Jon Osterman: developing new penetration codes for TrackJack

[Jon Osterman]

(cross-post alert: this post is being posted simultaneously on Don?s Domain and on the TrackJack support forum)

Viktor has asked me to develop new penetration codes for TrackJack.

this is a thorny issue that personally i don't feel warrants much attention. never do i shop around in CBJN or TrackJack or BlackJack Forum for games with good penetration; just look at the rules set and do your own legwork finding deep cuts. depending on dealer, shift, day of the week, phase of the moon, etc, etc, you will find a wide variety of cuts regardless of what anyone reports. some want more precise codes; i can't help but think that by using more 'precision' , TrackJack (and its competitors) are doing a disservice to the readers.

but never mind all that. i will list some of the concerns that come to mind about penetration codes. please let me know if i've neglected anything. maybe we'll take an informal survey to see what most people want; i thoroughly expect that no matter what is decided, many people are going to be unhappy (how does that go... something about pleasing all the people?)

concerns:

penetration codes should be short. no more than three characters. the code needs to fit neatly on the TrackJack report page.

penetration codes should be able to report a range of cuts.

penetration codes should be able to map neatly onto the SCORE charts for when that addition to TrackJack functionality goes live.

the existing codes should be able to map onto the new codes for an easy mass conversion.

i?ve already come up with an idea of how TrackJack penetration codes should look. before i finalize that notion and pass it on to Viktor (and you) i wanted to solicit your collective thoughts and incorporate them into the final product.

thank you for your attention.

[Don Schlesinger]

> this is a thorny issue that personally i
> don't feel warrants much attention. never do
> i shop around in CBJN or TrackJack or
> BlackJack Forum for games with good
> penetration; just look at the rules set and
> do your own legwork finding deep cuts.
> depending on dealer, shift, day of the week,
> phase of the moon, etc, etc, you will find a
> wide variety of cuts regardless of what
> anyone reports. some want more precise
> codes; i can't help but think that by using
> more 'precision' , TrackJack (and its
> competitors) are doing a disservice to the
> readers.

Yes and no. Pen changes daily, in every casino, everywhere. But, there ARE house guidelines and policies. You have to start somewhere. If you refused to report any pen at all, I think users would be outraged, and rightfully so. Some people are pressed for time and need to go straight to where the best games, purportedly, are. Trackjack can be a starting place for that guidance.
> maybe we'll take an
> informal survey to see what most people
> want; i thoroughly expect that no matter
> what is decided, many people are going to be
> unhappy (how does that go... something about
> pleasing all the people?)
The majority really should rule here. Trackjack needs to serve its users. If what you're doing isn't making the majority happy, then you're doing something wrong!

> penetration codes should be short. no more
> than three characters. the code needs to fit
> neatly on the TrackJack report page.

Einstein said, "Everything should be as simple as possible, but no simpler." The codes should be as short as possible, provided they can convey the information that users want. If you skimp on a character or two, and the proper information doesn't get conveyed, then you may be "neat," but you won't be effective.

> penetration codes should be able to report
> a range of cuts.

The house average (per type of game) should be the middle of the range, and the best and worst cuts observed should be the extremes. So, say, 60-80 would mean average pen is 70%, but you saw an occasional 60%, but also an occasional 80%.

> penetration codes should be able to map
> neatly onto the SCORE charts for when that
> addition to TrackJack functionality goes
> live.

Fine.

> the existing codes should be able to map
> onto the new codes for an easy mass
> conversion.

Maybe. Just worry about creating something that will be great for the future. If you try to conform to the past, to make things easier, you risk creating an inferior product.

> i?ve already come up with an idea of how
> TrackJack penetration codes should look.
> before i finalize that notion and pass it on
> to Viktor (and you) i wanted to solicit your
> collective thoughts and incorporate them
> into the final product.

I'd have five characters, as noted above, with the range expressed as a percentage, centered around the average pen. One problem: the range may not be symmetrical. What should we do if the average pen is 75% and we never see better than 80, but often see as bad as 65? 65-80 would be misleading, because the average (75) can't be surmised from the range.

Don

[Joe Miner]

> I'd have five characters, as noted above,
> with the range expressed as a percentage,
> centered around the average pen. One
> problem: the range may not be symmetrical.
> What should we do if the average pen is 75%
> and we never see better than 80, but often
> see as bad as 65? 65-80 would be misleading,
> because the average (75) can't be surmised
> from the range.

The "expression" should refer to a specific number with a tolerance. I would not like to see subjective symbols such as is currently used in BJF (g,b,u). What is a bad or unexceptional penetration for 6 deck may be good penetration for 2 deck. If pen is reported as an average percentage then we can decide if it is good, bad or unexceptional.

What is wrong with reporting a two digit number as a percentage that represents an average of the number of cards seen or if a cut card is used, the number of cards in front of the cut card?

[Don Schlesinger]

> What is wrong with reporting a two digit
> number as a percentage that represents an
> average of the number of cards seen or if a
> cut card is used, the number of cards in
> front of the cut card?

While there's nothing wrong with this, it doesn't give a good idea as to whether shopping can yield significantly better results, or not. A range gives an idea of the "volatility" of pen.

It answers the question: "Is this pretty much the norm at almost every table (where notches are used, for example), or is there wide variability around the mean that you're reporting?"

Don

[Joe Miner]

[Jon Osterman]

thanks to all who have given me some input on this matter.

here is what i propose: a four character code in the form of ##M# (a two digit penetration code, a modifier, and a 1 digit value for the modifier).

some examples...

75?5 would indicate an average of 75% penetration with a range of 70% to 80%

78?3 would indicate an average of 78% with a range of 75% to 81%

60?0 (reading the '0' as a '10') indicates a range of 50% to 70%

75+6 indicates generally 75% penetration but sometimes as good as 81%

60-0 (once again, '0' reads as '10') indicates generally 60% penetration but sometimes as bad as 50%.

legal modifiers therefore include: +, -, and ?
and the last digit is 1-0 meaning one through ten.

feedback?