This is something that has always made me wonder.
The Van Keelen Test says that you need 55% wins flat betting over 1k placed bets.
What if you were able to hit 52-53% wins flat betting over 1k placed bets consistently? And by consistently I mean if you are able to accomplish this with at least a 60% session strike rate, 6/10 times. Would you consider that an advantage?
Although I think that it helps being at or more then 50% strike ratio, the key for me is in variance of the W/L distribution.
Now comes my next question...
How many placed bets does it take to consider the variance of your bet selection? If you have 5k placed bets and you have never seen more then 10 losses in row and it has only happened once do you consider that a good thing?
I would like some opinions on this. And I also to think it would be interesting to see what others set their bench marks at when it comes to testing requirements.
'when do you have an advantage?'
When the casino tell you to go and play somewhere else! [smiley]aes/tongue.png[/smiley]
It's a seriously hard question to answer because of all the variables involved. It would be very noble to be able to win outright flat betting. However, I think this just puts a straightjacket on a player. Trying to work out any kind of advantage using progressions would just muddy the waters even further.
I don't think 1k or even 5k worth of results is going to give us an answer. There will be a small percent of players who come out ahead in their lifetime maybe only playing 5k worth of spins. It does not mean they had any kind of advantage other than good luck.
I don't think these tests like the Van Keelen test or Dr Wilson's acid test conclusively proof anything one way or another.
This may sound silly and not very scientific, However, I get a gut feeling fairly quickly if something is worth investigating further or not. I could be playing something and winning with it, however my gut instinct will tell me quickly if it's a disaster waiting to happen. I have paid the price a few times for not listening to my instinct. So who is to say instinct is not as valuable to us in gambling compared to percentages and what not.
Sorry Big EZ that I couldn't really specifically answer any of your questions. But I think that's in part because they are kind of unanswerable if you are playing strategies as opposed to taking a physical approach.
Let's hope someone else comes along and can help out.
Hi Big EZ,
That's a good question. A big part of the science of statistics is about determining when you have an "advantage". e.g., in testing a new drug, how do you know if it gives better results than a placebo? or in agriculture, does fertilizer A produce better quality crops than fertilizer B?
The logic behind classical hypothesis testing is quite simple. You assume that there is no difference (or advantage) in using the "treatment" (or system), this is called the "null" hypothesis. Then you calculate how likely it is that you got your results by chance alone, using probability. If this probability is small (say 1% or less), you reject the null hypothesis. In a sense, Bally is correct: you can never PROVE conclusively that your system/drug/fertilizer etc is a "winner", you can only show that on the present evidence and to a certain degree of probability, it is/isn't.
Whatever you're testing, in each case you are trying to draw a conclusion based on a sample from a "population". In most research scenarios, gathering data may be very expensive or difficult, and this is where your next question about how many bets are necessary comes in. A concept called the "power" of a test can tell you how much data you need in order to come to some conclusion within a specified level of significance (the level of significance determines how much coincidence you're prepared to put up with). So the more powerful the test, the less data you need.
There isn't a "one size fits all" test for all kinds of data; in fact there are a large number of tests for different distributions. These days, they're relatively easy to do with software, so you don't need to know much math or do hard calculations, but you do need to be able to recognize which test is appropriate and interpret the results correctly. Any decent spreadsheet will have a lot of statistical functions built in (including tests). There are lots of free resources online for learning stats, including a good one here (http://stattrek.com/tutorials/statistics-tutorial.aspx).
You might think that with roulette, the data is easy to come by and cheap, so it's easy to get a very large sample (and no sophisticated statistical test is necessary). That's true up to a point, but manual testing takes a lot of time, and even if you can write simulations your system may be quite complex and difficult to code. With a reasonably powerful statistical test, you may only need 100 placed bets or so before you can draw a conclusion (reject it, or do more testing).
Statistics is actually a fascinating subject and not nearly as dry as you might think, but it helps if you're a bit of a geek. :D
Well said Bayes. This is exactly why I like reading your posts. To the point, honest and simple. Only a very knowledgeable person can put such complex mathematical principles in a simplistic manner. I can understand how complex it is especially as we use statistical models in our robotics industry. It is interesting to see how statistics is applicable from such a complicated futuristic industry to old ancient games like roulette :)
One clue from there is, it is very easy to understand such intricate things when it is taught rather than reading it. Taking a clue from what Bayes has done, if the experts here are willing to share how they test their systems, that will really help newbies like me :) Thanks!
Now comes my next question...
How many placed bets does it take to consider the variance of your bet selection? If you have 5k placed bets and you have never seen more then 10 losses in row and it has only happened once do you consider that a good thing?
-Have developed a method/tested over 60 Wiesbaden sessions/about 20 k spins.......never pass 7 losses
in the row....won 59/60........and still am not sure if it is a good thing.
Rule of thumb usually is:-
"If it looks good......it (http://forum.quoteland.com/eve/forums/a/tpc/f/99191541/m/3291979166) is good"
as against the old adage:-
"If it sounds too (http://forum.wordreference.com/showthread.php?t=874951) good to be true, it probably is!"
The proof is in the pudding, Ivica.....and if your ahead with that particular bet, then its answered ALL your questions.
[smiley]aes/martini.png[/smiley]
This is probably a huge load of bull.........
Jl said that random has virtual limits. Either there is a limit or not, so let's change his verbage: Random has limits. I have come to believe this. I also believe where you stand and which way you look is important.
Example: If you are looking at 10 blacks in a row, you know the next black has as good a chance of hitting as red. You are looking forward. Now look backward. You make make this truthful statement: Black has certainly beaten the odds for the last ten spins! There comes a point where what black has done is remarkable.
Now imagine yourself being able to move into the future one spin. You could make this statement: If black hits again, it will have beaten the odds even more. And so on....
Now the question becomes: How long can black beat the odds? At some point (who knows where?) black will fall to red. Random has, in fact, reached its limits for that trot.
So the question becomes: Can we find a way to begin our betting at the time when random has just about ran the race and is ready to sit for a spell? I think it is possible. And not by waiting for ten blacks in a row.
I find, using the ExcelBot, that while random will beat me maybe seven of ten times at bat, the three times I win will move me ahead an infinitesimal amount. I have run perhaps 100,000 spins over the last few weeks and this has held up on play money. I have kept precise records of my wins and losses, so I will have a near-perfect idea of what to expect. I will advance the bankroll to ensure I do not crash from lack of funds.
So the question will be this: After 100,000 successful spins--not bets--can I see the same results with real money. I will tell you one something: If I don't I will begin to seriously suspect BV of manipulating their number stream.
What a load of bull.........
TwoCat
Sam,
Does this not become an earnest discussion about standard deviation and using it as a tool for bet selection?
Some of the best winning runs are flat betting after a sleeper awakens (not chasing the sleeper awaiting the win). Let's say a dozen or column sleeps for 15 spins (yes I know they can sleep for more than twice that...but I'm being practical and not wanting to sit forever waiting for a trigger) or maybe a better example is dozen has hit 3 times in 40 spins...your deviation from the expected norm is decently high (or your limit of random) and even flat betting the next 40 spins in this case I would wager that it comes back towards its norm enough to make money (an advantage) and I know it's not a huge sampling...but I would say I would have an advantage playing into that scenario over a number of sessions (yes I would play a slight progression...but as you say...random would essentially have it's limit in most cases...as the numbers would likely tend to fall back towards a norm)
We control when we bet, how much we bet and what we bet on...seems like that should be enough advantage in and of itself if you are a patient student of the game...and well funded. Just my opinion.
All very good responses guys....
Bayes,
I am going to look into the link you posted to see if I can figure out what I need to do.
Bally,
My gut tells me this one is good. Flat betting shows profits!! Im going to see if I can get over 50% strike rate with 10k placed bets. I manual test so that may take a little while longer.
FLAT,
That sounds impressive. How many placed bets over those 20k spins? And is it an even chance bet?
Sam,
I am on the same page as you. Random has limits and I think if you set the conditions right you can take advantage of it.
RouletteKey,
Patience is a big part of the game. But don't about self control either
They all like this/first one/---and you can see what it is all about.
May say why not 4 last lines in continuation......no,this is much better.
Quote from: Big EZ on July 26, 2013, 11:55:23 PM
Bayes,
I am going to look into the link you posted to see if I can figure out what I need to do.
After you've got a basic grasp of the theory you might like to try this free statistical software which gives you help in deciding what tests to use -
http://www.sofastatistics.com/home.php (http://www.sofastatistics.com/home.php)
Here's a nice little tutorial on the concept of hypothesis testing - no math involved!
http://evc-cit.info/psych018/hyptest/ (http://evc-cit.info/psych018/hyptest/)
You can use any bet selection method that you want to. All it needs to do is to
provide you with the state of current conditions. The most effective method is to
use your bet selections to see what the current conditions are.
There are three basic conditions. These three conditions are extremely good,
moderate, and extremely bad. I have found that targeting the extremely good
works very well for me. I also avoid placing large bets when either moderate
or extremely bad conditions are occurring.
From that, conditional management, CM, you can deliberately change the
effect that variance has on continuous play. I would call the method an
advantage. If you want to learn this technique then set out to identify these
three common conditions. Once you can see them you will never be able
to ignore them. The casinos can't prevent them from occurring. You can
take advantage of them.
Gizmo,
Would you consider a bad ncondition playing after 3-4 Lynchburg Lemomades?
Please advise.
Thanks
ND
Atlas Shrugged ! ! !
Bayes,
Thanks for that extra link. Still trying to figure it all out [smiley]aes/thinking.png[/smiley]
Gizmo,
The variance in my W/L sequence is very tight, most of the time. Its hard to determine conditional management as you put it.
For example I would consider an extremely good condition as winning on the first or second bet because that is win or break even situation, moderate would be winning between the 3nd and 6th bet, and extremely bad are the rare time over 6 losses are reached.
Even though there are extremely bad conditions, I believe they can only be extreme up to a certain point.
ND,
Maybe if its made with something better then Jack
Quote from: Big EZ on July 25, 2013, 10:58:12 PM
This is something that has always made me wonder.
The Van Keelen Test says that you need 55% wins flat betting over 1k placed bets.
What if you were able to hit 52-53% wins flat betting over 1k placed bets consistently? And by consistently I mean if you are able to accomplish this with at least a 60% session strike rate, 6/10 times. Would you consider that an advantage?
Although I think that it helps being at or more then 50% strike ratio, the key for me is in variance of the W/L distribution.
Now comes my next question...
How many placed bets does it take to consider the variance of your bet selection? If you have 5k placed bets and you have never seen more then 10 losses in row and it has only happened once do you consider that a good thing?
I would like some opinions on this. And I also to think it would be interesting to see what others set their bench marks at when it comes to testing requirements.
At 1000 trials, playing an even money bet (18/37 european wheel)
600/1000 hits means 7.2 standard deviations playing 18 numbers.
I would say say your bet selection has hit at a higher rate than a random selection goes.
You have already an advantage for the next 1000.
At previous 1000 trials you might have some positive or negative fluctuation, so, next 1000 test could be from 500/1000 to 700/1000
nice you decide you have got the advantage, a harder work is to set the exact edge.
kind regards