Author Topic: Bankroll Management  (Read 1223 times)

0 Members and 1 Guest are viewing this topic.

McCoy

  • New
  • **
  • Posts: 95
  • Thanked: 59 times
Re: Bankroll Management
« Reply #15 on: September 24, 2017, 11:15:22 AM »
Here are the results of using 2 staking plans betting on 4 numbers with an edge over 1500 spins. I cheated somewhat in this because I picked 4 numbers which were hitting above expectation. It doesn't matter because the point was to show the difference  in the results. The blue plot shows the result of using a classic d'Alembert raising the stake by 1u after each 8 losses and reseting to 1u on a new high bank. The red line uses 2% of bank. The starting bank is 250u for both. Notice the different scales on the vertical axes.  :)


« Last Edit: September 24, 2017, 11:16:53 AM by McCoy »
 
The following users thanked this post: kav, Reyth

MrPerfect.

  • Hero Member
  • ******
  • Posts: 1309
  • Thanked: 763 times
Re: Bankroll Management
« Reply #16 on: September 24, 2017, 11:28:42 AM »
Why would you cheat yourself in curve fitting?
  If you already started, why not pick up proper percentage of bankroll to bet on second case? Or these 4 number bet on average provide only 2 % of advantage???
 

McCoy

  • New
  • **
  • Posts: 95
  • Thanked: 59 times
Re: Bankroll Management
« Reply #17 on: September 24, 2017, 11:37:03 AM »
The point of the demo is to show the different results from using the two staking plans when you have an edge, that's all. I wasn't concerned with bet selection and I could have just artificially increased the payoffs but I wanted to use a more 'realistic' example. If you're interested here are the spins I used. I just looked at a barchart and picked some adjacent numbers which had a higher than average expectation. They were 26 - 29 inclusive.

 
The following users thanked this post: scepticus, Reyth, MrPerfect.

MrPerfect.

  • Hero Member
  • ******
  • Posts: 1309
  • Thanked: 763 times
Re: Bankroll Management
« Reply #18 on: September 24, 2017, 11:42:02 AM »
It's expected that different betting produce different results, but on your graph it looks too similar. Can you artificially duplicate a sample , so we can actually see a difference if there is any? If not , it looks like whatever you do, result can be same after many spins ( long run...)
 

McCoy

  • New
  • **
  • Posts: 95
  • Thanked: 59 times
Re: Bankroll Management
« Reply #19 on: September 24, 2017, 02:00:52 PM »
It's expected that different betting produce different results, but on your graph it looks too similar.

What do you mean? There is a huge difference, that's why I said notice there are two different vertical axes. Both plans were executed on the same set of spins but the 2% resulted in a final bank of 19,560u and the d'Alembert finished with 1,314u.
 

MrPerfect.

  • Hero Member
  • ******
  • Posts: 1309
  • Thanked: 763 times
Re: Bankroll Management
« Reply #20 on: September 24, 2017, 02:28:56 PM »
McCoy, your chart confused the hell out of me.
   I didn't realise it was 2 difervent vertical axix for different graphs. Probably posting separate charts would be more visual, how do think?
 

McCoy

  • New
  • **
  • Posts: 95
  • Thanked: 59 times
Re: Bankroll Management
« Reply #21 on: September 24, 2017, 02:50:09 PM »
Here is the plot with both axes having the same scale.

 
The following users thanked this post: MrPerfect.

Reyth

  • Global Moderator
  • Hero Member
  • *****
  • Posts: 3934
  • Thanked: 1262 times
Re: Bankroll Management
« Reply #22 on: September 24, 2017, 04:33:59 PM »
I base my bankroll on the largest expected drawdown. 

Why, in the first chart, do the up-down movements of the two betting methods not match?  Is the horizontal axis the same?

Awesome work btw!  I think you are on the right track with groups of hot numbers!  1500 spins, impressive! :D

I am very interested to see a losing graph using both methods!  The key to a losing graph will be successive drawdowns.
« Last Edit: September 24, 2017, 04:37:41 PM by Reyth »
 
The following users thanked this post: MrPerfect.

McCoy

  • New
  • **
  • Posts: 95
  • Thanked: 59 times
Re: Bankroll Management
« Reply #23 on: September 25, 2017, 12:20:28 PM »
Why, in the first chart, do the up-down movements of the two betting methods not match?  Is the horizontal axis the same?

Yes there is only one horizontal axis. The movements of the 2 plots do track each other up to a point but it's easier to see this if you take logarithms of the banks. This compresses the numbers and is a useful technique when the scales or ranges are very different.



Quote
Awesome work btw!  I think you are on the right track with groups of hot numbers!  1500 spins, impressive! :D

Thanks but I'm only showing the results of the different staking plans here. I remain convinced that hot numbers are the way to go but as stated above, I chose the numbers for these plots only after knowing the results.

Quote
I am very interested to see a losing graph using both methods!  The key to a losing graph will be successive drawdowns.

I used the same set of spins but chose the numbers which result in losses, numbers 19-22.



Here is the result of betting on those numbers using the same staking plans. In this case because they are losing numbers I increased both banks to 4000 u which was the worst drawdown.



In this case the 2% plan comes off worst. This is consistent with the theory which says that in games with a negative expectation it's better to be 'bold' than 'timid'. No staking plan can save you in this case but perhaps chasing losses can keep you afloat for longer.
 
The following users thanked this post: kav, Reyth

MrPerfect.

  • Hero Member
  • ******
  • Posts: 1309
  • Thanked: 763 times
Re: Bankroll Management
« Reply #24 on: September 25, 2017, 12:45:17 PM »
Nice charts, Bro. 
 
The following users thanked this post: Reyth

scepticus

  • Hero Member
  • ******
  • Posts: 2078
  • Thanked: 431 times
Re: Bankroll Management
« Reply #25 on: September 25, 2017, 02:16:15 PM »

Thanks  Mc Coy
As you say, though, these are pre-chosen  Hot / Cold numbers and your charts are for illustration only.  .
Hot and Cold numbers  have been discussed periodically in the forum and, logically, Hot numbers is preferable to Cold numbers . Within any  37 spins there will be Hot numbers so we KNOW that there will be Hot numbers WITHIN those 37 spins. The problem is to work out a profitable method of exploiting this INFORMATION.
Hopefully your future  analyses  will arouse much interest in the forum.
Incidentally , what do you think of Mathematicians' belief that we cannot use past winning numbers because the wheel " has no memory "  ?
 
The following users thanked this post: Reyth

Reyth

  • Global Moderator
  • Hero Member
  • *****
  • Posts: 3934
  • Thanked: 1262 times
Re: Bankroll Management
« Reply #26 on: September 25, 2017, 03:00:58 PM »
Can I answer too??

 

scepticus

  • Hero Member
  • ******
  • Posts: 2078
  • Thanked: 431 times
Re: Bankroll Management
« Reply #27 on: September 25, 2017, 04:30:17 PM »
Oh ! Well !  If you must ! ;D
 

Reyth

  • Global Moderator
  • Hero Member
  • *****
  • Posts: 3934
  • Thanked: 1262 times
Re: Bankroll Management
« Reply #28 on: September 25, 2017, 05:59:20 PM »
The "wheel has no memory"?  :o

What about the LotT does it have a memory? NO.

So who brought up this memory business anyway?

The fact is we are dealing with 2 forms of "closed circuits" where the possible results are limited along with the statistical reality that practically speaking, Perfectly Balanced ImbalanceTM is the constant norm; i.e. the LotT MUST produce repeaters and therefore certain numbers must come in less than expected (practically speaking).

The LotT is a constant force that applies to every spin and once we have over 37 spins, every spin is a new LotT session.  Since we know that certain numbers must appear more than others in a SINGLE LotT session, the SAME concept must apply to multiple sessions that are chained together over a long-term; the size of the multiple sessions is the size of our random sequence.

Now, since we must have this phenomenon of certain numbers that must appear more than others because it is a closed circuit, the SAME concept must also apply to the felt because it too is a closed circuit.  Therefore, some of these numbers MUST appear within a defined area of the felt more than any other as well.

So no memory required, just simple statistics -- there must ALWAYS be a set of numbers that comes in more frequently than others and some of these must ALWAYS group together on the felt in a defined space more than any others.

Your predicted response:

Great Reyth but the question for the Queen is, CAN WE USE THIS INFORMATION TO MAKE MONEY!???

Yes.  If we follow some of the wisdom of our AP friends regarding their betting techniques, and apply some of our own learning to compensate for the lack of a physical bias, the statistical edge should make us profitable; like Harry J says, "we are using the Physics of Statistics(R)".
« Last Edit: September 25, 2017, 06:08:28 PM by Reyth »
 
The following users thanked this post: kav

scepticus

  • Hero Member
  • ******
  • Posts: 2078
  • Thanked: 431 times
Re: Bankroll Management
« Reply #29 on: September 25, 2017, 08:20:18 PM »

Reyth
" The wheel has  no memory " is  a Stock Phrase  of mathematicians . Together with the HE these are  the reasons they use to argue  that the HE cannot be beaten (  deep breath ! )  " In the Long Run " .
What I am trying to do is counter their arguments .
But- Yes- the question is " Can we use this Information or is it Useless Information. " ?
Hopefully this is a question McCoy is working on.