Page 2 of 2 FirstFirst 12
Results 14 to 22 of 22

Thread: MJ: Standard Error in Simulations - Are You Being Fooled?

  1. #14
    Norm Wattenberger
    Guest

    Norm Wattenberger: CVCX Standard Error Update

    I've posted an update to CVCX that includes Standard Errors for EV, Win Rate and SCORE. The Customize Columns screen can be used to add or remove these stats. Note: SE will be somewhat lower than that displayed by CVData depending on penetration. This is because CVCX actually runs far more hands than you ask it to. For example, the 2 billion round sims in the library actually have over 3 billion rounds for some penetrations. This is because CVCX is a parallel (the parallel) simulator, simulating multiple penetrations at once and therefore must sim additional rounds at some penetrations to ensure that other penetrations have adequate rounds.


    Serious Blackjack Software

  2. #15
    MJ
    Guest

    MJ: Re: That was quick

    I thought there was no way to calculate the SE for SCORE? You sure work fast!

    If the SE for WR is $1/Hr, and the sample WR is $40/Hr, how do we use SE to get an idea of far we are from the theoretical WR?

    If we chart a bell curve with a '?' where the theoretical mean lies, and then draw +/- 3SE's from the center, can we then infer that our sample WR of $40 has a 99.7% chance of being within +/- $3 from the theoretical mean (whatever that number may be)?

    Nice job with the update! Many minor issues were fixed.

    MJ

  3. #16
    Norm Wattenberger
    Guest

    Norm Wattenberger: Re: That was quick

    > I thought there was no way to calculate the SE for
    > SCORE? You sure work fast!

    Just said it was problematic. And it is an estimate that only works with a large enough sample set. Conveniently, CVCX will not allow you to run a small sample set.

    > If the SE for WR is $1/Hr, and the sample WR is
    > $40/Hr, how do we use SE to get an idea of far we are
    > from the theoretical WR?

    > If we chart a bell curve with a '?' where the
    > theoretical mean lies, and then draw +/- 3SE's from
    > the center, can we then infer that our sample WR of
    > $40 has a 99.7% chance of being within +/- $3 from the
    > theoretical mean (whatever that number may be)?

    I think you're over-analyzing. As I mentioned, you will see polls taken that differ by more than double the SE. Albeit polls are inherently less reliable.

  4. #17
    MJ
    Guest

    MJ: Why did you remove Hands to Double from Columns?

    That was valuable info! If you want to remove something, remove DI.

    MJ

  5. #18
    Norm Wattenberger
    Guest

    Norm Wattenberger: Re: Why did you remove Hands to Double from Columns?

    It's a duplicate as it's in one of the widgets

  6. #19
    MJ
    Guest

    MJ: List of Suggestions for CVCX

    > It's a duplicate as it's in one of the widgets

    Here is the final list:

    HTD is NOT in any of the widgets. The closest thing is the Hours to Gain Widget. However, it does not contain HTD. You user could take the Hours to Gain 100% and then multiply that number by Hands/Hr to arrive at HTD, but why make the user do something that was already there to begin with? If it ain't broke don't fix it! I suggest you add it back to columns as an option. Trust me.

    Another great idea is to add another widget identical to Hours to Gain (maybe call it Hours to N0). It should tell the user how many Hours/Days is required to reach N0 given expenses or lack thereof. This would be useful, add it to the list.

    Oh before I forget, the Chart It Feature should be modified to allow comparisons of backcounting for two systems with unrelated counts (IE KO vs Hi-Lo). The only way to compare them right now is to set the wong in point at the "same" number, which makes no sense at all. With KO I enter at -4 and with Hi-Lo I enter at +2. Add it to list.

    Allow for comparison of one custom bet schedule vs another custom bet schedule on the Chart-It feature for 2 different systems.

    Allow for comparison of optimal vs custom bets for WR on the Chart-It Feature (right now it is just set for SCORE).

    Allow switching back and forth between 1 and 2 hands for KO-P.

    The cool thing about these suggestions is that they empower the user with the option to implement them if he so desires.

    Other than the aforementioned suggestions as well as integrating the Expense widget with all the other features, I think I am entirely out of suggestions. Nothing of major importance is left to be done. Thanks for listening and good luck!!

    MJ

  7. #20
    Norm Wattenberger
    Guest

    Norm Wattenberger: Re: List of Suggestions for CVCX

    > Here is the final list:

    > HTD is NOT in any of the widgets. The closest thing is
    > the Hours to Gain Widget. However, it does not contain
    > HTD. You user could take the Hours to Gain 100% and
    > then multiply that number by Hands/Hr to arrive at
    > HTD, but why make the user do something that was
    > already there to begin with? If it ain't broke don't
    > fix it! I suggest you add it back to columns as an
    > option. Trust me.

    > Another great idea is to add another widget identical
    > to Hours to Gain (maybe call it Hours to N0). It
    > should tell the user how many Hours/Days is required
    > to reach N0 given expenses or lack thereof. This would
    > be useful, add it to the list.

    Time to N0 and time to double are about the same.

    > Oh before I forget, the Chart It Feature should be
    > modified to allow comparisons of backcounting for two
    > systems with unrelated counts (IE KO vs Hi-Lo). The
    > only way to compare them right now is to set the wong
    > in point at the "same" number, which makes
    > no sense at all. With KO I enter at -4 and with Hi-Lo
    > I enter at +2. Add it to list.

    You can do this with dome work if you also have Excel or CVData(CVSpread.) Right-click on the chart and use Copy to copy the data. Of course this takes some effort. The problem with creating an automatic feature is in coming up with a way for the user to specify what he wants.

    > Allow for comparison of one custom bet schedule vs
    > another custom bet schedule on the Chart-It feature
    > for 2 different systems.

    Same problem. How does the user specify these options without greatly complicating the interface? I do have an idea and will give it some thought.

    > Allow for comparison of optimal vs custom bets for WR
    > on the Chart-It Feature (right now it is just set for
    > SCORE).

    Might be interesting.

    > Allow switching back and forth between 1 and 2 hands
    > for KO-P.

    Ahh, just got finished doing the impossible and allowing this feature for balanced strategies and the request comes up for unbalanced strategies:-) Knew it was coming - but frankly I haven't figured out how to do it.

    norm

  8. #21
    MJ
    Guest

    MJ: Re: List of Suggestions for CVCX


    > Time to N0 and time to double are about the same.

    Are you kidding me? Time to double is about DOUBLE time to N0 and even GREATER if there are expenses. Just see link below.

    MJ



  9. #22
    Norm Wattenberger
    Guest

    Norm Wattenberger: Re: List of Suggestions for CVCX

    Ahhh, you have a point when Kelly is not 1. Thanks for the demonstration.

Page 2 of 2 FirstFirst 12

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •  

About Blackjack: The Forum

BJTF is an advantage player site based on the principles of comity. That is, civil and considerate behavior for the mutual benefit of all involved. The goal of advantage play is the legal extraction of funds from gaming establishments by gaining a mathematic advantage and developing the skills required to use that advantage. To maximize our success, it is important to understand that we are all on the same side. Personal conflicts simply get in the way of our goals.