I want first of all to revisit the core premise that evolved when a friend persuaded me to travel to the (to me!) alien world of sports betting.
Since I first learned to shuffle a deck at the age of eight, I have been a card player, and occasional forays into horse racing after I hit my late teens failed to fuzz my focus on casino betting as the most reliable way to stay ahead of the game.
My pal "Peter Punter" contacted me early last fall with a request that I "do the math" on an idea he'd had that centered on the high win percentages claimed by some professional handicappers ("cappers") in the field of sports betting.
Pete reasoned that casino games - his favorite poison is baccarat - have a predictable and inescapable negative expectation, while top cappers can prove a year-round success rate of 70% or more.
That amounts to a player edge instead of a house advantage, and that has to mean money in the bank, said Pete.
My immediate response was to pour cold water all over the notion.
Handicappers pick mostly favorites to boost their chance of being right substantially more than half the time.
And since favorites almost always pay much less then even money on a right pick, the book is bound to make a long-term profit in spite of losing more bets than it wins.
Let's assume that favorites win 60% of 1,000 games and that the average return per win is 65 cents on the dollar.
The book would then collect $400 on 400 punter losses at $1 apiece and pay out 600 x $0.65 on wins: $390.
Add a dose of reality to those best-case numbers and it quickly becomes clear that my pal Pete had been gripped by a truly lousy idea.
He's not a guy who likes to hear the word NO! and he countered with confidence that surely target betting could overcome the bookies edge.
As things turned out, he was right.
But in the end, Pete went with some cockamamie progression that had nothing whatever to do with win targets, and he dumped thousands of dollars into an ever-deepening hole.
I, in the meantime, rejected the irrelevant win claims of pro cappers (who make a lot more money for the bookies than they do for bettors crazy enough to buy their picks), and opted instead to take a close look at underdogs.
Underdogs lose more often than they win, obviously...otherwise they'd be favorites.
But they also pay, at worst, even money on every win (assuming we never bet at odds below +100 or 1-1), and paybacks are routinely considerably more than 100%.
I began to build a database that has now topped 7,500 games from baseball, pro football, basketball and hockey schedules going back to the fall of 2008.
Backing ALL underdogs (including teams that didn't have a snowball-in-hell's chance of winning) we woulda had a win rate of 38%, giving the book an overall advantage of...24%!
With those numbers, we woulda needed an average payback of 62/38 = +164 just to break even.
And, of course, that didn't happen.
The overall average odds for underdogs for all 7,525 contests were +202, or a hair better than 2 to 1.
Average odds for winning "dogs" were +151 - not enough to break even on flat bets.
And the two words that count the most in that sentence are flat bets.
Bet fixed or random amounts in any gambling scenario and you are, in the end, doomed.
Which is where target betting comes in - and how spread saves our bottom line from being submerged in red ink.
In the 7-dog trial that I set rolling last November 1, I chose a minimum bet of $100 and created separate worksheets to track spreads of 1 to 5 (maximum permitted bet $500) and 1 to 50 (max $5,000).
I learned a very long time ago that against table games (blackjack especially) the only way to win is to spread as wide as you can afford, with your overall win percentage inexorably tied to your max/min multiple.
Most players bounce around within a 1-5 spread, which is why most players lose most of the time.
Combining target betting with a 1-100 spread is a good start, and the more zeroes tacked on to the larger of those two numbers, the better your chances of beating the odds.
The standard, knee-jerk response to this easily proven truth is that "no one can afford bets that high" and if that's true, then no one but the house can afford to win.
Some perspective here: Casinos (and bookies) have such a high level of confidence in the numbers in their favor that spreads do not bother them in the least.
As long as players bet randomly, the house will always win in the end.
Punters who deviate consistently from expected betting patterns - Martingale or double-up bettors especially - will either be thwarted by table and house limits, or will be dealt with in other ways.
Casinos think of consistent winners (as opposed to occasionally lucky players, who are good for business) as cheats, and respond to them accordingly.
The house has the edge against 99.99% of all players, but that's not enough in their book!
Readers who have stayed with me for the past seven months or longer know that while both variations of the "dogs only" concept (5x and 50x) prospered for the first few weeks, those limited spreads began to take their toll.
I should at this point explain that some of the tables and graphs that follow differ from interim versions in earlier posts because I ran a comprehensive error check after May 31 and found some anomalies that had to be fixed.
The corrections reduced the overall wins instead of the opposite, which should at least spare me from skeptical claims that the 7-month error check was self-serving.
The 5x strategy hit its peak on January 18 at $6,000 (a ten-week winning run for a method that played out day by day in real time)...and then the wheels came off:
The wider 50x spread approach stayed in profit for an additional three weeks, hitting $37,000 in winnings on February 10, and then it too succumbed to the bookies' edge:
The second of the two worksheets had linked cells that enabled primary parameters (minimum and maximum bet values and the response to individual series turnarounds) to be varied.
The first two factors speak for themselves. The third involves the strategy applied when one or more of the seven series achieves its win target, but others are still struggling to recover.
Does it make best sense to make each series an entirely independent entity, falling back to a $100 minimum bet regardless of the overall "seven dog" bottom line, or should each series help the others when the combined loss to date (LTD) exceeds 7x the minimum ($700)?
The first thing the 50x worksheet tells us is that a 1 to 50 spread is not enough. Surprise!
Here's what woulda happened if we had applied a 1 to 250 spread from the start:
We woulda been making money all the way from November 1 through May 23 before last months tail-end tail-spin started siphoning off our hard-earned profits.
Here's the effect of "spreading the load" by making each series bet more than the minimum after recovery, as long as the combined LTD exceeds $1,400:
Again, plain sailing for month after month, and then we hit the rocks.
There are countless skeptical mythematicians out there who will howl that it's easy to come up with a change in betting rules that wins after the fact against games (or hands) that have already been played.
But they will be wasting their breath in this instance, because the rules applied to the 7-dog trial outcomes shown in these charts are almost identical to the target betting strategy that I have been publishing for free on the 'Net with very minor updates and modifications for more than a decade.
Specifically, those rules call for a very wide spread (much wider than 50x, for sure), and after a fixed number of double-ups in response to a losing streak, we freeze the bet until a mid-recovery win, and only then effect an increase that matches the win target.
Check the archives predating November 1, 2009, and you will see minor variations on the following theme:-
After an opening loss, increase the bet in pursuit of a quick recovery of prior losses (the LTD).
After a preset number of successive losses, freeze the next bet (NB) value until a win.
In response to a mid-recovery win, boost NB to cover the loss to date (LTD).
If the recovery bet is lost, repeat it and maintain its value until the next mid-recovery win, at which point the LTD+ process is replayed.
It is a very simple but consistently effective strategy that depends on discipline, confidence, and adequate resources (aka money).
Those same three factors are success essentials for bookies and casino operators.
A book would very quickly fold if it had a loss limit that did not adequately allow for what some pedants like to call "standard deviation" (a temporary suspension of the known house advantage) - and the same applies to all casinos.
It is safe to state categorically that more often than not, tight betting spreads and shoestring bankrolls are far more risky and reckless than the seemingly aggressive approach that I have been publicly advocating since the start of the 8th decade of the 20th century!
Game operators of all stripes count on their customers to bet randomly within a limited range and to carry a bankroll too puny to survive the briefest spasm of so-called bad luck.
Luck has nothing whatever to do with winning in the long-term.
Of course, very few people have unlimited resources - and that's why a wide spread requires a minimum far lower than the $100 I applied throughout the 7-dog trial.
Here's a replay of the 7-dog/7-months data with a $10 minimum, the max set at 300x and the post-recovery minimum pegged at LTD/70, meaning that if the combined LTD for all series exceeded $70 after an individual turnaround, the minimum would double, jumping to 3x at -$141 and so on.
Dispense with the LTD/70 "load sharing" divisor, and you get a chart that looks like this:
Aha! Says the skeptic. You're just increasing the maximum after the fact to distort the real result.
OK, so here's what we woulda seen if the max had been 5,000x the minimum, permitting bets from $10 all the way up to $50,000:
As you can see, massive 5000xMin bets were never in play, and the average bet value throughout the seven months stayed the same as for 300x.
Next, a different way of looking at the effect of dropping the max back down to the 50x ($100 to $5,000) that still applies to the ongoing 7-dog trial:
At 5x ($100 to $500), the picture becomes sadder still...
The screen snaps above contain a wealth of useful information for readers motivated to make the most of it.
What we see is proof that spread is to gambling what location is to the real estate business (which happens to be much more of a risk than target betting against table games or a sports book!).
Each summary above breaks down final outcomes series by series and shows total action for the entire seven months of the 7-dog trial, along with average bet values.
At 1-5, the average bet is $420. At 1-50, it's $980. At 1-300 (or 1-5,000!), the average bet value is between $680 and $760, keeping in mind that the wide-spread charts above show a $10 minimum.
What really matters is that at 1-300, we end the seven months in profit - and we do it at less risk.
QED, as Pythagoras would put it!
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. One more piece of friendly advice: If you are inclined to use target betting with real money against online "casinos" such as Bodog, spend a few minutes and save a lot of money by reading this._
No comments:
Post a Comment