-
MJ: A quote from an MIT player
> For increased remaining card estimates to be of full
> value, you need to generate the indexes using the same
> accuracy.
Here is what I was told in an email from a member of the MIT Blackjack Team when I asked about index generation:
"As for our numbers plays, whether you round to the nearest 1/4 deck or 1/2 deck does not affect the true count at which a particular play is appropriate. How accurate your deck estimation simply affects your true count accuracy, not the index numbers themselves. Our numbers simulations calculated the precise CE numbers".
With all due respect Norm, this seems to contradict your comment above. I guess the MIT teams ran their simulations and calculated the CE numbers using EXACT card resolution and then required their BPs to use 1/4 deck estimation to try and get a very accurate TC. You and Don seem to maintain that index numbers must be generated using the same accuracy as your discard estimation to get maximum benefit. If that is the case, howcome the MIT Blackjack Team doesn't take the same approach? They have some guys that are quite brilliant as well! :-)
MJ
-
MJ: Re: Fascinating!!! Don and Norm please read my comments on discard estimati
> In your examples, you would have had your max bet out
> no matter which resolution you used, and it would have
> made little or no difference.
Not true. I am learning MIT style Hi-Lo. The way those guys calculate the bet is described below:
Bet = (TC-1)* unit. They also round (not floor) the TC to the nearest 1/2 unit.
The bet schedule is as follows:
TC <= 1.5 bet 1/2 unit
TC = 2 bet 1 unit
TC = 2.5 bet 1.5 units
TC = 3 bet 2 units
etc.....
Using this strategy would require a TC of +7 to place a maximum bet. Clearly, this increased accuracy would mean wagering different amounts at a TC of +5, +5.7, and +6.7.
> Quarter-deck resolution in the shoe game is, more or
> less, overkill and quite useless. Half-deck might
> slightly outperform full-deck. Norm and I did some
> studies on this recently. I think half-deck won 1-2%
> more, tops.
Frankly, I'm surprised 1/4 deck estimation does not generate more SCORE then 1/2 deck. In the second half of the shoe it might a difference, especially in the last 1/3 of the shoe as my example demonstrates. Now what if the penetration was slightly deeper and the RC was even higher then +10?
MJ
-
Don Schlesinger: Re: A quote from an MIT player
> Here is what I was told in an email from a member of
> the MIT Blackjack Team when I asked about index
> generation:
> "As for our numbers plays, whether you round to
> the nearest 1/4 deck or 1/2 deck does not affect the
> true count at which a particular play is appropriate."
Not true.
> How accurate your deck estimation simply affects your
> true count accuracy, not the index numbers themselves.
> Our numbers simulations calculated the precise CE
> numbers".
"Precise CE numbers" is quite meaningless. Over and over, people fool themselves into thinking they've invented a better mousetrap. If you play differently from the way you generate the indices, you actually have less accuracy, not more.
> With all due respect Norm, this seems to contradict
> your comment above.
Right. It contradicts what Norm said. But it most certainly doesn't make him wrong. :-)
> I guess the MIT teams ran their
> simulations and calculated the CE numbers using EXACT
> card resolution and then required their BPs to use 1/4
> deck estimation to try and get a very accurate TC.
Foolish waste of time. It could be easily demonstrated by simulation that such an approach wouldn't buy you a penny more SCORE than calculating indices to half-deck resolution and then playing that way.
> You and Don seem to maintain that index numbers must be
> generated using the same accuracy as your discard
> estimation to get maximum benefit.
We don't claim it; it's a fact.
> If that is the
> case, how come the MIT Blackjack Team doesn't take the
> same approach? They have some guys that are quite
> brilliant as well! :-)
You're asking the wrong people, aren't you? :-) Shouldn't you be asking the MIT people?
Don
-
Don Schlesinger: Re: Fascinating!!! Don and Norm please read my comments on discard estimati
> Not true. I am learning MIT style Hi-Lo.
Be my guest. In case you don't know, they didn't invent Hi-Lo. :-)
> The way those
> guys calculate the bet is described below:
> Bet = (TC-1)* unit. They also round (not floor) the TC
> to the nearest 1/2 unit.
To each his own.
> The bet schedule is as follows:
> TC TC = 2 bet 1 unit
> TC = 2.5 bet 1.5 units
> TC = 3 bet 2 units
> etc.....
There is an advantage to calculating TC to a half-point, to gain a better betting ramp.
> Using this strategy would require a TC of +7 to place
> a maximum bet. Clearly, this increased accuracy would
> mean wagering different amounts at a TC of +5, +5.7,
> and +6.7.
It is probably a bad mistake to wait until +7, using Hi-Lo, to place a max bet. Regardless of resolution, you are dividing by whole decks. In a 6-deck shoe game, TCs of +7 or higher occur, for 4.5/6, 0.84% of the time. So, what's the point?? You say you're using a certain spread, but you actually attain the max bet once every 119 hands? Who's kidding whom??
> Frankly, I'm surprised 1/4 deck estimation does not
> generate more SCORE than 1/2 deck.
Again, you seem to equate "I'm surprised" with "you must be wrong." I'm not trying to convince you of anything. Run the sims yourself. Then, report back.
> In the second half
> of the shoe it might a difference, especially in the
> last 1/3 of the shoe as my example demonstrates. Now
> what if the penetration was slightly deeper and the RC
> was even higher than +10?
You can formulate all the "what ifs" you like. I'm telling you that it doesn't matter. But, don't take my word for it.
Don
-
Don Schlesinger: Max bet
Just looked at a few optimal bet ramps. In a 4.5/6 game, max bet should be placed at +4. In 5/6, at +5. Waiting for +7 is ridiculous.
Don
-
MJ: Is that for flooring or rounding....
I stated that the MIT Team uses rounding for the TC. I would venture to say that the optimal bet ramps you posted are for FLOORING the TC. Obviously, with flooring you have a greater advantage per each unit of TC as compared to rounding. That is why your ramp mandates placing a max bet at a lower TC.
Now, do you have any optimal bet ramps for when you ROUND the TC? If so, please go ahead and post it. Otherwise, the comparison you are trying to make is invalid. Also, was the statistic you sighted of a TC frequency of +7 occuring 1 in 119 hands for a floored or rounded TC? If it assumes flooring, do you know how often a TC of +7 would occur when using rounding? Thanks Don!
MJ
> Just looked at a few optimal bet ramps. In a 4.5/6
> game, max bet should be placed at +4. In 5/6, at +5.
> Waiting for +7 is ridiculous.
> Don
-
MJ: Forgot to ask
> It is probably a bad mistake to wait until +7, using
> Hi-Lo, to place a max bet. Regardless of resolution,
> you are dividing by whole decks.
If resolution is 1/4 deck, how do you figure I am dividing by
"whole" decks?
MJ
-
Francis Salmon: Re: Max bet
> Just looked at a few optimal bet ramps. In a 4.5/6
> game, max bet should be placed at +4. In 5/6, at +5.
> Waiting for +7 is ridiculous.
> Don
I assume the MIT team defined a unit as being the smallest advantage bet.The bet for negative situations would just be a fraction of this.
Otherwise the bet ramp would indeed be ridiculous and I don't see how they could have made all their money like that.
Francis Salmon
-
Jay: KO Even More Simplified?
Norm, thanks for your post and sims. It seems to back up the simplified version of KO I have been using the last few years. The difference is that I have been using the pivot point (+4) instead of +2 as my index point. I only use a few indeces, and it plays for single, double, and 6-deckers:
12 vs 2, 3
15. 16 vs 10
Insurance
For 6 deckers w/ surrender
15 vs 9, A
I do make an adjustment on single deck, changing the key/pivot to 3/5 from 2/4. I start out with 2 units, then keep it at 2 at +2, double at +3, half it at +1 or less.
Some of your 6-deck indeces of course are basic strategy for 1 & 2 decks. Would my described strategy increase in efficiency if I changed the index point to Pivot -2? Thanks again.
Jay
-
-
MJ: Correct
> I assume the MIT team defined a unit as being the
> smallest advantage bet.The bet for negative situations
> would just be a fraction of this.
That is why they bet 1 unit at TC = +2 (.5% advantage) and 1/2 unit when the TC <= 1.5 (do not have advantage). These figures are for rounding and not flooring.
MJ
-
Norm Wattenberger: Re: KO Even More Simplified?
> Would my described
> strategy increase in efficiency if I changed the index
> point to Pivot -2? Thanks again.
Hard to say. I ran 'scores' of sims to find the sweet spots. Theory, calculations, and intuition are all of great value. But when it comes down to fine-tuning simplification; trial and error is still required.
-
Francis Salmon: Re: A quote from an MIT player
> Here is what I was told in an email from a member of
> the MIT Blackjack Team when I asked about index
> generation:
> "As for our numbers plays, whether you round to
> the nearest 1/4 deck or 1/2 deck does not affect the
> true count at which a particular play is appropriate.
> How accurate your deck estimation simply affects your
> true count accuracy, not the index numbers themselves.
> Our numbers simulations calculated the precise CE
> numbers".
> With all due respect Norm, this seems to contradict
> your comment above. I guess the MIT teams ran their
> simulations and calculated the CE numbers using EXACT
> card resolution and then required their BPs to use 1/4
> deck estimation to try and get a very accurate TC. You
> and Don seem to maintain that index numbers must be
> generated using the same accuracy as your discard
> estimation to get maximum benefit. If that is the
> case, howcome the MIT Blackjack Team doesn't take the
> same approach? They have some guys that are quite
> brilliant as well! :-)
> MJ
You and this MIT member are making valid points and it's sad to see that these are only met with platitudes and contempt from Norm and Don.
They delude themselves into thinking that they can correct the imprecision of an index by adding more inaccuracy in deck estimation and TC calculation in general.
Please,keep your critical mind.It's needed on these pages.
Francis Salmon
Posting Permissions
- You may not post new threads
- You may not post replies
- You may not post attachments
- You may not edit your posts
-
Forum Rules
Bookmarks