Norm Wattenberger: Advantage by True Count

[Ouchez]

> Don right from the start took a patronizing
> attitude, dismissing my ideas out of hand
> and refusing to debate. He has done so in
> the past whenever I came up with an idea he
> didn't share.So,I felt entitled to express
> my displeasure about this and I don't think
> you can call this a personal attack.
> When I spoke about simulation freaks, I
> didn't mean you personally, I meant those
> who think, that because of the damatric
> increase of computer performance over the
> last few years, they can solve every issue
> by simulation and forget about the
> mathematical craftmanship.
> When you denied the linearity of Blackjack
> in spite of the fact that counting is based
> on just that I couldn't help thinking that
> this was because of your daily work and
> tried to express this in an amusing
> way.What's wrong with that?
> I made you even a compliment but soon
> afterwards you called one of my statements
> nonsense in the subject line. I call this
> rude and irreverent whether you like it or
> not.If you want to be treated with respect,
> you should start with showing respect for
> others including their ideas, especially if
> these ideas are based on blackjack truismes
> that have long been established.
> For me, the findings of Dr. Thorp and Peter
> Griffin are still valid in spite of the
> so-called new "discoveries" and
> they will still be valid in a 1000 years.So,
> don't try to rewrite the laws of blackjack.
> It would be blasphemy.

> Francis Salmon

or attacks in this thread as all sides should be respected. I know that Francis has been around for a very long time and is a very serious player, you might call him "The Real Deal".

Don and Norm have their strengths and are true "BJ Celebes", having added much to the game in different theaters of expertise.

So, all I am saying, is let's respect one another and those ideas presented to beat the game of BJ.

I know I like to hear all sides.

Ouchez.

[Norm Wattenberger]

> The correct index for insurance can be found
> by solving some mathmatical equations for
> any number of decks.I did this by writing a
> special program and my numbers coincide
> exactly with Wong's numbers.

Wong uses sampling by replacement which is known to be inaccurate. For the same reasons that your numbers are inaccurate. They depend upon "representative subsets," a shortcut method that is very fast but incorrect. Stanford had an excuse. He wrote his software when machines were a tiny fraction of current speed. And they were 'accurate enough' for his time. That excuse is no longer valid. We can develop far more accurate indexes. Not that it matters much in shoes

[Brick]

Norm if you were to run the same sims over again is it possible the curve will show different results at tc intervals due to the randomness?

[Francis Salmon]

I don't know what sampling by replacement is and it is certainly not the method I used. An optimal index for high-low has to obey certain mathmatical conditions related to the count and to the missing ace. When you solve this you get exactly one optimal insurance point for a given number of decks.There is no need for simulation at all.
By the way, did Wong himself recognize that his numbers are inaccurate or are just you and your cronies saying that?What would be a more accurate number for 6 decks?

Francis Salmon

[Norm Wattenberger]

Possible. I only ran one billion. I would need to run 20 billioin to know for certain.

[Brick]

Sorry,I seem to be asking too many questions today,but what would happen if you re-ran only a billion rounds. Would the results be the same?

> Possible. I only ran one billion. I would
> need to run 20 billioin to know for certain.

[Norm Wattenberger]

Very close on the lower counts. A bit different on the higher counts.

[MGP]

> The correct index for insurance can be found
> by solving some mathmatical equations for
> any number of decks.I did this by writing a
> special program and my numbers coincide
> exactly with Wong's numbers. For 6 decks
> it's 3.0 (not floored!)and it's valid for
> any level of penetration. You don't need to
> bother about subsets here.

I'm sorry but here you are wrong. If you want exact indices, even for insurance - you have to use all the subsets. We covered this over on BJMath awhile back. The post with the final 100% pen values and their comparison to Griffin's and others' linear estimates are included in the post.

I showed already in my other post on the Beginner's page that insurance calculations are not linear but they are close to linear.

This is not speculation, estimation or a bias - it is a fact and I can produce any numbers you need to see that for any number of decks with hi-lo. Only a simple 10-count is perfectly, linear. Those values can be found in the thread as well. Other counts cannot be linear because of the non-linearity of the subsets.

Sincerely,
MGP

Exact Insurance Indices

[Francis Salmon]

Thank you for the link.The figures coincide to the decimal with mine and Wong's. This proves that the algebraic solution which assumes linearity of distribution is absolutely valid.
It confirms me in my belief that the assumption of linearity is generally justified for all practical purposes.
It's strange, I get scoffed at for using decimals and then I see people calculating EV's and indexes to the tenth position after the coma!
If some here tell me my decimals are of waste of time, what do they say to you?
I don't see what should be wrong about what I said before. I never claimed perfect linearity and you know that, I just say linearity is a good assumption.

Francis Salmon

[BJinNJ]

What are the implications for this data
regarding Kelly betting? It seems that
an extra .1 advantage is 20% higher than
the 'rule of thumb', and it could affect
proper betting.

> Many of the books talk about a rule of thumb
> that advantage increases .5% (or some other
> number) for each True Count. Before we had
> optimal bet calculators, this rule of thumb
> was very useful in determining bet size. I
> think that it is important to understand
> that, like all rules of thumb, these are
> just a rough estimates. Of course I like to
> show everything in color So, I created the
> chart below. In this chart, the gain in
> advantage from one TC to the next is plotted
> on the Y axis and the ending TC is on the x
> axis. That is, each point shows the increase
> in advantage from one TC to the next higher
> TC. Four different sets of rules and
> penetration are displayed. As you can see,
> the gain in advantage from one TC to the
> next tends to bounce around. HiLo Ill18 was
> used.

> Note, the exact same strategy, indexes,
> deck estimation accuracy, and TC calculation
> methodolgy were used in all four sims.
> Varying these would of course add to the
> variability in the chart. I am not saying
> that the old rule of thumb is useless. But
> since the rule is stated so often as gospel,
> I thought it worth pointing out that it
> isn't precise.