Using Bone Tracker for Smarter Betting?

[Wrangler]

I'm no math wizard. Thats why I really appreciate Bone Tracker, but I also get a little overwhelmed with all the data it provides.

I recently had someone ask me if I knew about the Kelly Criterion. Basically it comes down to understanding your edge and betting with it. So I started digging through Bone Tracker to see what it would tell me.

In the toss stats page, I can the possible outcomes 2-12, and if I were below or above expectations.
Even with an unremarkable SRR of 6.8, I was able to see that there is a number that seems to have a distinct advantage.
For me, it's the 5. Expected outcome is 11.1%, my outcome is 13.6% giving me a 2.5% advantage on the 5.

Forget the Kelly Criterion for a moment. If I'm reading this right, this is where I should be placing bets when I'm the shooter, right? At least as a starting point?

Can we take it one step further and look at the Box Number stats?
When I look at the Across stats, expected is 67%, but I'm at 71%. A 4% advantage.
Should I be using an across strategy?

Forgive me if this belongs in a different group, but I'm trying to become a smarter better and not just following the hot betting strategy of the week.

Thanks,
Dave

[Bankerdude80]

Wrangler, you are on the right track and your interpretation of the data is correct, however 478 rolls is still a very small sample. Once you have two books of 720 rolls for BoneTracker you will have a more accurate picture of where your true edge lies. The rule of thumb is to use data from 2 or 3 rolling books (720 rolls each book) of data of your most current books to use in any analysis whether developing a betting strategy or overall analysis. As DI's our tosses evolve over time and signature numbers can change.

In my experience, anything over 2.5% from random is a good marker in identifying your signature numbers. Percentages greater than 0 and less than 2.5% over random are more an indication of what your signature numbers could be, it's just that the current data hasn't fully borne them out.

I wouldn't make the jump to across betting just yet based on this small sample of data. Finish out the book, and try to get another book complete before making changes to your betting strategy.

[dork]

What B-dude said. Roll at least two full books if not three or four. Accumulating data for BT is much like "watching your portfolio"--you can do it on a day-to-day or week-to-week basis, but that short span isn't enough to give you a true picture of what's happening. Short spans might indicate a hand's trend, but a true signature number won't show itself in such a short time (or book). Ignore the data that BT calculates with each throw (or even a set less than 2 books); there's no "provable" trend in such a small sampling. When I go back 5-8 books, I wonder, "what happened to my toss?" because those "archived" results usually don't resemble my most recent books.

[220Inside]

I third all of the suggestions and advice from Bankerdude and dork. More data is needed, but your analysis is on the right track. Although across strategies do generally water down your strongest edges, as is evidenced by the below average results for 9's and 10's.

[Wrangler]

Thanks guys. I’ll keep it up and see where it leads.

[DarthNater]

Wrangler,
Definitely agree with all the above comments and let me toss out the following for when you have another book of rolls. Review the table above your screenshot and look at your on axis and one die off axis results. I always want to know my on axis frequent numbers as I would rather bet them, than bet across, keep charging, DN8R

[Parson]

Your on the right track. Still need more tosses to rule out varience etc.... those are not iron cross stats, but if they percentages hold up, i like the 5,6,8.

[coaster]

Wrangler,

I too agree with much of the above from the other guys, especially the need for more data.
Actually, with the info given and your interest in the across bet you might have a play for an early regression back to your top 3. Keep an eye on that prospect.

Parson, i can tell you've been putting some effort into learning more. 5,6,8 would be my top 3 also for these tosses. For quite a while 6,8 was my starting point followed by adding in my next best number after a few hits. That simple play worked pretty well for over a year playing basically like that. Way more positive trips than negative.

Wrangler, also keep in mind you have 23 other group sets that might contain one or more that give you a higher srr to work with. Also within the group you are using you have 23 others that might give you other numbers that better fit the way you might want to bet.

Keep up the good work. It gets easier to pick stuff out the more you work with it.

Coaster

[220Inside]

Wrangler,

I too agree with much of the above from the other guys, especially the need for more data.
Actually, with the info given and your interest in the across bet you might have a play for an early regression back to your top 3. Keep an eye on that prospect.

Parson, i can tell you've been putting some effort into learning more. 5,6,8 would be my top 3 also for these tosses. For quite a while 6,8 was my starting point followed by adding in my next best number after a few hits. That simple play worked pretty well for over a year playing basically like that. Way more positive trips than negative.

Wrangler, also keep in mind you have 23 other group sets that might contain one or more that give you a higher srr to work with. Also within the group you are using you have 23 others that might give you other numbers that better fit the way you might want to bet.

Keep up the good work. It gets easier to pick stuff out the more you work with it.

Coaster

Just a quick note that the SRR Group stuff that Coaster mentions is not available yet in BT. This new version of BT is still being worked on be Nate and me and won't be available for a while yet. It builds out on the ideas that Coaster developed and incorporates them into BT as another way of looking at your toss results. It's pretty powerful stuff.

[Moe Bettor]

It is amazing stuff even as it is. Can only speak of my experience in that even when I had a short hand with a couple of my numbers showing and I went seven out, my set came back in force after popping number after number. And it is a way to know that you are on your game too when they start coming early. For me it took a while to find the LZ on a table that had previously been a lot less bouncy.

[coaster]

Even though the set chart is not included in any currently available version of BT it is available on the forum. Nate posted one of his first compilations of my word files in the topic Face Matching for Easier Dice Set Comparison & Analysis on Sat Apr 04, 2020.

If you look at the chart, across the top row you will see the 24 group sets listed. Under them are the 24 sets that will always give the same 7 out for the same toss no matter which one you use. The cool thing about these 24 vertical sets is they give many different signature numbers depending on the set but always the same 7 out. In other words looking down column 1 if I used set 1122 and a seven was tossed it would have been a seven with every set in that column.

On Apr 20th I added a pretty detailed explanation to the topic describing how I find my best group set (my best SRR set) using the chart and a few other things. So if you are trying to find your own this could be some help with that.

Coaster

[DarthNater]

Right you are Coaster !

Two weeks ago on occasion, I changed sets 4 times in 17 tosses on a 40+ hand in an effort to hit an All. They were all seamless transitions done with no concern in that I kept popping secondary and primary box numbers in an effort to toss an elusive yo.

Not trying to brag, but rather trying to emphasize the importance of knowing those sets that not only give you the same facing matching characteristics, but also the same SRR for the greater than normal number distributions.

We’ve also discussed some of the leverage of Face Matching in the PUG sub forum, so Wrangler, give those a look when you have some time, DN8R