System question

suicidekings · 12-07-11, 05:08 PM

Originally posted by celzus

Hi all, I am completely new to betting. I've been looking around at the different systems and I found one which seems to me to be viable, except people say it's not

Could somone please explain why. Basicaly it's a version of martingale. Every time you lose you bet double of what you lost and so for losing 1 unit you bet 2 and get four. This covers the first one lost and gives me a profit of 1. Now the problem is of course if you lose too much in a row. But let's say I have a bank of 31 units. I can lose 1-2-4-8-16 which means five times until I am bankrupt. However, if I can go through 31 series without getting 5 losses in a row I would still be at an even zero. (31 wins - 31 units lost in bank). Now if I added a unit to each lost bet so that, for example if I lose 2, I bet 7 so that I get profits on each bet lost, so that when I cover my loses and I also add a profit for each bet. This would mean, i think 1-2-3-4-5=15 units to get profits for each bet lost. (this is done so I don't get profits each series but each bet I make) Assuming I have a reserve bank of another 46 units, I can't see how, probability-wise... I don't make profit. (I just break off when I get to 5 losses in a row and the chances of not making that back in the second bank seem impossible) I am sure I've got this wrong somewhere... (does someone know the math for it?) Basically how often can you expect 5 losses in a row in 46 bets assuming they are 50-50 bets? Thanks!

In a truly 50-50 proposition, the odds of you getting X many bets in a row correct is the same as getting X many bets incorrect in a row, yes? So the odds of a particular series losing is 0.5^5 = 3.125%

A conventional chase to cover losses and still win 1u would be (including the juice on -110 odds): (#1) 1.1u to win 1u, (#2) 2.31u to win 2.1, (#3) 4.85u to win 4.1u, (#4) 10.19u to win 9.26u, (#5) 21.39u to win 19.45. Losing 5 in a row = -39.84u, not 31u, and the expected value of a series = EV = 96.875%*1u + 3.125%*-39.84u = -0.276u. This means you lose money over time, even under perfect conditions.

No matter how you modify the system, you can't escape the juice. If you reduce the risk at each level to win less, just trying to get back to even, you then stand to win less on each wager, which makes the whole endeavour even less worth your time as the final goal is ultimately maximizing profit, not spending your time and energy trying to dig yourself out of the hole.

The biggest problem with your idea is that it completely neglects variance. Over the long term, 50/50 propositions are going to come out relatively even, which falsely implies that over smaller data sets the results will be around 50%. The reality is sports vary a lot and you can very easily go 2-13 over a few days and then go on a 13-2 run right after that. This happens to EVERYONE at points, and in this case, you would have 2 series losses on a run like this. You're actually proposing a much riskier methodology than just making straight wagers at a fixed unit size (or Kelly %-based units).

suicidekings · 12-07-11, 05:36 PM

If you have 46u to work with, here's an alternate suggestion:

Breakeven on -110 odds is a winning percentage of 52.4%. Let's target a winning percentage of 53.5%. Kelly criteria says you should be risking 2.35% of your bankroll with this edge, which means risking 1.1u per wager to start (recalculated after every bet). If you can do this, on average, after 50 bets your bankroll should be at $4900, after 100 bets it should be $5228, etc. If you were to lose 5 bets in a row, you'd be out ~5.5u instead of 39.4u.

Doesn't that seem a lot easier and less stressful than a chase system?