Thursday, October 22, 2009

A Eureka! moment for dognostics, perhaps: If the numbers don't convince you to change your betting habits, try simple logic instead.

_
Skepticism about my by the numbers underdog betting concept is on the rise, led by die-hards with years of experience of backing teams that inside knowledge says have the best chance of winning.

Of course I have no quarrel with that. Why would I? Educated, smart betting has been my mantra for as long as I can remember, and I am in awe of anyone who takes the process seriously and sticks to a rigid set of rules.

As I commented to one doubter recently, my problem is that I barely know a baseball bat from a banana or a football from a flying fish, so all I have to guide me is "the numbers."

One thing we can all I agree on, I hope, is that the great majority of sports book punters put their faith in and their money down on favorites (one pro at Harrahs Stateline the other day said at least 85% of his customers bet that way).

Favorites are called favorites because most people favor them. Makes sense. And of course it all starts with handicappers who look at all the factors that a smart bettor considers, and use their knowledge and good judgment to come up with odds that are a distillation of all that wisdom.

Thing is, with so much money pouring in on favorites, the bookie biz would be in big trouble if the team with the shortest odds won every time.

Favorites have to lose for the bookies to make money. And the only way that can happen is for underdogs to win.

One thing I am still chewing on, and have yet to find a satisfactory explanation for, is the way odds on favorites and against dogs so often break what I was at first assured were hard and fast rules regulating the size of the rake that bookmakers take from each sporting event.

I started out with major league baseball, so let's look at that.

In theory, bookies take a "dime a time" so that if the odds on the favorite are -140, meaning that a $140 bet will earn $100 if the team wins, the odds against the underdog will be a "dime" shy of the exact opposite, or +130 (a $100 bet wins $130 instead of $140).

In hockey, I see, it's a "double dime": -130 on the favorite usually means +110 on the dog.

In reality, a comparison of odds applying to favorites and underdogs, especially when the sample runs into thousands of games, reveals repeated instances in which the bookies shave far more than a dime or two off each side.

Why, I wonder? It can't be greed, surely? Or sleight of hand? Perish those thoughts!

All I am trying to say here is that underdogs must win in a substantial number of contests in order for the bookmaking industry to make big money.

So betting against the mob (the 85% mentioned above, not the mob with a capital 'M'!) seems like a sensible plan to me.

I have started to track all the betting action on certain days, to see what "woulda" happened if I had backed favorites instead of underdogs, and already have several instances where more winning picks than losing ones still left me in the hole, while the reverse woulda been (or was) true for dogs.

This is all very interesting.

No one out there who believes in studying all the different factors that come together to make a team a winner should be insulted or offended by what I am proposing.

It's just another betting option, is all - one that has a lot of appeal for someone like me who, until recently, paid no attention to any sport unless one of my kids or grandkids was out on the field.

That meant soccer, mostly. And I see the bookies take bets on that!

Betting update: A mixed NFL bag on Sunday with two wins and three losses (+2/5), a good Monday (+2/3), 0/2 Tuesday, then +2/3 again Wednesday. Today's picks are all hockey dogs: Boston Bruins +145, Dallas Stars +130, Phoenix Coyotes +115, Tampa Bay Lightning +135 and Atlanta Thrashers +110. If the two shortest dogs (Chihuahuas?) come home, I'll almost cover all five bets, and three winners will put me ahead. The same could not be said if I bet against these teams at -170, -145, -130, -155 and -135. Target betting does not need to me nearly as aggressive against even+ payouts as it does against table games, and that's a good thing.

An important reminder: The only person likely to make money out of this blog is you, Dear Reader. There's nothing to buy, ever, and your soul is safe (from me, at least). Test my ideas and use them or don't. It's up to you.
_

Tuesday, October 20, 2009

What are the bookies (or the handicappers) saying to us when they make a team a 1 to 100 favorite? Only lunatics need place a bet?

_
With my 2009 MLB database only a few hundred games short of completion, I have become used to what seems to be a spin-the-bottle approach to the critical task of setting the odds on baseball games.

One day, a club is as sure-fire a winner as it can get, at least according to the odds. And less than 24 hours later (sometimes against the same opponents!) it's a long shot.

I have barely made any progress with my college football database, and am tempted to abandon it because of a range of odds that defies explanation.

In baseball, with its 30 pro teams in 15 pairs and its relatively similar ability and performance range, dogs start at even money, more or less, and occasionally go up to 3.5 to 1 (+350).

As all you experts already know, that's not the case with the motley rabble playing under the NCAA banner.

Is there anyone crazy enough to put money down at -10000? Maybe, but even for the most daring of gamblers, that's a lot of risk for almost zero return.

Not much reaction to my "dog project" so far, but I have evidently ruffled a feather or two by seeming to suggest that a careful analysis of teams and players, game locales and conditions, wind speed and chances of precip, is redundant.

What I am actually suggesting is not coming from me at all: It's the numbers that tell us that in the absence of expert knowledge and a love and understanding of the game (whatever sport it might happen to be) betting the odds can be very effective.

Put another way, no one should let common sense and good judgment get in the way of making steady money betting on ball games!

"Doggy power" is so pervasive not just in baseball but in pro-football and hockey and in baskeball too, I'll bet, that I will make a prediction or two ahead of the completion of my databases.

My guess is that when all is said and done, it won't matter how you shake up the stats...chasing dogs will always be a winner, not necessarily betting flat, but certainly with target betting applied.

I heard an interesting baseball snippet on NPR the other day, on my way up to the casinos in Stateline. The Washington Nationals were quoted as easily the worst team in the MLB lineup, but somehow they managed to win 59 of the 162 games they played in 2009.

That's a fairly crappy win rate, even for the team at the bottom of the heap. But with the odds-makers routinely rating them at +250 and above, backing them as underdogs throughout the season might have made a tidy profit at, say, $1000 a pop.

Let's do the numbers, as they like to say on the radio show that caught my ear. Those 59 wins amounted to 36% of all games played, or one win per 1.75 losses. Bad outcome if the odds on the Nationals as bottom dogs averaged less than +175, but I very much doubt that was so.

From what I have seen so far, I'll wager that if you applied pretty much any selection criterion, backing your picks only as "dogs" throughout the season, you would end up making money. Not a lot, maybe, but probably more than the expert laying down his dough on the team as favorites.

How about backing only teams playing in the mid-west time zone? Or teams with left-handed batters? Or the same team all season? Or only teams with the letter "Y" in their name? Or...

You get the picture. The dog factor is so reliable that you can count on it to bring home the bacon no matter what you do.

Watch this space!

Wednesday, October 21:

OK, time for me to eat a little crow, courtesy of the redoubtable Washington Nationals!

Even though the databaseball is about 900 games short of the full 2009 regular season schedule, formulas are already in place to analyze each team's performance.

The Nats were already on track for an embarrassing year when they bungled their way past the 100-game mark.

They lost 67 of 102 games and were favorites 15 times last season (against their own grounds-keepers, perhaps?). As favorites, they won 5 games, which is eerily consistent with their overall 33% win rate.

A dog-lover would have put money on the Nats 85 times during the season, winning 29 of those wagers for a total flat-bet take of $4,537. He would have lost the remaining 56 bets (66%, which is becoming an awfully familiar losing number in this scenario!) at a cost of $5,600, leaving him $1,063 in the hole for the year.

Bummer! That means that while my optimistic prediction in yesterday's post might hold true for any MLB team that lost more games than it won, the stats have to draw the line somewhere and the awesomely awful Washington Nationals are under that line!

Plug in, say, the L.A. Dodgers, and the picture brightens greatly:

Los Angeles Dodgers

111 Games 30 Bets
67 Wins 60%
44 Losses 40%

Won D 17 2067 $122
Lost D -13 -1300 $767

Dogs 33 17 57%
Max U 175
Min U 109

Fav 78 47 60%
Min F -260 0.38
Max F -104 0.96

Won F 47 3135
Lost F 31 -3100 $35

The dog fancier could only make 30 bets on the dodgers in 111 games, keeping in mind that the MLB database is not yet complete.

The Dodgers had a win rate of 60% in the sample to date, and consistent with that WR, 57% of the "dog" bets (17 wins) brought in $2,067 offset by losses of $1,300, giving us an end of season (EOS) profit of $767 based on flat bets of $100 apiece.

To me, the most critical piece of information above comes in the last two lines of the summary: 47 wins as Favorites would have brought in $3,135 in the season so far, and losses as Favorites would have swallowed almost all of that ($3,100) leaving $35 in folding green to show after 78 bets at $100 a pop!

Ouch!

So...the dog-devotee nets $767 or a whopping 26% of his ($3,000) action, while the fave-follower sees $35 back, or 0.45%.

Food for thought, I reckon!

My faith in the Washington Nationals may have been misplaced yesterday, but the bottom line superiority of dogs over faves remains unsullied!

An important reminder: The only person likely to make money out of this blog is you, Dear Reader. There's nothing to buy, ever, and your soul is safe (from me, at least). Test my ideas and use them or don't. It's up to you.
_

Wednesday, October 14, 2009

Remember the giddy delights of day trading on Wall Street? Welcome to a safer, more profitable alternative: dog trading!

_
Quite a while back, I was introduced to a steel-hearted amateur day trader who claimed he was making north of $90,000 a month buying stocks online and dumping them the moment they reached either a win or a loss target.

If he's still at it, I doubt his current profits add up to more than $1,000,000 a year, as they did back then.

And my guess is that like most people left with money to gamble, he is sitting tight and waiting for the financial sector to quit leaking red ink all over his quixotic portfolio.

Meanwhile, my "dog" project (which I can claim as my own since it does not remotely resemble what my friend Pete suggested, and he has in any event either abandoned sports betting or opted to keep me in the dark!) remains a work in progress.

It's work because the task of building a full-season MLB database, then doing the same for college football, is time-consuming and tedious drudgery.

But I can't complain, because every time I plug in a new set of verifiable outcomes, I can see astonishing progress in the application of target betting rules to wins that always pay better than even money, and losses that are, in a sense, controllable.

Obviously, we can't know the final outcome of a given game in advance. But because we are not backing favorites, we are never tempted to "risk more to win more" and recoveries are greatly aided by returns that can at times exceed 2 to 1.

The chart below looks suspiciously like those reportedly distributed to his victims by the world's first $50 billion conman, Bernie Madoff.

(And it still seems strange to me that sane people would believe his reports while the market was bouncing up and down like a 300lb linebacker on a trampoline!).

The more frantic graph below it tracks flat-betting $100 on the exact same set of "dog" selections.

Both suggest (but do not prove, remember!) that avoiding favorites altogether when you venture bets in a sports book is a much smarter choice than stampeding with the herd in the direction of "certain" winners.



Big-hearted and perhaps foolhardy as I may be, I am not about to go public here with details of my bet selection process.

But anyone with access to a complete database of 2009 season results and a couple of active grey cells should be able to figure it out in an hour or so!

I am putting all this time and effort into building complete and accurate sports databases because the spreadsheet platform offers virtually unlimited opportunities to ask pertinent questions and get answers that mean something.

The "all" claim in the chart captions above is not yet accurate, because I have learned that even with 1,316 games keyed in, I am barely halfway through the 2009 MLB season.

So far, querying the "databaseball" file has come up with several surprises, none of them hard to take.

What I see already is a real prospect of creating the "Pari-mutuel Fund" I mentioned in an earlier post, keying minimum bet values to the bankroll as it fattens from a combination of wins and investments.

Again, this is not the place for details, but try to guess what would happen to the sky-reaching green line above if the $100 minimum bet now in place were to be stepped up in $25 increments on a time line or as the bankroll grew.

I could run that test in about five minutes, but I am more than happy with the results seen here and don't want them to seem even more like an extract from a Bernie Madoff portfolio!

Comments I have received so far have been underwhelmingly positive, and that is to be expected.

Even to me, this line of research sometimes seems too good to be true.

If this were possible, said one comment, someone would have thought of it before. And since I have never claimed to be a mathematical genius (unlike some of my longtime critics!), I'd say that's a fair bet.

But would a well-funded sports book investor using a long-term strategy like this be foolish enough to publish it? I doubt it!

Another tentative comment refrained from doubting the accuracy of my projections, but suggested that "bookies won't allow it" if I am right.

How could they stop it, I wonder?

Target betting against casino table games faces an uphill struggle less because of "the numbers" than as a logistical consequence of house interference and the resultant need to stay almost constantly on the move.

Those issues are not relevant to my "dog trading" plan.

Bets will get high from time to time, but their average value in the model to date is a few cents less than $150 and the return on overall action (the "hold" as they call it in a casino pit) is an astounding 12.6%.

Risk is far less than for blackjack or baccarat (-$2,425 so far).

And the bonus element (pay-outs exceeding even money) make such long-established target betting elements as a wide spread, a win progression and limiting losses in a downturn not just unnecessary but counter-productive.

Sports bookies pay little or no attention to bet values or patterns, as long as any "local" limits are not exceeded.

And on the rare occasion a bet that's likely to raise eyebrows is required, it can easily be spread across two ore more books.

The model from which my current numbers are drawn sets the maximum bet at 5x the minimum ($100 for now), and that's a fraction of the cap applied in sports books in Nevada and online.

I should remind everyone that the results seen above are not "real" because they assume that each bet was placed consecutively rather than in blocks.

But statistically, that is not a relevant factor.

When I am done with the MLB database (in about a month, if I'm lucky!) I will break out the "dog" bets into multiple lines as I did before, and demonstrate that the combined profit will confirm what you see above.

Then I will pull together college football odds and outcomes, confident that what I have discovered so far is not purely a baseball phenomenon.

The current NHL, NBA and NFL seasons will provide me with fresh data day by day, and by the time the boys of summer are back in action in 2010, I will have a lot more data to work with.

Maybe before then, I will dig back in the archives for odds and scores from 2008-2009.

Then again, maybe not. My typing fingers (all three of them) are numb from the work I have done so far.

But as I said, I'm not complaining. Or only a little...

An important reminder: The only person likely to make money out of this blog is you, Dear Reader. There's nothing to buy, ever, and your soul is safe (from me, at least). Test my ideas and use them or don't. It's up to you.

Monday, October 12, 2009

Am I in a dog daze? The numbers from three different sports all say loud and clear that I'm barking up the right tree! Roll on NBA 2009-10...

_
Somewhere in a sneaky sub-clause in the Law of Large Numbers is a rule that says that the effects of an apparent but illusory trend will, if it has no merit, always diminish as the size of a data set expands.

Within a couple of days of accepting my friend Pete's challenge to adapt target betting to sports book wagering, I suggested to him that following tipsters' picks and therefore putting money mostly on favorites might not be the best way to go.

And I also urged him to stop placing bets in blocks, instead relying on multiple lines or linkages that would enable him to set win and recovery targets based on prior outcomes.

Pete chose to ignore my advice, and since he was betting his money, not mine, it was his right to shrug off my dire three-word warning: You. Will. Lose.

Since then, I have analyzed almost 1,500 actual games, mostly from the MLB season just ended but with season-to-date numbers from hockey and pro football making an important contribution.

And my advice to Pete and anyone else who might be more receptive than he is remains the same.

Forget favorites, at least until your bankroll has some serious heft behind it.

And depend on target betting, the most effective method of money management known to betting men, to keep you out of the red.

I have already posted MLB data that shows target betting profiting day after day in spite of the fact that, as we would all expect, underdogs are more likely to lose than favorites.

The method prevails if we bet ALL "dogs" on a given day. But who would want to do that?

Instead, I used a middle-ground selection process to cherry-pick betting options solely from the odds offered at sports books in my Nevada neighborhood.

Final results for games played out before I started all this with Pete at the end of July were available to me, of course, but I did not start keying them into the "databaseball" until I had made all my selections by the numbers.

By season's end on October 4, I had 355 "cherry dogs" (a mixed metaphor that I'll improve on one of these days!) consisting of 172 winners and 183 losers.

The overall win for target betting woulda been $9,120 or 13% of total action of just over $72,000.

Nice, but not such a huge surprise, given that "dogs" almost always return better than even money on a win.

For the record, flat-betting $100 on each of my cherry-picked underdogs would have brought in a profit of $2,770 on bets totaling $35,000, an overall outcome of +7.9%.

I ended up cherry-picking just 153 favorites, and found even that many a chore simply because I could not see the sense in risking a hundred bucks to win less than $100 in return.

I was wrong.

Flat-betting my choices at odds consistently shorter than even money, risking $100 each time, woulda LOST $325 overall in spite of the fact that favorites trounced underdogs more often than the other way around (81 wins vs. 72 losses).

Target betting told a happier tale: +$1,597 from action of $34,732 (+5%). Better than losing, I'd say.

It's early days in the NFL and NHL seasons and the NBA schedule is still a few weeks away, but here's what I have so far.

NFL "cherry dogs" (35 winners out of 61 games) +$2,430/$7,630 = +32%.

NFL "cherry faves" (14 winners out of 18 picks to date) +$720/$2,080 = +35%.

NHL cherry dogs are not doing so well after 40 season-to-date games and are $65 in the hole (-1%) with an outstanding LTD of $910. That sad state will not last for long, I can guarantee...

NHL cherry favorites are doing even worse, with 12 picks down a total of $885 (-49%!). Again, the season is young, and the picture will change to black numbers soon enough.

I owe quite a debt to a punter known to me simply as "Dave," a regular at the sports book at Harrahs Stateline (a pale shadow of its former self since the parent company's feeding frenzy swallowed up neighboring Harveys along with almost every other casino in Nevada!).

Dave pulled out a pen and scribbled a few numbers on the corner of my day's NFL schedule: "Dogs 51.5%, faves 48.5%." I had no idea then what "dogs" were, but Dave quickly put me straight and added, "Favorites will break your heart."

From then on (and this was maybe two days into Pete's sports book project) I weighed the advantages of larger but less frequent wins against a greater number of smaller pay-outs, and have been looking for Dave ever since.

There's a guy who really knows what he's talking about!

I mentioned Dave's doodle to a sports book manager in Carson City, and his response was terse: "That's crap!"

From what I have seen so far, I don't think so.

From a bookies' standpoint, of course, business is best when most punters back the short shots, in spite of the standard industry claim all wagers are welcome.

The other day, a couple of NHL contests were so heavily weighted that it took a $1,200 bet to win $100.

Neither favorite lost, of course (the scores were 44-7 and 33-14).

But they gave the bookies a lot of action for very little risk, and every so often in games like that, Christmas comes in the form of an upset, and Brinks has to send an extra truck to haul the money to the bank!

An important reminder: The only person likely to make money out of this blog is you, Dear Reader. There's nothing to buy, ever, and your soul is safe (from me, at least). Test my ideas and use them or don't. It's up to you.
_

"Whatever happened to your blackjack experiment? Guess you crashed and burned and don't want to talk about it!"

_
Good question!

And it gives me just the kick in the pants I need to wrap up a few loose ends before devoting the next few posts to sports betting without fear of interruption!

I abandoned the "one-thumb" blackjack game on my trusty iPod with my win cut back to a healthy "$386,489.75" (beats me how you can get a 75-cent return on a $5 minimum bet, but that's the number I'm left with) because after more than 20,000 rounds, the house edge remained north of 3.0%.

Twenty thousand rounds!

Seems like I have way too much time on my hands. Or thumb.

But the experiment was worth the multiple hours I spent on it because it showed what target betting can achieve, even against a blackjack game far tougher than anything found in a real casino.

The undoing of my seemingly relentless race towards a cool million bucks was my application of a stop-loss forced on me by the reality that the iPod game imposed a 1 to 1,000 ($5 to $5,000) spread limit.

As my (handful?) of regular readers know, a wide spread is essential to long-term success against casino table games, and 1 to 1,000 is at the bottom end of the recommended range.

The whole purpose of spread limits is to cripple a losing player's chances of winning his money back, not (as many house-trained mathematicians claim) to limit the casinos' liability.

And that last sentence is not a contradiction, by the way!

Casinos know that 999 out of every 1,000 players bet in a way that is essentially random, most of them limiting themselves to a spread as tight as 1 to 5, with an occasional looney-tune well-heeled enough to attempt to recover all his prior losses in one or two extra fat bets.

A lot of high rollers think of themselves as being too rich to lose, and some of them are able to "buy the pot" over and over again before the house edge leans extra hard on them and cleans them out.

The casinos know that most gamblers on a losing streak are either too lazy or too excited to step away from the table when the game goes against them.

So table limits serve to rein in the potential benefit of a short-lived player winning streak, making recovery all but impossible.

A fat cat loser with big money behind him stands a much better chance of getting back in the black if he stops betting when he bumps up against the table limit, and goes looking for a game with much higher limits.

But that would be contrary to the nature of most humans.

And if there is one thing casinos know a whole lot about, it's human nature.

After twice pushing my win past $700,000 on my iPod blackjack game, only to fall foul of the 50 max bet (50 x $5,000 = $250,000) stop loss I set myself, I figured it was time to move on.

There was no higher-rent alternative for me!

In real play, I would have faced an overall house edge of less than 1.0%.

The iPod game, which even in its revised form breaks several real-play rules, had 9841 player wins, 10572 dealer wins and 1763 pushes.

That's a house edge of 3.3% with pushes included, or 3.6% without them...and neither figure shows any respect for blackjack reality.

I have to say that I am surprised that any self-respecting game developer would create an application that falls so far short of honesty and accuracy.

But MobilityWare is not the only culprit.

I fooled around for a while with a nifty-looking iPod version of 3-card poker, and was amazed to find that it allows "the house" to rake in the player's bet on a push.

No, no, no!

That might sound like a trifling error to some of you.

But when a maximum bet is on the line, theft is not an option that casinos should be permitted. Beating an honest game is tough enough.

Three-card poker has a real house edge of close to 7.0%, making it virtually impossible to beat in the long run even without the standard 1-40 spread limit ($5 to $200) imposed in most Nevada casinos.

It is, by any standard, a blatant sucker trap.

But it can be both profitable and entertaining in the short term, so long as any losses are transferred to safer games such as blackjack or baccarat when recovery is thwarted by those piddling table limits!

I admit I am disappointed that I didn't get to see my one-thumb blackjack win click over to seven figures.

But a big part of long-term success in casino betting (never, ever, call it gambling!) is knowing when to move on...

And my enthusiastic interest in sports book betting is all about moving on without abandoning casino table games entirely.

I urge you to come back often to check on my frequent updates on that topic.

For now, I'll simply look back to yesterday (Sunday, October 11, 2009) and the fate of the five NFL picks I posted last week.

As with the hockey picks (not pucks!), three of the five went south and two brought home the bacon.

Using the odds that prevailed when the bets were placed, the two winners brought in $275 and the three losers cost $300, handing the bookies a flat-bet profit of $25 on action of $500 (5.0%).

But target betting is not about flat betting.

My next post will deal with the MLB database, up from 700 or so baseball games to more than 1,100 with at least 1,000 more to come from the 2009 schedule.

It will show that target betting can be used to win steadily on both underdogs (the best bet) and favorites, day after day, month after month, sport on sport and season after season.

Flipping yesterday's NFL outcomes to three winners and two losers, backing the favorites all the way, we'd get flat-bet wins totaling $300 and losses of $315.

Oops!

As I often say, long-term profitability means winning more when you win than you lose when you lose, because chances are you will lose more bets than you win.

Here's a betting opportunity that permits you to lose even when you pick more winners than losers.

No wonder sports betting is so popular...it's a masochist's dream!

An important reminder: The only person likely to make money out of this blog is you, Dear Reader. There's nothing to buy, ever, and your soul is safe (from me, at least). Test my ideas and use them or don't. It's up to you.
_

Saturday, October 10, 2009

Sports book rule No.1: Never place a bet that pays less than even money. No.2: Don't be afraid of longshots (sometimes they actually win!)

_
My MLB database, which I have never claimed to be definitive, representative or anything else-tive, confirms everything I have been saying since Peter Punter and I first started working together on this sports book project.

And while it is inevitable that someone out there will accuse me of the baseball equivalent of Monday morning quarter-backing, the truth is that I have been ruthless in keeping my analysis of the numbers objective and uncontaminated.

Translating my protocol to real life would result in a scenario that goes something like this:

AFTER YEARS of frustrating ups and downs playing table games in casinos all over the USA, Bob Bettor (Pete's first cousin) decides there has to be a better way to make a living as a gambler, and for the first time in his life takes a look at the sports book.
HE KNOWS ZIP about baseball, basketball, hockey or football (or golf, horse-racing, NASCAR or underwater motorcycling), but he's been a numbers guy all his life. So he can look at any day's lineup of sports events and assess odds without getting all emotional about teams, players, pitchers, coaches, stats or the difference between home and away.
FIRST OFF, he recognizes that while backing favorites might result in more winning tickets than losing ones, the odds offered on the likeliest winner in any match-up are too short to permit a long-term profit.
THE NEXT potential hazard to pop up is that flat-betting the same sum every time is unlikely to succeed in the long run, and enabling a variation of target betting by placing just one wager a day is certain to be fatally boring.
BOB SUSPECTS, but can't be certain, that maybe the best way to go is to select at least four underdog prospects in each sport every day, splitting them into separate lines, series or streams, and using target betting to determine each day's wager values.
KEEPING IT SIMPLE is essential, because he won't be able to take his laptop into the sports book every day (he doesn't know why, but that's the rule), so he settles on four lines a day, and calls them Green, Blue, Purple and Yellow.
HE ZEROES IN on baseball as his sole sports interest for the time being, and places his first four bets on August 12, 2009, betting the same way every day until the MLB season officially winds up on October 4.
BOB CAN'T BE SURE yet, but he reasons that "dogs" with relatively long odds (more than $140 paid on every $100 wagered, for example) might not win often enough to justify a large bet on them, so he looks for odds in the even-money to 1.3 to 1 range whenever possible.
HIS LONG-TERM AIM is to quit playing blackjack, baccarat and other casino "house games" and move his money to a milieu where no one will be looking over his shoulder or tracking his play, and what other players do can't hurt him.
HE'S NOT WILDLY OPTIMISTIC, but then Bob never is. He just hopes that this time, Cousin Pete has come up with a viable way to make money in a casino without having to endure smoke, spilled drinks and dim lighting. Hell, he has sports books by the dozen within a half hour's drive from his front door, and he can get a whole day's betting squared away by mid-morning!
IF BETTING "DOGS" proves profitable, the sky's the limit on bets, because even though small books have low caps on wagers, an occasional large bet can easily be spread across as many tickets as it takes to get it covered.
THERE'S A CHANCE, he thinks, that sports betting could turn out to be more reliably profitable than any other form of "speculation" he has tried in a lifetime of bucking the odds (and that especially includes ventures involving Wall Street and/or foreign exchange!).
PROFITABILITY could further be enhanced by creating what might (for the sake of a good pun) be called a Sports Betting Pari-Mutuel Fund, with at least half of all winnings being ploughed back into the bankroll every day, enabling the minimum bet value to grow proportionately. More about that some other time!
THE QUESTION that has yet to be answered involves the outside chance that there is something unique and unrepeatable about 737 MLB outcomes that I used for my first database (731 bets in all when cancellations were deleted). Do 731 games constitute a "representative" sample? It depends whom you ask...

To finally cut to the chase, all four of Bob's lines ended up making money for him in spite of the fact that less than half of their combined 216 picks were winners.

The GREEN line won $1,330 or 15.61% of total action, and had a loss to date (LTD) of $195 unrecovered and therefore "pending." GREEN went just $100 in the hole before recovering!

BLUE won $1,845 overall (13.95%) with a LTD of -$405. BLUE dug a hole $1,160 deep before its dogs got back on track.

PURPLE was $690 or 5.5% ahead, with no LTD left hanging, and exposure of $2,765.

YELLOW had a final win of $695 (7.45%) after completing a recovery, falling $1,345 at one point before being rescued by its slate (sled?) of dogs.

The four lines COMBINED had a win of $4,563, with the GREEN and BLUE LTDs of $600 awaiting turnaround.

Flat bets of $100 on the 216 selections "woulda" achieved a win of $1,188 or 5.5% of the four-line combined action. The maximum exposure for the flat bettor was a skimpy $160, presaging its relatively modest overall profit.

For those of you who prefer visual aids, here's what the MLB results look like expressed in colors and wiggly lines:


One handy thing Bob and I learned from this exercise is that we were both wrong about "long dogs," or dachshunds, as we now think of them.

The range applied for the test described here was from even money (+100) to 1.4 to 1 (+140), with an occasional exception.

With bets limited to four a day, that gave us an overall flat-bet win of 5.5% in spite of there being more lost dogs than won ones.

Removing the four-a-day limit and betting every dog within the range would have resulted in 459 bets from August 12 through October 4, with 188 winners (41%) and 271 losers.

In spite of the fact that most dogs didn't have winning days, flat-betting $100 a pop would have brought in a 54-day profit of $1,750 or 3.8% of total action.

Making +140 the lowest acceptable number instead of the highest, and placing a bet on any dog rated +140 or higher "woulda" reduced the number of wagers to 273, leaving the win percentage pretty much unaffected at a paltry 40%.

The happy surprise is that the win value woulda jumped to $3,520 or a hefty 12.9% of total action, thanks to win payoffs ranging from 1.4 to 1 to 3.1 to 1.

Dogs rule!

As I see it, the bad news is that I now feel obliged to key in outcomes and odds from the entire 2009 MLB season, then moving on to NFL and NHL stats.

That's a lot of work. But worth every minute, I suspect.

An important reminder: The only person likely to make money out of this blog is you, Dear Reader. There's nothing to buy, ever, and your soul is safe (from me, at least). Test my ideas and use them or don't. It's up to you.
_

Friday, October 9, 2009

How can you win in the end when you win more often than you lose but you win less when you win than you lose when you lose?

_
Two winners out of five bets isn't cause for popping corks, but what would have happened if I had picked the favorites instead of the "dogs"?

My by-the-numbers picks lost me $300 but softened the blow somewhat by returning $265 in wins, leaving me $35 in the hole on the night.

Reversing the bets would have reeled in wins totaling $300, soured by two losing bets that cost me $305. Net loss, in spite of the fact that I had more right picks than wrong ones: $5.

NEW YORK ISLANDERS +140/Ottawa Senators -160
PITTSBURGH PENGUINS +125/Philadelphia Flyers -145
CHICAGO BLACKHAWKS +135/Detroit Red Wings -155
ATLANTA THRASHERS +140/St. Louis Blues -160
COLORADO AVALANCHE +115/Nashville Predators -135

This little exercise does not prove anything, but it serves to illustrate why I had powerful reservations when my friend "Peter Punter" first came to me with the idea that "cappers" held the key to making steady profits at sports betting.

Pete pointed to tipsters' win rates of 56% or better, and reasoned that a "game" with an overall positive expectation rather than the negative expectation that we are all used to when we risk money in a casino had to be the perfect platform for target betting.

Two months later, I am still not sure Pete's wrong.

But I don't think backing favorites, which tipsters tend to do to goose their win rate, is the way to go.

Favorites win more often, probably, but when they crap out, they cost a lot more money than underdogs do.

Worse, when they win, they pay less than even money, resulting in the numbers seen when final scores were in for my five NHL picks for Thursday, October 8.

From July 26 through October 4, Pete kept me posted on all his bets, giving me the opportunity to apply postmortem tests to see what coulda, woulda, shoulda happened if a variation of target betting had been applied to his picks from Day One.

The first problem was that from the far side of the world, Pete was stuck with having to make all his bets at once on any given day, with some of the events playing out simultaneously while he slept.

So it would not have been valid for me to take all the outcomes, and run them through a target betting model as if the bets were consecutive.

Almost from the start, I recommended that Pete split his wagers into multiple lines, perhaps limiting himself to five picks each day, and linking each first bet to the next day's first bet, each second bet to the next day's second bet, and so on.

Still and all, target betting's usually stellar performance would have to be hurt by the less than 1-to-1 return on the majority of the cappers' picks.

Pete chose to go a different route, betting as little as $20 when he did not have much confidence in a particular tipster, and more than $3,000 when he did, and using a cancellation system to set his wager values when he was not sure which way to go.

Today, I took all of Pete's sports betting data and separated most of the bets into four lines or series linking first, second, third and fourth bets each day.

Target betting did not cover itself with glory in this experiment, coming out of it after 158 bets with a final win of $800, or 1.75% of its overall action of $45,610.

That was disappointing, given that in spite of all the rude things I said about them in my e-mail exchanges with my friend Pete, the "cappers" managed to deliver an overall win rate for those 158 picks of 56.33%, providing a positive expectation of 6.33%.

Pete did not do so well. His actual result from 158 bets placed during a two-month trial run was a loss of $6,611 or -11.37% of total action of $58,168.

So at the end of the day, more right bets than wrong ones could not keep Pete out of trouble, precisely because of the higher cost of losses and lousy rate of return inherent in backing favorites.

Yes, favorites win more often. Usually.

But target betting depends on winning more when it wins than it loses when it loses to overcome the effect of negative expectation. And more wins than losses is not a whole lot of help when wins average $100 and losses cost $150 or more apiece.

I am posting all this in advance of running an identical four-bets-a-day test on outcomes in my baseball database, which has 737 games played from August 12 through October 4th.

To keep the process objective, I will dump all of the "dog picks" that I actually made while evaluating Pete's sports betting proposition, and will simply select the four underdogs offering the shortest odds above even money each day.

It should be interesting...

An important reminder: The only person likely to make money out of this blog is you, Dear Reader. There's nothing to buy, ever, and your soul is safe (from me, at least). Test my ideas and use them or don't. It's up to you.
_