Thread: The truth about betting spreads: Why wide wins and tight tanks!

(Update at 6:00 PM PST: My apologies to the 70+ people who have read this since I posted it: I let a minor glitch slip by me, which would be a firing offense for the kids who work for me, and now the figures below are not 100% accurate. The principle is unaffected with a 1-10 spread losing 170u on action of 30,500 and a 1-5,000 spread winning 4,700u on action of 85,000 but the details are just a little off. I can't change the images below because there's a limit imposed by the BF board's exemplary software, but the data on the website link will be 100% on the ball very soon. Again, my apologies. I'm just glad it was me who found the mistake, not LL or II, although they never look at data and so would have remained blissfully unaware!)


There's an enduring myth that betting a very wide spread is dangerous foolishness and that the only way to win is to be cautious and rely on the right choice between Banker and Player to help us win more often, rather than letting bet values take care of the inherent house advantage.

I have said over and over that a wide betting spread will routinely result in smaller bets, on average, and therefore reduced risk over the long haul, so I will try to keep this short.

In order to focus more clearly on spreads, I created a new Excel model which will accept plugged-in data from any source (as long as it's in the right format - north to south, one B, P or T decision per line) and will give a detailed breakdown of the effects of variations in the low-high $ range and other critical factors. Most strategy evaluations provide an end result without corroboration or detailed methodology, and that's bad for all of us.

First time out, I took about 3,000 outcomes logged from sessions against an online baccarat sim, and pasted them into the new model. Sure, I know as well as you do that online casino table games are all crooked as a dog's hind leg, but it's quicker and easier to practice against a funny-munny game than it is to grind through real-time, real money shoes.

As it happens, the shoes I played were heavily Player biased, so I set the new model to bet Banker all the way, defying the odds in my usual reckless way. I then pulled screen shots to provide instant visuals that say I'm not that reckless after all. I have attached a few below as a sampling, and will provide more on my website as soon as I can.

First, betting Banker with a 1-10 spread (1:10 for you purists):

Target-Bovada 1 to 10 spread.JPG

We get action of 76,000u (average bet 25u) and an overall loss. At 1-50...

Target-Bovada 1 to 50 spread.JPG

...we still lose, and overall action jumps 5x, which is pretty much what one of the house shills posting to BF would tell you to expect and run away from, fast. But if we change the spread to 1-500...

Target-Bovada 1 to 500 spread.JPG

...we see what we're all looking for: a pretty decent win, and a dramatic reduction in the total action and therefore the risk.

Against this particular data set, my recommended max spread of 1-5,000 would not have had much effect on the overall win, but it would have further reduced the total action:

Target-Bovada 1 to 5,000 spread.JPG

The reason this works so consistently is that when you have a low ceiling on your bet values, a rough run will push you to your max, and then keep you there, quite possibly for hundreds of rounds (or until your BR runs dry). Low max, long haul is one way to think of it.

Casinos know that if you spread wide, you will bet more in the short term but will recover your losses far more rapidly when standard deviation pushes a temporary house spike back into compliance with expectation. That's the one and only reason for table limits and house limits. (House shills will tell you otherwise, but they're paid to lie to you).

Spreadsheets provide the most reliable and accurate platform for strategy tests primarily because they enable a line-by-line analysis of the effects of small changes in a betting method. My raucous opponents on this forum scoff at corroborative numbers, probably because they never provide data themselves. Perhaps they don't know how?

Something to keep an eye on in all these summaries is the difference between Target's average win values, and its average loss values. I have been preaching for half my lifetime that if you know you're going to lose more often than you win (and you will, over time) you need to bet in such a way that you win more when you win than you lose when you lose. My critics scoff at that, but don't ever put their money where their loud mouths are and disprove what elementary arithmetic shows to be a consistently achievable goal.

The link for the dedicated SpreadWideToWin page should be active pretty soon - keep trying if you want to know more about the mathematically proven benefits of running as wide a spread as your BR will allow.

Keep winning!

Seth T.

Blog
Website
Comm-comp XL file on SkyDrive