Losing streaks in terms of an EC

[Bayes]

Great chart you posted there, Bayes. Thanks.

Let's bet 3 numbers and stop on 1 win,

if you bet 3 numbers 12 times you profit/break even/lose.

If you bet 3 numbers 16 times you profit/break even/lose/lose.

Question:
If I opt for betting 16 times should I increase the stake for the last four bets making it profit/break even/profit/lose?

Or flat bet and maintain the original scenario of profit/break even/lose/lose?

You're welcome, Skakus.

I would choose to flat bet the 3 numbers, unless I had some strong indicator that the numbers would be likely to hit higher than expectation.

[Skakus]

...unless I had some strong indicator that the numbers would be likely to hit higher than expectation.

Ah yes, of course. That would make all the difference.

Damn! I've been testing so long with 3#'s for 12 spins now I'll have to start the whole friggin thing again this time with 16 spins!

[Skakus]

Ah yes, of course. That would make all the difference.

darn! I've been testing so long with 3#'s for 12 spins now I'll have to start the whole friggin thing again this time with 16 spins!

I've decided to push on with my test as is. When I'm finished I'll do it all again flat betting 3#'s for 16 spins (as per Bayes' chart for a 2 EC loss), then I'll do it all a third time adding in the extra progression chip on the last 4 attempts.

It's a substantial test so it will take a while but it will be interesting to see by how much the results have been impacted.

@ Bayes, could you do me a favour and check my maths with this,

I am 25% through my current test and have made 6243 bets on 3#'s, and I have had 543 hits. I calculate the z-score @ 2.66

Is this the correct z-score, and is it any good or is the sample still too small?

In your opinion what's an adequate sample size for 3 number bets?

Thanks.

[Bayes]

I make the z-score 1.73 with those values.

z = (w â€" np) / √[np(1â€" p)]

w = 543
n = 6,243
p = 3/37 = 0.081
1 â€" p = 0.919

z = (543 â€" 6243×0.081) / √(6243×0.081×0.919) = 1.731

Sample size of 10,000 - 12,000 should be ok. 

[Skakus]

 
Oops!

Must wear my glasses, it's supposed to be 563 hits, not 543. Sorry.

I should end up with about 25,000 bets so we will see.

[Bayes]

should have gone to spec savers

[Skakus]

 
That was a very disturbing video, thanks Bayes.