers on the list below 30 percent and most are above 40 percent.

Mark: My thanks to Ed for going deep into what he does. No one will ever criticize Ed Bain for lacking the work ethic. By isolating the first four races following a layoff or debut, Ed is assured a lightly raced horse, already an advantage under most racing conditions. By isolating claimed horses, Ed embraces the projection of change, another built-in advantage in a pari-mutuel where most people are betting on continuity.

Readers can learn more about Ed Bain’s work at www.edbainslayoffsandclaims.com.

Mark Cramer’s one-of-a-kind book of racing fiction, Scared Money, which sold out in its first edition, will be re-released by DRF Press in September. Visit his Web site: www.altiplanopublications.com.