Why Practice Makes Perfect

[Spike!]

Why practice? Why is it  so important? Here's why: Variance.

Variance is: "A measure of the spread of a statistical distribution about its mean or center. The distribution of your results over a set of sessions, or the swings in a positive or negative direction." In other words, variance is how far up and how far down a system can go, on average. The only way you can know this about your method of play is by lots and lots of practice.

How will you know whats going on with your play unless you practice? If you don't know what the variance for your system is, how can you go to the casino with any confidence. In practice you learn what to expect without wagering any real money. You get to see your system at its worst and at its best. The more you practice, the better you get at playing your game.

There are some players that believe you should play your game as much as humanly possible, in the casino and out. Its all one big session, whether you're at the table or using your wheel or actuals at home. A session in the casino is just an extension of a practice session at home, you know what to expect and are never surprised at the results. How many players do this? Not many. If in practice you get huge swings of wins and losses, why would you want to play that for real money? Keep searching for a method that has tight variance under most conditions, otherwise save your cash for something more worthwhile, like a boat or hard drugs. Stay away from the casinos.

[The Spiders Kiss]

Good post Spike, thanks

TSK

[Bayes]

To the extent that reducing variance is a worthy goal, I agree with Spike, but to suggest that you can do it by 'practice' is absurd and misleading.

Keep searching for a method that has tight variance under most conditions,

Why not just say so in the first place?

'practice' is necessary and appropriate if you're learning to, say, play the violin. A teacher can point the way, but you have to put in the practice yourself to see results. You can study the techniques of the world's greatest violin players, but that won't put you in their league unless you put the hours in.

It's different with roulette (at least, the kind of systems we're talking about here). Roulette isn't a 'skill' in the sense that playing a violin is. When learning to play the violin, there are many continuous variables which you need to pay attention to. What I mean by 'continuous' is that a small change in any of them can have an effect on the sound you make. You could also call the 'system' of VIOLIN + PLAYER an 'analog' system. For example, the angle and pressure of the bow on the string, the pressure of your finger on a string etc.  This makes playing the violin more of an art than a science. Same with other PHYSICAL skills like gymnastics.

Roulette has potentially many 'variables' too, but it doesn't require complex hand-eye co-ordination, split second timing, or simultaneous monitoring of self-directed activity in the way that the above mentioned activities do. Any 'skill' required for winning at roulette is purely skill in DECISION, not a 'physical' skill.

You either have a bet which wins (or reduces the variance), or you don't. If you really could reduce the variance (even by a small amount), you would have the holy grail. Suppose you found a method which limited your longest losing run on the even chances to only 7. If everything else stayed the same, you would win flat betting.

Programmers might like to try the following experiment: get a few thousand spins, then remove all black streaks greater than 7, and run the simulation again betting on Red. What is your edge? you could try this with different length streaks to see how the edge varies.

It's possible to systematically search for 'solutions' because the number of outcomes (and 'inputs') are strictly limited. In this sense ROULETTE + PLAYER is a 'discrete' system rather than an analog one.  You don't have to conduct your research systematically, you could just play around, hoping to stumble on something worthwhile. But in any case, calling it 'practice' is not appropriate. You can keep searching for a method which wins, and if you find it then 'practice' is redundant.

If you have a combination lock, and have forgotten the combination, but after trying many numbers you eventually hit on the correct one, do you then have to 'practice' every time you want to open the the lock? of course not, that would be a meaningless and absurd suggestion.

[Jack Wad]

Nice to know Spike's still full of C. R. A. P. .

(Classic Roulette Advice Posts)

Keep up the good work,


Jack Wad. .

[Bayes]

In any case, trying to shorten the variance on its own isn't enough, for the same reason that just looking at the probability isn't enough. You can reduce the variance merely by betting on more numbers. You need reduce the variance in relation to the expected variance in order to make a profit. Is there any reason why it should be easier to do this than increase the probability in relation to the payout? it really amounts to the same thing.

[Rheti]

You have a Wheel and a Sheet.

Let's divide the Wheel in 3 sections of twelve.. forget the zero.

Divide the sheet in 3 pieces of twelve like:

Angels: first six last six +zero =13
Q3: 10-11-13-14-17-18-20-21-25-26-28-29 + zero = 13
Other: 7-8-9-12-15-16-19-22-23-24-27-30 +  zero = 13

So we have Angels-Other 25 numbers
Angels-Q3 25 numbers
Other-Q3 25 numbers

Propability 25/37 = 67.6% .

The 3 sections are the Sources, the 3 sequences of the sheet the target. It's a follow-up system.

Let us take a look at 5 days of german spinning.

ANG-OTH   DAY-1      
         
Som van freq   Kolomlabels      
Rijlabels   S1   S2   S3
L   26   19   23
W   68   52   51
Eindtotaal   94   71   74
         
actuals   2.6   2.7   2.2
recovery   2.3   2.3   2.3


ANG-Q3   DAY-1      
         
Som van freq   Kolomlabels      
Rijlabels   S1   S2   S3
L   30   32   24
W   64   39   50
Eindtotaal   94   71   74
         
actuals   2.1   1.2   2.1
recovery   2.3   2.3   2.3


OTH-Q3   DAY-1      
         
Som van freq   Kolomlabels      
Rijlabels   S1   S2   S3
L   35   17   22
W   59   54   52
Eindtotaal   94   71   74
         
actuals   1.7   3.2   2.4
recovery   2.3   2.3   2.3


ANG-OTH   DAY-2      
         
Som van freq   Kolomlabels      
Rijlabels   S1   S2   S3
L   11   21   18
W   35   50   48
Eindtotaal   46   71   66
         
actuals   3.2   2.4   2.7
recovery   2.3   2.3   2.3


ANG-Q3   DAY-2      
         
Som van freq   Kolomlabels      
Rijlabels   S1   S2   S3
L   19   17   20
W   27   54   46
Eindtotaal   46   71   66
         
actuals   1.4   3.2   2.3
recovery   2.3   2.3   2.3


OTH-Q3   DAY-2      
         
Som van freq   Kolomlabels      
Rijlabels   S1   S2   S3
L   16   29   26
W   30   42   40
Eindtotaal   46   71   66
         
actuals   1.9   1.4   1.5
recovery   2.3   2.3   2.3


ANG-OTH   DAY-3      
         
Som van freq   Kolomlabels      
Rijlabels   S1   S2   S3
L   17   17   16
W   39   51   42
Eindtotaal   56   68   58
         
actuals   2.3   3.0   2.6
recovery   2.3   2.3   2.3


ANG-Q3   DAY-3      
         
Som van freq   Kolomlabels      
Rijlabels   S1   S2   S3
L   18   24   18
W   38   44   40
Eindtotaal   56   68   58
         
actuals   2.1   1.8   2.2
recovery   2.3   2.3   2.3


OTH-Q3   DAY-3      
         
Som van freq   Kolomlabels      
Rijlabels   S1   S2   S3
L   19   26   23
W   37   42   35
Eindtotaal   56   68   58
         
actuals   1.9   1.6   1.5
recovery   2.3   2.3   2.3


ANG-OTH   DAY-4      
         
Som van freq   Kolomlabels      
Rijlabels   S1   S2   S3
L   25   22   20
W   41   54   58
Eindtotaal   66   76   78
         
actuals   1.6   2.5   2.9
recovery   2.3   2.3   2.3


ANG-Q3   DAY-4      
         
Som van freq   Kolomlabels      
Rijlabels   S1   S2   S3
L   18   27   31
W   48   49   47
Eindtotaal   66   76   78
         
actuals   2.7   1.8   1.5
recovery   2.3   2.3   2.3


OTH-Q3   DAY-4      
         
Som van freq   Kolomlabels      
Rijlabels   S1   S2   S3
L   17   26   25
W   49   50   53
Eindtotaal   66   76   78
         
actuals   2.9   1.9   2.1
recovery   2.3   2.3   2.3


ANG-OTH   DAY-5      
         
Som van freq   Kolomlabels      
Rijlabels   S1   S2   S3
L   28   39   20
W   41   59   51
Eindtotaal   69   98   71
         
actuals   1.5   1.5   2.6
recovery   2.3   2.3   2.3


ANG-Q3   DAY-5      
         
Som van freq   Kolomlabels      
Rijlabels   S1   S2   S3
L   14   26   25
W   55   72   46
Eindtotaal   69   98   71
         
actuals   3.9   2.8   1.8
recovery   2.3   2.3   2.3


OTH-Q3   DAY-5      
         
Som van freq   Kolomlabels      
Rijlabels   S1   S2   S3
L   26   29   25
W   43   69   46
Eindtotaal   69   98   71
         
actuals   1.7   2.4   1.8
recovery   2.3   2.3   2.3


What's the variance ?

Day-1 : recovery 2.3  lowest 1.2 factor = -/- variance of -/- 1.1 //  highest 3.2 = +/+ 0.9
Day-2 :recovery 2.3   lowest 1.4 factor = -/- variance of -/-0.9  //  highest 3.2 = +/+ 0.9
Day-3 :recovery 2.3   lowest 1.5 factor = -/- variance of -/-0.8 //   highest 3.0 = +/+ 0.7
Day-4 :recovery 2.3   lowest 1.5 factor = -/- variance of -/-0.8 //   highest 2.9 = +/+ 0.6
Day-5 :recovery 2.3   lowest 1.5 factor = -/- variance of -/-0.8 //   highest 3.9 = +/+ 1.6

The only thing is how to find the trigger ? lol

[Noble Savage]

Nice to know Spike's still full of C. R. A. P. .

(Classic Roulette Advice Posts)

Keep up the good work,

Very true.

[I have cookies]

Nice to know Spike's still full of C. R. A. P. .

(Classic Roulette Advice Posts)

Keep up the good work,


Jack Wad. .

[Spike!]

>>To the extent that reducing variance is a worthy goal, but to suggest that you can do it by 'practice' is absurd >

Thats the other side of the coin. Screw practice, just come up with a method and run to the casino with it. Either way, thru practice or thru playing, you'll learn a lot about your methods variance.

[Noble Savage]

Um, your variance can be calculated and/or simulated easily by knowing the details of your bet (such as its edge, probability, and payout).

You just discovered the term "variance" Spike, thanks to NumberSix.

[Jakkalsdraai]

I like the saying "Correct practice make perfect"

No point in practicing something the wrong way......Just enhances bad habits. So yes practice makes perfect if it's the correct methods.

I saw on Discovery I think about that study about whether geniusses are just that or whether they were forced into a direction and because of thousands of hours they become masters. The studies concluded that it is all about the "time" that was put in. Something like 7000-10000 hours and you are there!

Cheers
Jackal

[suitcase charlie]

5 hours a day for 5 years brings you up to 9125 hours.
If you have not come up with something worth the effort by then, it could be time to look for another hobby.
Seriously though, if you think about the game of roulette, a lot of the stuff we end up studying is just the same old. It is very difficult to come up with original ideas and keep it real without reverting back to all the old premises which never worked. I mean if you were working on a progression for streets 20 years ago and today you are looking at a progression for dozens, then you would have to question how far forward you have really came.

[Number Six]

Spike might have known about variance for a while, but what he doesn't understand is why it is impossible for it not to apply to him since whatever method he uses that considers past spins will perform no better or worse than random betting. The question is, can the variance of a classical mechanical bet be reduced in any way, and if not, can it at least be tamed?

The variance is typically exacerbated by negative progressions that very quickly spiral beyond all control when the wins dry up. It really doesn't take much deviation for these things to fail utterly. Variance is always reported to be much much worse that it really is or should be. So, can the effects of variance be reduced rather than the variance itself?

[Spike!]

You just discovered the term "variance" Spike>>

I first heard about variance in 1977 in BJ. Search the GG archives and I have threads I started on variance dating back to 2006.

[Spike!]

>>It is very difficult to come up with original ideas and keep it real without reverting back to all the old premises which never worked.>>

I fight that all the time. Unlearning bad habits is difficult, and a constant challenge. I think its laziness, its so easy to revert back and pretend you're not.

[Spike!]

-