Significance

[formula_2002]

Significance
You can test your results by using statistical formulas, calculating the standard deviation, calculating the Z score, and then look up the z-score value to determine the probability for achieving your results.

Quirin and Mitchell go into the calculation somewhat as does of course most books on statistics.

Or, perhaps you can also try the following as I have, using a data base (I use dbase) and a spread sheet (I use excel).

I wanted to know the probability of achieving z- scores >=1.9, when the win rate is 21% (approximate win rate for odds>=2.5 to 1 and <=3.4 to 1) given that the number of horses with those odds =100.

First, I generated a 10000 records data base of random numbers 1 to 100.

Then I wrote a program that appended another data base with the total the number of times the record value fell between 1 and 21 and grouped the results into sets of 100.

Second, I imported the appended data base into excel and calculated the standard deviation and the z scores.

Then I simply summed the number of times the z-score exceeded 1.9.

I did this for both the random number data base and my horses racing data base.

The results were;
Random number data base, 1.9 was exceeded 3 times (4%)
Horse racing data base, 1.9 was exceeded 5 times (5%).

So without any handicapping, it was possible to produce statistically significant results 5% of the time. (what?)


The following is the random number excel file.
The table compares the z scores with the Quirin formula results.

NW
TEST
SET_100 WINS DBT Z SCORES
100 12 1 -2.55 -2.21
100 14 1 -1.98 -1.72
100 15 1 -1.70 -1.48
100 15 1 -1.70 -1.48
100 15 1 -1.70 -1.48
100 15 1 -1.70 -1.48
100 15 1 -1.70 -1.48
100 16 1 -1.42 -1.23
100 16 1 -1.42 -1.23
100 16 1 -1.42 -1.23
100 17 1 -1.13 -0.98
100 17 1 -1.13 -0.98
100 17 1 -1.13 -0.98
100 17 1 -1.13 -0.98
100 17 1 -1.13 -0.98
100 17 1 -1.13 -0.98
100 17 1 -1.13 -0.98
100 18 1 -0.85 -0.74
100 18 1 -0.85 -0.74
100 18 1 -0.85 -0.74
100 18 1 -0.85 -0.74
100 18 1 -0.85 -0.74
100 18 1 -0.85 -0.74
100 19 1 -0.57 -0.49
100 19 1 -0.57 -0.49
100 19 1 -0.57 -0.49
100 19 1 -0.57 -0.49
100 19 1 -0.57 -0.49
100 19 1 -0.57 -0.49
100 19 1 -0.57 -0.49
100 19 1 -0.57 -0.49
100 19 1 -0.57 -0.49
100 19 1 -0.57 -0.49
100 19 1 -0.57 -0.49
100 19 1 -0.57 -0.49
100 19 1 -0.57 -0.49
100 20 1 -0.29 -0.25
100 20 1 -0.29 -0.25
100 20 1 -0.29 -0.25
100 20 1 -0.29 -0.25
100 20 1 -0.29 -0.25
100 20 1 -0.29 -0.25
100 20 1 -0.29 -0.25
100 20 1 -0.29 -0.25
100 21 1 0.00 0.00
100 21 1 0.00 0.00
100 21 1 0.00 0.00
100 21 1 0.00 0.00
100 21 1 0.00 0.00
100 21 1 0.00 0.00
100 21 1 0.00 0.00
100 21 1 0.00 0.00
100 21 1 0.00 0.00
100 21 1 0.00 0.00
100 22 1 0.28 0.24
100 22 1 0.28 0.24
100 22 1 0.28 0.24
100 22 1 0.28 0.24
100 22 1 0.28 0.24
100 22 1 0.28 0.24
100 22 1 0.28 0.24
100 22 1 0.28 0.24
100 22 1 0.28 0.24
100 22 1 0.28 0.24
100 22 1 0.28 0.24
100 22 1 0.28 0.24
100 23 1 0.56 0.49
100 23 1 0.56 0.49
100 23 1 0.56 0.49
100 23 1 0.56 0.49
100 23 1 0.56 0.49
100 23 1 0.56 0.49
100 23 1 0.56 0.49
100 23 1 0.56 0.49
100 23 1 0.56 0.49
100 23 1 0.56 0.49
100 23 1 0.56 0.49
100 24 1 0.84 0.73
100 24 1 0.84 0.73
100 24 1 0.84 0.73
100 24 1 0.84 0.73
100 24 1 0.84 0.73
100 24 1 0.84 0.73
100 24 1 0.84 0.73
100 25 1 1.13 0.98
100 25 1 1.13 0.98
100 25 1 1.13 0.98
100 25 1 1.13 0.98
100 25 1 1.13 0.98
100 25 1 1.13 0.98
100 25 1 1.13 0.98
100 26 1 1.41 1.22
100 26 1 1.41 1.22
100 27 1 1.69 1.47
100 27 1 1.69 1.47
100 27 1 1.69 1.47
100 28 1 1.97 1.72
100 28 1 1.97 1.72
100 29 1 2.26 1.96
100 30 1 2.54 2.21
AVERAGE 21.01
STDEV 3.543

[formula_2002]

Oh the possibilities!!!

A profit was returned in 90 of 648 sets of 100 races, where the odds were >=2.5 and <10.5.

That a 13% possibility of showing a profit with no handicapping.

Joe M

[Tom]

Originally posted by formula_2002

That a 13% possibility of showing a profit with no handicapping.

Joe M

You get no profit without betting. Run all the simulations you want, if you don't have real money on the selections, all you have is fun with numbers.