Rick G's Variance System Put To The (Back) Test

Reply #7

jimmy4164 — Mon, 06 Sep 2010 18:02:40 GMT

Rick,

As I did in Fooled By Randomness with simple betting schemes like betting on Yesterday's Winners, I left everything identical in this run except: I looked back three(3) days instead of one(1) to get the Pick3 and Pick4 numbers to calculate the Key Number. For me, this is enough to pass final judgement.

Backtest of Variance System looking back 3 draws for the Key Number predictor.

Here's what happened in PA between 1988/01/04 and 2010/09/04.

(8201 Draws - 27 Bets per Draw - Stra... [ More ]

Reply #6

Rick G — Mon, 06 Sep 2010 12:30:21 GMT

Jimmy, we're on the same page. Anything less than or equal to .50 was rounded down. .51 - .99 was rounded up. 10.4898 would be rounded off to 10.

Reply #5

Scott311 — Mon, 06 Sep 2010 04:37:44 GMT

You two should get a room

Reply #4

jimmy4164 — Sun, 05 Sep 2010 22:56:22 GMT

Rick,

This is to see if we're on the same page with our rounding methods. I got a PM questioning my calculation of your Key Number for 2010/08/09. Based on Pick3: 8-7-1 and Pick4: 8-4-8-2.

I calculated a Key Number of 0-0-0

The PMer got 0-0-1 by rounding a 10.4898 Floating Point Variance to and Integer 11.

My MS Basic statement:

IVAR = INT( VAR + 0.50 )

results in an Integer 10 when VAR==10.4898.

How are you rounding?

-------------------

This is probably an appropriat... [ More ]

Reply #3

Rick G — Sun, 05 Sep 2010 18:21:00 GMT

Jimmy, I consider it a legit test. Box results would not improve the bottom line. A good system must do demonstrably better than chance with straights.

I might have another one for you to check out soon...lol. The only thing is that it does involve some player judgement that may be tricky to include in a test. Will work on that angle.

Thanks again for your help.

Reply #2

jimmy4164 — Sun, 05 Sep 2010 17:42:01 GMT

Thanks for the encouragement Rick.

It appears that it doesn't perform better than chance, am I right?

It would appear that way. Although I don't expect to learn otherwise, I'll wait a while before passing final judgement.

Reply #1

Rick G — Sun, 05 Sep 2010 17:04:50 GMT

Jimmy, thanks much for your hard work. It appears that it doesn't perform better than chance, am I right? If that's the case then your effort has not been wasted. That's OK. I'm working on something else that might be better.

That's the beauty of throwing ideas out here at LP...feedback.

Thanks again

Rick G's Variance System Put To The (Back) Test

jimmy4164 — Sun, 05 Sep 2010 16:50:51 GMT

Rick G proposed a novel system for predicting Pick3s using the previous day's draw of the Pick3 AND its corresponding Pick4 He uses the Population Variance of the concatanation of the 7 digits from these two previous draws to produce his prediction for the next Pick3 draw in the form of a Key Number. It is recommended that, within your budget, you choose from the 27 (3 3 3) numbers generated by varying each of the digits of the Key Number by 1.

Here is his Post describing the system. This... [ More ]