Bold strategy: useful in roulette?

[reddwarf]

Hi all,

After studying roulette and related mathematics quite intensive for some months I came to the following conclusions:

1.  the house edge is there to stay, there is no way we can get rid of it
2.  even without a house edge, casinos have the edge because they have a virtual unlimited bankroll, so a Gamblers ruin is about to occur -even in a positive expectation game you might end with a tremendous loss.   
3.  It has been mathematically proven that "bold play" is the optimal roulette strategy: this strategy basically minimizes the numbers of spins needed to reach a certain goal, or go bust.

So: I'm looking for a strategy where we make the most money in the shorthest time with a reasonable change of winning.  The idea that I currently have is to combine bets:

a.  when playing even changes, it takes the longest to reach a certain goal.  The upside is that there is a high probability I will survive long enough
b.  when playing single numbers I have the largest odds

The goal is a certain amount of money.  Once reached we stop.

The idea is that the EC' are used to ensure that we either gain a lot (single number wins), or break even (EC). 

Has anyone experience with this kind of approach?

When I've figured out how I can add snapshots I will show some graphs of tests that I ran.


greetings, rd

[No More Bets]

My whole feeling in terms of roulette is that you have to be very bold.

You sometimes have to be aggressive and gamble..

Just ensure the dealer is on YOUR side first.     

[mr.ore]

Your idea is absolutely correct, by chance I had the same idea few days ago. There were some articles regarding this topic which I posted there lately to this forum also in this brainstorming section. I did not see there were someone else with the same idea, just now I saw "bold play" in the title of your post. I just proposed method how to do what you are thinking about and posted it to that post.

nolinks://vlsroulette.com/brainstorming/interesting-reading-regarding-betting-policies-could-this-improve-our-gambling/

Well, I have not been studying roulette math for so long, so I don't understand all what is written in that article, but I believe, that something like basic strategy for roulette can be created. I am actually quite surprised this idea is not more wide spread.

I did look again into that article after reading on Czech motorbikers forum that some biker were trying with his friend system which is almost perfect, but you cannot use it in casino because you need computer to compute next bets and they would not allow it to have a notebook with you. He written that on on-line roulette it was winning for some quite long time, and then it lost, of course. After reading that, I realized there must be mathematically optimal strategy to play roulette, but no one plays it. Everyone is obsessed with making "holy grail", that nobody even tried to find "basic strategy", which might lose.

Please read my post in which I linked above, and also the articles. If you understand that math and know, how to compute that equation, I would like you to show me how to do it. Just now, I am thinking of making a program which will brute force through all possibilities to find optimal way of play.

A lot of people are trying to find "holy grail", yet they don't even have proper basic strategy! Let's combine forces and find at least that, because I'm almost sure it exists.

[reddwarf]

Hi mr.  ore,

I will study your posts.  In the meanwhile: do you know "hxxp: nolinks. bjmath. com/bjmath/thorp/tog. htm" written by E Thorpe? (The guy who cracked Blackjack!!!.  Especially chapter 4 is interesting.

greetings, rd

[mr.ore]

I already have got chapters 2,3,4 on my disk, but I did not studied them much. Some time ago when looking into chapter 4 when I saw Kelly criterion I decided not to read it since in roulette you can not have an advantage, and so that thing cannot be applied. And I do not understand a lot of what is there anyway, it's quite difficult to grasp.

I'm more into programming, in the end it will have to be done "conventionally", in that literature, there is some math condition which I don't understand, and which can determine, whether the strategy is optimal or not. This is to be coded. If you could tell me how to compute that condition, I could try to use genetic algorithm (best friend lazy programmers have today) to generate random bold policies, and letting it find most optimal ones(not guaranteed to find them, but probability increases with time the algorithm is running). It is not optimal way of findig them, but it is much easier to programme it and run it overnight. The result will be an optimal table telling us how to reach target bankroll with highest probability.

All what is now needed for me is how to decide, whether a policy (which is just a table on what odds and how many units bet according to our actual bankroll and target and table and allowed spread) is better than another one, and I need to be able to do compute it relatively quickly. Now I could do that by running simulation for say ten millions spins, but genetic algorithm needs to do that for each generated candidate policy, and it would take eternity.

On the other way, the policy table can be also drawn like an oriented graph, and the problem can be transferred to finding shortest path in it, for which there exists Dijkstra's algorithm, well, it needs to be little enhanced to take a care of cycles, but we do not want many cycles in strategy anyway, because they are "eating" our much needed probability. The problem is, that because roulette has not all odds we need for optimal bold strategy by it's definition, the optimal bold strategy for this game will contain some cycles. It can't be avoided, but we can minimize their impact.

Why to bother with this game anyway? There are a few reasons:

1) There is no basic strategy for this game, or, it have been found tens years ago by math people, but have been forgotten and not used. Now it seems to me there exist an infinite number of optimal strategies, for each target bankroll at least one. They must exist! For case when target bankroll is 3 units, it is sure that the optimal strategy exist, so there can at least be some for higher targets.

2) From its very nature, if we found the basic strategies, it would once and definitely prove that "holy grail" does NOT exists. Because one of conditions of optimal policy is, that there is NO better strategy, it cannot exist. Such a strategy is a supremum in a set of all strategies to reach that target. But such an optimal strategy is sure to lose because of house edge, therefore it is not "holy grail", but because there is really no better way of play, it is proved that "holy grail" cannot be found.

3) Many people play it, an many like it. So it would be good for them to know how to play. For other games there are known basic strategies, like for poker or blackjack.

4) Variance in roulette can be controlled by player, no other game can be so customized to the player favourite ways of play. There may be policies, which will not be absolutely optimal, but may be optimal for what player wants. There would be at least one optimal flat betting policy to reach the target, so flat bettors can improve probability to reach their target, while still flat betting. Or there may be nice policies for faster short term profits, which would be still better than many common progressions. You can say - I want to play only outside bets, then there would be at least one optimal policy for that. This would be perfect for a game with le'partage rule, the optimal way how to play it, inside betting for such a game is not optimal anyway!

5) People who believe in visual ballistics and dealer signature could also bet by such a basic strategy, and their only concern would be where to bet, they would have to know well only the wheel and dealer and believe in their own skills, on which they are betting. They would be sure, that they are playing the best way they can. After finding those policies, there would be nothing more left to mastering the game of roulette than to improve player's "predicting" skills, if it is even possible.

[reddwarf]

Hi mr.   ore,

I had a go at the article you posted.   My math is a somewhat rusty, but what I gather from  it is the following:

it is proven that bold play is optimal for a subfair game (roulette).   The closest thing that comes in the neighbourhood of bold play is the strategy proposed in the article: select to bet which leads closest to the target with minimal betting amount. 

If we accept this proof (why not?) than the recipy is fairly simple:

we can only encounter as many betting decisions as the target amount minus 1.   For example: goal is 10 units, so the above strategy will lead to 9 decisions (what to bet if i have 1 unit left, 2 units left etc).  

The thing is however: due to the limited betting options the optimal play in roulette has lower succesrate than bold play.  But still there is an optimum.

How to determine the optimimum? I have to study a little while longer to determine what the optimalization rules are.

This evening I will test it this strategy with the extra assumption that only for low odds (2:1 and 1:1) i can bet more than 1 unit (this gives for N=10, the same policy as in the article). 

greetings rd

By the way: the reason why I like chapter 4 of the article I mentioned: in it is a loophole.  There is a set of bets that do not fullfile all requirements for a failed system.  I'm not sure if this can be exploited or not, but somehow I have the idea that it is related to a specific form of bold play.  here is the thing: if the summed subsequent betting amounts are less than the betting amount of a bet that was won, than 2 out of 4 reasons why a roulette bet is worthless do not apply anymore. . .

[reddwarf]

-