Question on BoneTracker

[heavy]

Well, I posted an elaborate (not) answer but my rural internet provider crapped out on me in mid upload, so I am just going to defer to the experts, Maddog and Mad Professor. Do you have to be "mad" to use Dice Tool? Of course not. But you do have to look at things a little differently than most of us do. DT was the brainchild of Stanford Wong and MP, and Maddog adapted it for use in BoneTracker. I'm sure that somewhere in the explanation the word correlation will come up - a favorite term of SW's - he also uses it in blackjack. But alas, I will say no more. Gentlemen?

[Mad Professor]

Hi Pointman,

If DiceTool is indicating a 1.2% edge/Bet on the pass/come; then it means that your roll-sampling has overcome the -1.4% house-advantage by 2.6%, and you now have a net-positive advantage of +1.2%.



MP

[heavy]

Would someone like to elaborate on the actual calculation of your advantage in Dice Tool? How does that 1.2% advantage relate to any percentage of improvement shown in BoneTracker proper?

[Maddog]

Pointman, just a “point” of caution. Use the DT HA/PA results carefully and as indicators, not as actual expectations.

DiceTool provides a model, an approximation based on your tracked foundation frequencies. Use it as an indication of what is possible and then test the theory with either practice/mock betting or through tracking in-casino play.

The whole DiceTool Foundation Frequency(FF) “mapping” versus the BoneTracker “tracking” thing is really easier than my poor convoluted explanations make it seem. Let me take a stab at a better explanation by using a specific example.

We know that DiceTool maps your FF values to a provided dice set. Lets use a P6 permutation as an example to highlight the mapping process. We tell DiceTool that we are setting the dice to 3/1-3/1, either through the entry on the Roll-Data tab or through the override on the DiceTool tab. Ok, based on this DiceTool can figure out the Primary Face combinations of the Set. So what are the PF combinations?
~ Tops: 3-3
~ Facing: 1-1
~ Bottoms: 4-4
~ Backs: 6-6

Now what does DiceTool work with? It works with FF values. Individual dice rolls are assigned to an FF category. For example we may tell DiceTool that we had 100 PF results in our data set. Do the FF values tell us how many of the PF hits were 3-3’s? No, all DiceTool knows is that there were 100 PF’s. What DiceTool does is takes an average, or an approximation, by evenly dividing the 100 PF’s across the 4 PF combinations. Following our example, DiceTool allocates the 100 PF’s like this:
~ 25 3-3’s
~ 25 1-1’S
~ 25 4-4’s
~ 25 6-6’s

DiceTool performs similar mappings for the Double Pitches (DP), the Single Pitches (SP), and so on. But for now let’s stick with just the Primary Face hits…

How does this mapping result in identical calculated Edge Values for different permutations? Take another P6 permutation as an example: 1/4-1/4. What are the Primary Face combinations for this Set?
~ Tops: 1-1
~ Facing: 4-4
~ Bottoms: 6-6
~ Backs: 3-3

Do you notice that although the order is changed, it is the same list of PF’s as we had with the 3/1-3/1 set? And, again, if we follow the FF mapping strategy of DiceTool over 100 PF results we get the following:
~ 25 1-1’S
~ 25 4-4’s
~ 25 6-6’s
~ 25 3-3’s

Notice the mapping produces the same results. This is because the two-dice PF combinations are the same and only the positions of the combinations are changing. DiceTool maps these permutations to identical results.

Will your results match DiceTools results directly? No! If you review your specific PF face results (on the TossStats Tab), it is highly unlikely that you will find an even distribution of your PF results over the four PF combinations.

In other words, if you looked at how many 4-4 combinations you really tossed, you will probably find that you tossed more then, or less then, the 25 4-4 combinations that the DT Model allocated in the above example. What this translates to is more or less eights and so the probability of tossing an eight is actually more or less than what DT models. If the probability of tossing a number changes, then the Edge results change. Edge is a calculation of probability compared to amount wagered vs winnings.

But, on the other hand, if you check how far different your actual count is from the averaged count, it is also highly unlikely that you will find a wild swing between the PF results either. This is why I am comfortable with the results that DiceTool produces.

I’ll say it again; DiceTool provides a model, an approximation based on your tracked foundation frequencies. Use it as an indication of what is possible and then test the theory with either practice/mock betting or through tracking in-casino play.

[heavy]

And THAT is why I didn't try to explain it. Thanks Maddog. Once again, "You the man!"

[wudged]

Hey Pointman, care to share what other tools you are using?

[heavy]

There are a couple of different roll tracking programs out there. SmartCraps is a program Scoblete and the GTC crowd tout. I've heard some good things about it but I have a personal issue with the guys who run that camp and would not purchase it solely because they would derive income from it. Yeah, I know. I'm just made that way.

Dice Coach's partner and webmaster, Pablo, developed an on-line roll tracking program called TossTrack. You can access it through Dice Coach's website at http://www.dicecoach.com/tosstrack/

One of our forum members developed his own windows-based tracking program through MS Access. Apologies, but I can't remember the name of the program or whether it is generally available for download anywhere.

Most folks I know who have tried the other programs say they prefer BoneTracker. It's certainly a powerful program. Best of all, Maddog gives it away for free.

[wudged]

You may want to check out WinCraps. You can dump your rolls into it, and program betting strategies into it (it's a proprietary thing developed exclusively by/for WinCraps, but if you're writing vbscript already you shouldn't have much problems other than getting accustomed to its syntax/intricacies.) There's a shareware version that includes pretty much all features except for hyperdrive, which lets you go through thousands of rolls with your betting strategy in a matter of seconds. The full program is only $20, so it's probably cheap relative to SmartCraps.

[acpa]

WinCraps is very good and there already many betting programs to go with it.

Noah

[heavy]

Well, Wudged and acpa have taken you exactly where I'd have gone. Wincraps is great for wargaming betting strategies. It's relatively simple to take your roll results from BoneTracker and import them into WinCraps. Writing scripts for betting strategies is a little more tricky, but most of the "traditional" betting strategies have already been scripted and are available for download on the WinCraps site. You'll want to Google "Cloud City Software" to find Steen's fine program.

[Maddog]

…What I like about BT so far, is it uses my rolls. It's weak in betting. The betting is ok but it can't take into account pressing etc (unless it can and I've missed it)…

That is a good observation Pointman, and no, you haven’t missed it, BT does not provide much by way of betting.

Although there have been contributions from our board members over the years such as MP/Wongs DiceTool and Allyoops Bet Summary Tool, the best tool around for doing bets war-gaming and experimentation that I have ever seen is Steen’s WinCraps (http://www.cloudcitysoftware.com).

~~~~~

BoneTracker’s primary purpose and focus is to provide tracking of the development of your dice tossing skill. It helps validate that you have developed a true influence over dice outcomes, or just a pretty toss with random or near random results.

We, as DI shooters, have to prove that our results truly deviate from random and we need to prove that the influence we impart is consistent and that the influence is consistent over time. Sometimes one of the most difficult things we do is take a critical look at our own dice results to determine if the results were truly influenced or the result of expected variance.

BoneTracker provides us a tool to track and measure so that we have a basis to answer this question.

Knowing that BT is focused on validating your influence, the first question should not be “how do I know what to bet?”, but should instead be “how do I know if my toss is working?”

BoneTracker is based on the On-Axis dice influencing model. It shows how your practice tosses perform under this model and if you have the skill to develop an advantage based on the on-axis theory of dice-sets. It will show you how your dice toss translate into dice results in terms of patterns (or lack of patterns), in terms of left vs right die performance to point out potential grip issues and to assist with overall toss improvement.

BoneTracker shows the numbers that you toss the most with a given dice set; this in turn helps you decide where to place your bets.

So a more concrete example: If over the course of 2 or 3 months of tracking your SL OH 3Finger-Front-Grip toss, using a standard permutation of the 3V set, you can see a consistent improvement of tossing the eight over random expectations, then you have a very good indication that you should be betting the eight. It should be that straight forward.

Starting from that base premise, you then expand out into looking at DiceTool to identify leverage, other dice sets to target numbers, Toss Stats for continuous improvement, etc.

BTW: you can read some of the background that lead to the development of BoneTracker here: Maddog's Journey

[Jonah]

[snip]

One of our forum members developed his own windows-based tracking program through MS Access. Apologies, but I can't remember the name of the program or whether it is generally available for download anywhere.

[snip]

Sharkbyte's Toss Analyzer :

For users without Microsoft Access:
http://download.cnet.com/Toss-Analyzer/ ... 01729.html

For users with Microsoft Access:
http://download.cnet.com/Toss-Analyzer- ... 01725.html

BTW continues to work with Microsoft Access 14.0.4760.1000 – this is part of Microsoft Office Professional Plus 2010.

[heavy]

THAT'S the one I couldn't remember. Thanks, Jonah.