New article: http://www.roulettician.com/articles/article2.html

Apologies to anyone who signed up for email subscriptions. It doesn't seem to be working, but I'm looking into it.

Thanks guys for the nice feedback.

Hi Guys,

It's been a while. Sorry if you've tried sending me a pm but I disabled receiving messages, so don't take it personally. 

The site is up and running-

http://www.roulettician.com

It's work in progress and I'll be adding articles regularly (aiming for one per week), plus software in due course. I'll post here with updates.

I'd like to say a big thanks to sqzboz who has helped a lot with proof-reading the site and making helpful suggestions - thanks B! 

Quote from: Drazen on May 05, 2015, 05:00:05 PM
Is this what you are using for your real play, as I see notification you said you will stress about its use?
Drazen

Sometimes I use this, but mostly I use a kind of matrix based on all possible permutations of 5 outcomes, which also includes staking options (which I'll be making available on the web site when it's finished). In this software the counts measure the number of misses. L2 & L4 track the longest runs since multiple hits, so it generates long losing streaks, and the default bet is for the direction to go up, but you can change this by clicking "Bet Opp." (if you want to keep track of the wins and losses). L3 tracks the longest sequence without a double win (two in a row). This can go to 60 or more and  can generate long choppy sequences. There are 9 options in all and I find it best to keep the pattern changing (leave "Fix Ptn" unchecked), although sometimes it's useful to fix it if you have a very unusual sequence and want to play it beyond the default point where it automatically switches to the next pattern. Or you can just use it as a point n figure plotter which tends to generate extreme sequences, betting either for the trend to continue down or whatever you like, based on subjective criteria.

Quote from: Drazen on May 02, 2015, 08:12:26 PM
This has been mentioned in the past and a few years ago Bayes actually made a DOS program to transfer EC bets to look like a stock charts for easier understanding.

Hi Drazen,

You'll probably find the attached easier to use. It's based roughly on an old program I uploaded to rouletteforum.cc which tracked a lot of bet selections and selected the worst performing ones according to the number of hits since a win and double win (the center radio buttons). You might have more success with this because of the graphical element, and also it tracks the last 60 spins which is 20 more than the old program.

[attachurl=1]

[attachimg=2]

I'm not!

HTH.

QuoteBayes, you and the other math guys are ever circling but don't know how to "go in for the kill".

So, are you saying the math is wrong?

I thought the point of Kav's question was to get us to think "out of the box", but to what end? did he have something specific in mind? What is the take-away value from this exchange?

You can dream up all kinds of novel approaches, but novelty shouldn't be an end in itself, IMO.

Actually, I've just realized that I didn't take the house limit into account in my simulation. If the reverse marty is limited to 7 steps (house max 100 units), then obviously you can't bet your entire bank whenever it exceeds this.

I think the question as posed by Kav wasn't very clear, but in general, the correct solution is to flat bet. This would also be confirmed by the mathematical expectation, which would tell you that in the long run you will make a 20% return on investment (assuming 60% win rate). It simply cannot be any less than that, without assuming unrealistic scenarios such as infinite house limits.

sqzbox,

I wrote a little simulation to test this. If you use a standard reverse marty assuming 7 steps (a house limit of 100 units) and also given a 60% win rate, the marty makes far more than flat betting. However, I disagree that losing your bank isn't a case of "minimal" profit, and in fact, if you bet your entire bank on each spin (whatever it may be) you will lose it every time. Even betting half your bank doesn't guarantee you will lose under these conditions; most of the time you will lose it but every now and then you make a huge profit which makes up for all the previous lost banks.

Theory says that  with a positive expectation you should be timid (bet a small % of your bank).

Kav,

I was thinking about this and it seems to me that a reverse marty would do the trick, as you had suggested. With the proviso that there are no house limits, it wouldn't take long, even with a 60 % win rate, for you to lose your entire bank, assuming you're betting a large fraction of it on each spin.

It seems you need no house limits both to lose with a reverse marty when you do have an edge and also to win with a standard one when you don't.

Without more detail it would be hard to say what kind of progression would result in minimum profits. If, for example, your longest losing run was 3, then a two-step martingale might do it. but I suspect that sqzbox is right and that it would be impossible to "achieve" anything less than if you were flat-betting, given that wins outnumber losses. So sorry, but you could always give the extra profits away to charity. 

horus,

One of the many things I intend to post on my blog is analyses of this kind of thing. Does the sum of the last X gaps have any bearing on future gaps?

The hot-number guys say bet when Sum(last X gaps) < N
The "maturity of chances" guys say bet when Sum(last X gaps) > N

Doing some research using various values of X and N might turn up something. Or maybe not... 

One thing common to all bets (streets, splits etc) is that the chance of at least one hit in a cycle is around 65%, so the W/L distribution is very similar to betting on two dozens, except that in the higher odds bets you have the chance of multiple wins in a cycle.

I posted the detailed stats somewhere on one of Vic's threads, can't remember the title of it though.

Quote from: TwoCatSam on April 11, 2015, 07:26:01 PM
I trust Bayes and if he said it, it's true.

Sam!  thanks for the vote of confidence. Yes, Skakus'  was the best confirmed result I've seen in all my years on the forums. It cannot easily be dismissed as luck either, because the challenge was a one-off. Maybe skakus will give you the system if you ask him nicely.

It's an interesting read, but you wonder what the point is. I mean, so what if you know what the cause or "force" is that makes roulette outcomes the way they are? even if you knew that the ultimate cause was quantum physics or whatever, or could precisely measure the initial conditions like ball velocity, the geometry of the wheel etc, it still wouldn't explain why the sequence of outcomes has no obvious pattern to it.

LGH was short on specifics, but he was a fan of hot numbers and gives the following guidelines for roulette:

For single numbers-

• If it hits twice with the last 10, follow once or twice
• If it hits 3 times within 20, follow up to 10 times. If it hits within that 10, continue to follow 10 times until the number cools off.

For splits-

• If it hits 3 times within 10 follow up to 5 times

For streets-

• If it hits 3 times within 8 follow up to 4 times

For Sixlines-

• If it hits 3 times within 5, follow twice

For dozens/columns-

• If they hit 3 in a row follow twice

For EC's-

• If 5 or 6 in a row, follow once

All bets that w are continued on any warmed-up cycle.

He also says that his favorite bet is all 3 EC's at the same time, but doesn't give any more detail.

It would be pretty easy to test all this, I haven't.