Quote from: BusyBeaverHP on August 19, 2014, 06:47:14 AM

Quote from: dooglus on August 19, 2014, 05:54:59 AM

Quote from: BusyBeaverHP on August 19, 2014, 04:15:28 AM

Quote from: dooglus on August 19, 2014, 03:32:56 AM

Are you some kind of a crazy person?

I'm glad you asked.

Random.org is a True Random Number Generator that I use under certain conditions as a Hi/Low decider to ensure that I don't have any cyclical betting patterns that may turn into a weakness under exposure to a large number of rolls.

I actually SHA-256 hash my seeds and truncate a part of it to fit into the PD3 client seed. Like I said in another thread: OCD about entropy.

Notice the Excel sheet, I keep record of the periodicity of 7+ streak failures to investigate my strategy with a series of value derivations versus the mathematical expectations. The mathematical expected odds of running into a 7+ streak failure given 49.5% odds is about 1 in 119, my strategy in practice after 20,000 rolls is about 1 in 220 currently.

I have a similar 20,000 roll record of PD2 that also reflected this trend (at 1 in 238, exactly 2x). One day when I have met my set BTC quota, I may just share the details my findings with the world...

So that's a "yes"? Tongue

I recently turned 0.018 BTC into 1.444 BTC on prcdice using a stupid bunch of 49.5% and 45% martingales and a bunch of 90% all-ins. That's an increase of over 80 times. Then I turned it into 0 BTC using the same strategy.

I'm willing to share my findings with the world: It's just the way she goes; sometimes she goes, sometimes she doesn't go.

The same craziness that launched us into space and then invent magical internet money, oh yes... that kind of crazy Wink

I actually have a powerpoint spelling out all the ifs and thens of my betting procedure that I share with real-life friends.

With data recording the outcome of my betting procedure, I'm prodding at the low-tiered losing streaks and seeing how frequently they show up versus the mathematical expectation. Let's say that in 100,000 rolls, I'm supposed to find 800-900 7+ streak failures, and in my experiment, I only find 400-500.

If the number holds up to the current trends, that is a significant advantage because that smaller failure platform means fewer numbers of higher-tiered failures can be supported on top of that, and from there, I can recalculate my wager based on observed risks.

Now once I've collected enough data, I should be able to use something like the Student's t-test to determine how significantly different it is from the normal distribution of a null control dataset. It will be published free of charge after I have retired from this gambling thingy.

It's data-driven, and it'll be awesome...

You are an idiot. Go take some Coursera intro to stats course and stop spewing bullshit mixed with math 'buzzwords'.

broolstoryco

Posts

Merits