+ Register

[Bayes]

3652 units is the minimum for a progression that puts you in plus when you hit, it covers 185 spins

Thanks soggett. Let's put this in context; 185 spins is equivalent to around 6 misses on an even chance - just 6!!!

So IMHO it's HIGHLY unlikely that this system could have achieved a 20k profit even as a one-off. I would leave well alone. 

[soggett]

Thanks soggett. Let's put this in context; 185 spins is equivalent to around 6 misses on an even chance - just 6!!!

So IMHO it's HIGHLY unlikely that this system could have achieved a 20k profit even as a one-off. I would leave well alone. 

no problem

are you sure it is just 6? 

wow
[revealb]

I got to find that chart of yours
[/revealb]

[Bayes]

are you sure it is just 6? 

It seemed too low to me too, then I realized I had made a mistake because I used the standard deviation of 185 misses on a number instead of the actual probability. It should be this:

Chance of 185 misses on a number is (36/37)¹⁸⁵

Chance of N misses on an EC is (19/37)^N

So equating the two terms and solving for N gives:

(36/37)¹⁸⁵ = (19/37)^N

185 × log (36/37) = N × log(19/37)

N = 185 × log(36/37) / log(19/37) = 7.6

This has to be rounded down to 7 because you can have no more than 185 misses, and if you round it up to 8 it goes over 185.

[Chrisbis]

Jules, am I right in surmising that the:-

Chance of N misses on an EC is (19/37)N

is the internal count of all the possible felt/table bets that would equate to an EC (Even Chance) bet? (equaling 19)

(including betting 18 inside numbers, betting 3Lines, betting 6 streets, betting 2 dozens & 2Lines with ratio values, and all other combinations)

[PLIP50]

Bayes
Could you explain that in laymans terms please!

Like how many times out of say 100 picked numbers
Would a number not come out in 185 spins!
Kind regards Nick

[soggett]

It seemed too low to me too, then I realized I had made a mistake because I used the standard deviation of 185 misses on a number instead of the actual probability. It should be this:

Chance of 185 misses on a number is (36/37)¹⁸⁵

Chance of N misses on an EC is (19/37)^N

So equating the two terms and solving for N gives:

(36/37)¹⁸⁵ = (19/37)^N

185 × log (36/37) = N × log(19/37)

N = 185 × log(36/37) / log(19/37) = 7.6

This has to be rounded down to 7 because you can have no more than 185 misses, and if you round it up to 8 it goes over 185.

yes, i thought it was somewhere around 8
still it is to small, too bad

[Bayes]

am I right in surmising that the:- is the internal count of all the possible felt/table bets that would equate to an EC (Even Chance) bet? (equaling 19)

(including betting 18 inside numbers, betting 3Lines, betting 6 streets, betting 2 dozens & 2Lines with ratio values, and all other combinations)

Chris, no it's simpler than that. For a loss on any EC, the number of ways you can lose is 18 + the zero, which gives 19. Probability is the number of ways an "event" can happen divided by the total number of equally likely events, which is 37. Hence the probability of a loss is 19/37. It's the same principle for the single number, so the number of ways your number CANNOT hit is 36, and the total number of equally likely events is again 37, so P = 36/37.

[Bayes]

Bayes
Could you explain that in laymans terms please!

Like how many times out of say 100 picked numbers
Would a number not come out in 185 spins!
Kind regards Nick

Hi PLIP50,

Your example would actually be more complicated to work out. It might be easier to think of it in terms of the number of times, on average, that you will bust. So in the case of a single number it will go MORE than 185 spins before it hits 1 time in 157 attacks. I was just comparing this to an EC because more people are familiar with the losing runs on a EC than they are for other bets. Hope this helps.

[XXVV]

@Bayes


Thanks for the crystal clear illustrations.
All very well for a situation where no 'skill' or interface with influence on possible outcome is concerned, but what about say, based on large bet samples, the player developed a theory that demonstrated say a +5% edge, thus overcoming the house edge at say -2.7%, and enabling a small net positive tendency.


There are numerous linear methods that can achieve that, and I would nominate some, given appropriate filters and 'triggers' to stop /start ( I know you are not hot on 'triggers), then a progression of this nature may be useful.


Also, but I won't go into it here, what if there could be a mental influence on outcome to shift the results say 10% your way.


In fact perhaps you could illustrate what 'edge' would be necessary achieve 'consistent' victory and thus avoid the risk threshold of 185 spins.


Or even with say a 10% edge might there still be 'whales' out there that would cause wipeout?


Or worse, with progressions is there always the lurking possibility it could be a wipeout day?


I suspect the latter.


Best wishes
XXVV

[esoito]

I have to say that when I wrote to him a while back, the question the author was most evasive and unhelpful about was concerning the number of losses versus wins, based on his experience of using the software.

He gave me no data about that.

(Perhaps I needed to rephrase the question.)

[Bayes]

Or worse, with progressions is there always the lurking possibility it could be a wipeout day?

XXVV,

With respect to the martingale in particular, it's not just a possibility, it's a certainty. Under no circumstances whatsoever are you ever justified in risking so much over such a small number of spins (and it's always a "small number of spins" relative to the odds), and that's just what the marty does. It's the crack cocaine of gambling systems, and the lure of that quick fix is just too tempting for many. A slick web site and fancy software doesn't make up for the fact that it's just a very very bad idea.

Even with a healthy edge (say 5%) over the casino, a martingale can still get you into serious trouble, and in any case, if you do have an edge, there are much better options such as Kelly betting or staking a small % of your current bank.

[Gizmotron]

Wow Bayes, that's as beautiful as Milla Jovovich in Ultraviolet.

[Edit: For those who want to know about MJ see the OFF TOPIC section -- where it belongs.]

[Bayes]

Gizmo, thanks for the compliment.

In fact perhaps you could illustrate what 'edge' would be necessary achieve 'consistent' victory and thus avoid the risk threshold of 185 spins.

XXVV,

Interesting question. There are a number of ways you could work this out but let's suppose that instead of the probability of a bust being 1 in 157 (which is the true probability), it's a much smaller probability of say, 1 in a million. That would give us the assurance that it would be an extremely rare event.

So from this assumption, we have the equation:

(1 − x/37)¹⁸⁵ = 10^-6

In which 'x' represents that portion of the wheel which you'd need to cover to get the probability of 1 in a million that you would not get a hit in 185 spins. I won't go through the steps of solving this, but the answer is 2.66, so let's round it up to 3. This means that if you were to bet a street (3 numbers covered), the chance would be at least 1 in a million (actually higher than 1 in a million, because we've rounded UP) that you would not get a hit in 185 spins. Now let's suppose that you were paid the usual single number 35-1 instead of  street odds. If you work out the expectation of this, it will give you the edge required in order to overcome the 185 spin threshold (there is still a very small chance of a bust, but it's negligible). So for the purpose of obtaining the required edge, we assume we're getting paid for betting on a single number, but the probability of that number hitting is actually that of 3 numbers (3/37).

Expectation  = probability × payout

So, E(X) = 3/37 × 35 − 34/37 = 1.92
So your edge would need to be a massive 192% in order to beat the 185 spin threshold. 

It's a sobering thought.

[Bayes]

I have to say that when I wrote to him a while back, the question the author was most evasive and unhelpful about was concerning the number of losses versus wins, based on his experience of using the software.

He gave me no data about that.

(Perhaps I needed to rephrase the question.)

Perhaps you could rephrase it thus: "When are you going to stop misleading people with your specious nonsense?" 

At the foot of the page he proudly proclaims -

Voted Best Roulette Software Tool 2012!

eh? and where exactly did this comparison of roulette software tools take place? and did anyone vote other than you?

He compares other betting progressions and criticizes the martingale, then follows with this:

The Martingale roulette system and the Roulette Number system share just one similarity, and that is that a single win recovers all previous losses and leaves you with a profit on top. Apart from that, the Roulette Number system is much better. With the Martingale system you can lose your whole bankroll within 10 to 15 losing bets, while with the Roulette Number system you can have up to 230 losing bets (230 tries) before your bankroll is exhausted.
Even if you start with just $0.01 using the Martingale roulette strategy, you'll have accumulated a total loss of $40 with only 12 consecutive spins. Also, each time you win you'll rake in only $0.01 profit. On the other hand, with Roulette Number you can have 185 tries (spins) before you use up the same $40, yet each time you will win from $0.01 up to $0.35. That's a very substantial difference compared to the Martingale system.

Yeah, right. 

What he fails to point out is that a marty on an EC will stretch to 12 bets for your $40, but as we've seen, $40 betting a single number will only get you the equivalent of 7, or at most 8 bets. It's comparing apples with oranges. 

[XXVV]

@Bayes


This is fantastic work thankyou. Sobering indeed, to the point of a cold shower on a mountain summit
at 5am (being say one hour before sunrise- traditionally the coldest time). If we add a 100kmph wind and the chill factor, and that we have no protective clothing, its bleak.


Okay will aim for a +200% edge in ongoing work!


Will have to review all this data.


Fortunately real life means we can work in short bursts, select the table at which to play where we can review past spins ( some argue this has no meaning but I totally disagree as I have demonstrated so often all spins are interconnected within the context of a random distribution cycle), choose when to enter and exit, and apply the most effective methodologies and strategies at will. The challenge is on.


Also, and perhaps the best way forward, avoid progressions. Minimise loss.


Thank you for your comments. Much food for thought.
XXVV