Popular pages:

Roulette System

The Roulette Systems That Really Work

Roulette Computers

Hidden Electronics That Predict Spins

Roulette Strategy

Why Roulette Betting Strategies Lose

Roulette System

The Honest Live Online Roulette Casinos

Reducing Volatility @ TwoCatSam

Started by rob567, May 08, 2008, 01:58:49 PM

0 Members and 2 Guests are viewing this topic.

Noble Savage

All your conclusions are valid.

The Kelly criterion does work though, only if you have an edge. It's supposed to optimize capital growth on an already winning game, so there was no point testing it on a losing game (I wonder how you made the calculation with no available edge).

Quote from: reddwarf on March 31, 2010, 05:40:52 AM
After that, I'm done with gambling: I lost (luckily) only a few hundred dollars.  And many, many hours simulating million of spins

I hear you.

That's why I hate it when people give others false hope with no real knowledge of what they're talking about. It's us guys who spent endless hours with the game, seen millions of spins and tested every single approach who know. That's what backs up our conclusions, verifiable proof (along with every scientific source on earth), not some "bumblebees can fly so I can beat the math", "man beat gravity so the math is beatable", or some self-contradicting gibberish about reading randomness.

reddwarf

Yep i agree.

Although I do like it when people challenge "common beliefs" and knowledge.  By the way: I tested kelly criterium in combination with the common belief that after a couple of times "red" "black" must occur (in other words an edge must appear).  I simulated many spins, in which I varied the fraction of bank roll used for bet as a function of the streak length.  So actually I tested several things at once.  Anyway: the result was negative.


Danger Man

Quote from: reddwarf

Although I do like it when people challenge "common beliefs" and knowledge. 


Unfortunately where money is involved people are willing to believe anything.

reddwarf

Yep, you are right. When money is involved people tend to believe a lot: not believing would, in the case of roulette, mean a lot of studying, and when doubting testing.

I think it was Einstein who also sad something like: "stupidity is trying the same thing again and again, hoping for a different outcome". In that sense I was stupid. Now it's just a nice hobby to test new methods/systems and believe it or not, I learned a lot about programming, statistics (I am a physicist), randomness, finance.

Noble Savage

Quote from: reddwarf on March 31, 2010, 01:18:55 PM
In that sense I was stupid. Now it's just a nice hobby to test new methods/systems and believe it or not, I learned a lot about programming, statistics (I am a physicist), randomness, finance.

Same here. :)

Proofreaders2000

"That's what backs up our conclusions, verifiable proof (along with every scientific source on earth), not some "bumblebees can fly so I can beat the math", "man beat gravity so the math is beatable", or some self-contradicting gibberish about reading randomness."---Noble Savage

With games of chance you have unfair payouts--the casinos would not profit unless the game was heavily in their favor.  And yes, it would be better not to gamble looking at the maths, a cogent arguement.

With that said, when I test a system, my goal is to test by hand (at least 5 hours a day) for 30 days to see how well it holds up.  If there is a decent return the following month, then I try with a small bankroll to see what happens.

As for the bumblebees remark, science ultimately is a finite resource.  It is a testament to what mankind doesn't know.

The Holy Grail does exist.  If someone doesn't seek it, we'll never find it.   I will never stop looking.



Jish


Monte Carlo

I am Monte Carlo.

I can't believe you are still trying to figure this out 2 years later.  Incidentally the concept was merely a preamble to explaining a much more complex idea.  Like explaining buoyancy before ship building.  Nevertheless;

1.  RedDwarf you are wrong.  You have calculated the wrong math equation.  It is not about the probability of losing 2 times in a row.  It's about the probability of the number of possible repeats within a cycle, this is the volatility.  The casino developed the spread on betting of an e/c bet based on this number but using simple rather than combinatoric probability.  But in doing what I am talking about you have to pay a price, a 50% increase of the house edge.  In return I know the range of volatility has been reduced by 50%.  The question then becomes how can you exploit this?
2. This theory does apply to a roulette solution but the theory is not an answer in itself.  And no system wins 100% of the time.  I just win a hell of a lot more than I lose.
3.  I never raised the issue of Kelly criterion of wage management.  At least I don't remember doing so.  In any case it's not relevant to understanding the basic concept presented since it is not a solution in itself.

Good luck I will check up in a couple of years and see how you are doing.

Monte Carlo



reddwarf

Hi Monte Carlo,

Thanks for your reply. Yes, if your definition of volatility is the number of repeats in a cycle, you are absolutely right. The thing is however: how to utilize this, so again you are right again: it is not a system in itself, but it could help in building a system.

OK, I'll give this one another try than (I actually stopped testing systems etc as I thought it is fruitless, but your claims are interesting though)

c'ya in two years time



airkyd

i dont you can reduce volatility.  playing roulette you can expect and you probably have seen  R R R R R R,  B B B B B B or even R B R B R B R B . . . .  is this a shock to you? should hope not ! however over 1000 spins regardless of the mentioned patterns you will still expect something like a 50/50 red/black percentage.  considering that all spins are RANDOM and the  PROBABILITY of any number/colour is even to each other, reducing "volatility" is relative.   

what i am saying is that  BBBBRRRRRRRRRBBBBB is volatile over 50 or so spins . . .  but over 1000 spins this is insignificant. 

i guess you can reduce volatility by being consistent over many many spins!

Monte Carlo

Remember the data set remains the same with every spin counted but because of the push which returns your bet you have reduce the data set in relation to  wager only.   (Ignoring zero for argument) The odds of black or red occurring remains the same at ½.  The odds of r/o or b/e with r/e and b/o being the push is ¼.  This means that even though you have a 1/1. 04E6 chance of 20 reds in a row.  The odds of 20 r/e are 1/1. 09E12.  To reduce the odds to the same one in a million chance you will have to have 10 r/e in a row but this only applies to your wager.  Which is half the volitility of the data set as a whole.   Since we have created two separate volatility curves pushing a greater number of stochastical outcomes into lower tier brackets only in relationship to wagers, a changeover in trend line should occur earlier and be exploitable through prudent wage management.

reddwarf

It is clear to me that if volatility is defined as more of the same, that volatility can be reduced. In the extreme case of volatility =0, we basically have created a conditional game which can always be won in when the example of Monte Carlo is used (combined bets). The question therefore is: does a reduced volatility change the "nature" of randomness, and if so, how can it be exploited, because there is a catch:

Lets assume we play a game where we simultaneously bet on two EC's (color/parity for example). It is true that the volatility has been reduced but:
1. it came at the cost of increased outcome space (we now have 4 possible outcomes: r/e, r/o, b/e, b/o)
2. and, depending on the betting strategy, it also comes at the cost of increase house edge as the percentage of "zero's" versus the outcomes that either generate win or loss has increased.

So my question to other readers is: how to avoid that the "game" collapses again in a regular loosing game: for example if we would ignore the break-even events in out betting strategy, we basically created a new game with two possible outcomes (win/loose) with an increased house edge, and even worse, with an unreduced volatility: the game with reduced volatility "collapsed" onto an game that is even worse than the original. (this means that we can not ignore the break evens and treat them as "nothing-happened" situations

What I understand from Monte Carlo's post is that he claims that a slight conditionality (in other words, past spins do influence future spins) is introduced. I'm willing to accept that this might be possible, even from a mathematical point of view. But at this moment it is still unclear to me how he ensures that the game does not collapse.

A question to Monte Carlo: can you please explain "we have created two separate volatility curves" to me? Do you mean with the two volatility curves the old one and the new one? Or two new ones?

Red dwarf


Carpanta

While playing inside bets volatility can be reduced playing numbers favoured by a hit at least. Discard those numbers that havent had a hit coz they can be cold numbers for a long time. Produce the right strategy to play those numbers that have shown recently. Rightstrategy means tracking tools that will tell you which of those "hot numbers" are most likely to repeat next spin.

reddwarf

Hmm,

this is not what I'm looking for, so I will ignore it.

It is a mathematical fact that we can create "less" random sequence by combining two perfect random sequences. This "mixing" must meet specific requirements. (By the way, people who believe it is not possible, just google on "allison mixtures").

The "less" random sequence has nonzero auto correlation. How must correlation be interpreted?
1. if there is positive auto correlation, above average performance tends to repeat (when the sequence is purely random and people wrongly believe that data is positively auto correlated, they talk about "hot games", "hot numbers" etc)
2. if the set is negatively auto correlated, below average performance tends to repeat (when the sequence is purely random and people wrongly believe that data is negatively auto correlated, they believe that a color is due because it did not come up for some time: gamblers fallacy!)

So now back to Monte Carlo's remarks: a reduced volatility will lead in the extreme to a negative auto correlation (volatility =0), but again, this can never be achieved. But still it there might be a way to achieve a "mild" form by combining bets. So the goal is to transform the purely random sequence into  less random sequence, which is indeed possible for sequential bets, but still the question remains if this can be utilized.

Conclusions thusfar: nature of randomness can be changed by very specific combination of bets (at least sequential bets). If this can be utilized remains to be seen

Still investigating and looking for an answer (but not the commercial kind!)

reddwarf

-