Combinations & % Error Challenge

         FILE :OHIO ROLLING CASH5 
   COUNT OCCURRENCES OF NUMBERS IN FILE
 
    10/04/04 TO 05/28/08  1193 RECORDS

  11. 176   31. 161   18. 151   37. 141   
  21. 175   36. 160    1. 150   17. 139   
  15. 172   23. 159   14. 150    6. 138   
   7. 171   33. 159   29. 149    2. 137   
  30. 170   24. 156   32. 149    9. 137   
   3. 169   13. 154   10. 148   28. 136   
  19. 166   27. 154   20. 148   39. 136     
  38. 165    4. 153   12. 147   25. 135   
  35. 164    8. 152   26. 146   22. 133   
  16. 162   34. 152    5. 145   

A=11 21 15 07 30 03 19 38 35 16 31 36 23
B=33 24 13 27 04 08 34 18 01 14 29 32 10
C=20 12 26 05 37 17 06 02 09 28 39 25 22

Tried your suggestion and there were 21 distribution patterns.  Nothing that I could use to pick combinations in a unique way.

   1. A  B  B  C  C =176
   2. A  A  B  B  C =171
   3. A  A  B  C  C =139
   4. A  B  C  C  C =118
   5. A  B  B  B  C = 99
   6. A  A  A  B  C = 98
   7. A  A  B  B  B = 52
   8. A  A  A  B  B = 48
   9. B  B  B  C  C = 44
  10. B  B  C  C  C = 44
  11. A  A  C  C  C = 43
  12. A  A  A  C  C = 33
  13. A  A  A  A  B = 24
  14. A  A  A  A  C = 24
  15. A  B  B  B  B = 22
  16. B  B  B  B  C = 17
  17. A  C  C  C  C = 16
  18. B  C  C  C  C = 15
  19. B  B  B  B  B =  5
  20. C  C  C  C  C =  3
  21. A  A  A  A  A =  2

Having separated numbers in winning combinations into groups of high/low/medium, active/normal/inactive, hot/normal/cold and other ways and seeing all the different mixes has less than 25 distribution patterns and are were dominated by nine or fewer patterns that looked similar regardless of the type of mix convinces me that it almost impossible to eliminate a small group of numbers that will not have any of the winning numbers most of the time.

In the Ohio Rolling Cash5 file which I sighed for this thread, 51% of the time at least one number from the previous drawing was in the winning combinations and 10% of time it was two or more.  With its payout of $1 for a match2, $10 for a match3 and $300 for a match4 had I simply replayed the previous winning combination every drawing, I would have won back $582 of the $1,192 it would have cost to play.

I am please to noted that you took the time to look.  While the results were disappointing, having that knowledge is better than not having it.

I once had a chemistry professor in college who told us that 99% of all the experiments turned out to be failures.  Had someone documented those failures, future chemists would not waste their time re-inventing the wheel, or the experiment in this case.

I think LP would benefit if there was a forum for "failed" attemts.  There might be a gem or two of benefit to someone else.

Hey, that would make a good poll.  "How many failures have you experienced while developing your method?"

Anybody up to do that one?

How would you define "failure"?

I have tried systems that had the winner but were too expensive to play. (haven't we all?) I have had systems that didn't work well with other systems (non-confirming). Usually, afterwards, I had created some code that I could use in other places.

LP has a 2if5 wheel with 25 combinations that uses all the numbers in a 5/39 game. Granted playing $25 with a win expectation of $1 doesn't look all that great, but it costs nothing to experiment with the wheel. The wheel is unbalanced; 1 number will appear 7 times, 6 numbers 4 time, 30 numbers 3 times, and 2 of the numbers will appear twice. It might be easier to break this wheel down and create a system that raises the win expectation than using a system that is more expensive to play.

Hey, that would make a good poll.  "How many failures have you experienced while developing your method?"....qoute:GASMETERGUY

If success is winning more than you spend every time you play then lottery players fail almost every time they play and that's the way the games are designed or else the states couldn't keep 50% of the sales for themselves and offer a large jackpot to the winner to attract more players.

When I test a theory, what ever happens is not a failure to me, the results answer a question I asked or disprove a myth that I read or heard.  Almost any system can win something if you're willing to buy enough combinations.  Beating the odds seem to be more a matter of luck than how one pick  combinations, at least that's my theory until I or someone else proves other wise.

I'll continue to play combinations picked using a system because even when I lose I like to think I'm in control of my own destiny.

If that's directed at my post, I didn't say anything about playing pick 3 with 5 even digits. I said playing pick 3 only with even numbers, of which there are 500. Play 1 of them (straight) and your chance of winning is 1 in 500 if the winning number is even. FWIW, if you played even numbers boxed you could also win on odd numbers that have an even digit, such as 176 paying for 167, 617, 671, and 761.

If I was going to play numbers that can be made with (all) 5 even digits I can play a lot more than the 65 you list:
1: 000
2: 002
3: 004
4: 006
5: 008
6: 020
7: 022
    024, 026, 028
8: 040
    042
9: 044
    046, 048
10: 060

You said: " For a simple analogy imagine playing pick 3 with only even numbers. If the winning number is even you have a 1 in 500 chance of winning instread of 1 in 1000. Since there is only a 50% chance that the winning number will be even your overall chance of winning is .50 * 1 in 500, which is 1 in 1000. There is no net gain."

Ok. Always amazing how we can see exactly what we want to see.

One interesting thing was I learned something about the Maybel
Quik Wheeler.  I entered Straight 0-2-4-6-8 and Force Doubles and
got 65 combinations for my trouble.

Went back and unchecked Force Doubles and got 125.  Got to stop
doing things when I'm half asleep. Zzzzzzzzz 

1: 000
2: 002
3: 004
4: 006
5: 008
6: 020
7: 022
8: 024
9: 026
10: 028
11: 040
12: 042
13: 044
14: 046
15: 048
16: 060
17: 062
18: 064
19: 066
20: 068
21: 080
22: 082
23: 084
24: 086
25: 088
26: 200
27: 202
28: 204
29: 206
30: 208
31: 220
32: 222
33: 224
34: 226
35: 228
36: 240
37: 242
38: 244
39: 246
40: 248
41: 260
42: 262
43: 264
44: 266
45: 268
46: 280
47: 282
48: 284
49: 286
50: 288
51: 400
52: 402
53: 404
54: 406
55: 408
56: 420
57: 422
58: 424
59: 426
60: 428
61: 440
62: 442
63: 444
64: 446
65: 448
66: 460
67: 462
68: 464
69: 466
70: 468
71: 480
72: 482
73: 484
74: 486
75: 488
76: 600
77: 602
78: 604
79: 606
80: 608
81: 620
82: 622
83: 624
84: 626
85: 628
86: 640
87: 642
88: 644
89: 646
90: 648
91: 660
92: 662
93: 664
94: 666
95: 668
96: 680
97: 682
98: 684
99: 686
100: 688
101: 800
102: 802
103: 804
104: 806
105: 808
106: 820
107: 822
108: 824
109: 826
110: 828
111: 840
112: 842
113: 844
114: 846
115: 848
116: 860
117: 862
118: 864
119: 866
120: 868
121: 880
122: 882
123: 884
124: 886
125: 888

BobP

Hi,

  By using an individual number method, for Pick 3, you get the following:

1-leaves 729 combinations -%error=27.1 (ie-eliminate the number 8 from all combinations being considered)

2-leaves 512 combinations-%error=48.8

3-leaves 343 combinations-%error=65.7

4-leaves 216 combinations-%error=78.4

5-leaves 125 combinations-%error=87.5

6-leaves 64 combinations-%error=93.6

7-leaves 27 combinations-%error=97.3

8-leaves 8 combinations-%error=99.2

9-leaves 1 combinations-%error=99.9

Even though this is a very effective method in a Pick 5 type of game, it ends up being a 10 to 1 reduction method in Pick 3 and pick 4 games. 

Does anyone have a Pick 3 or Pick 4 method that doesn't end up being a 10 to 1 reduction method? (10 combinations for every 1% error)

Failure is testing a theory only to discover the concept did not pick the winning numbers a sufficent number of times to make the financial effort profitable.

I agree most heartily with your statement about systems that were too expensive to play.  These systems, in my humble but sagacious opinion, can be labeled "failure".

So what is "success"?  Sucess is a system that is reasonable, that results in less than 100 number sets and is correct ( it predicts a winner) at least one time out of ten.  (And if anyone has such a system, I would be more than willing to invest in your project.)  I doubt such a system exists.  I have no doubt that such system will be developed by someone at some time in the future.  There is just too much research going on right now for such a system not to be found.  We may not have PhD's or Nobel prizes but we have something most people do not have....we have imagination. 

Feel free to tweak the above characteristics of "success".  Success, like failure, means different things to different people.  One man's failure is another man's success or something akin to that.