Hi all,
I happened to see that someone had posted statistics from a multimillion spins investigation on "another" board.
Because I like you people here so much, I post these tables so that you don't have to do the same kind of boring and time-consuming tests again, just because your amount of spins isn't 12M...
And NO; I didn't run some thousands of billions of spins to get to these figures: I used my calculator ;D
Very simple.
Please compare these figures to tests that you have seen or done!
(The "50/50" values are true 50% chance - like coin-tossing)
SINGLE ZERO TABLES
How many trials do we need, MATHEMATICALLY, to find a sequence of AT LEAST length X?
X 50/50 "Even" Cl/Dz DblStr
2: 3 3 8 37
3: 7 7 28 233
4: 15 16 89 1,445
5: 31 35 277 8,916
6: 63 74 858 54,991
7: 127 154 2,648 339,118
8: 255 317 8,167 2,091,238
9: 511 654 25,186 12,895,976
10: 1,023 1,345 77,660 79,525,192
11: 2,047 2,767 239,454 490,405,361
12: 4,095 5,689 738,320 3,024,166,401
13: 8,191 11,696 2,276,491 18,649,026,147
14: 16,383 24,043 7,019,184 115,002,327,911
15: 32,767 49,423 21,642,486 709,181,022,129
16: 65,535 101,592 66,731,001 4,373,282,969,801
17: 131,071 208,830 205,753,922
18: 262,143 429,264 634,407,931
19: 524,287 882,377 1,956,091,122
20: 1,048,575 1,813,778 6,031,280,963
21: 2,097,151 3,728,322 18,596,449,640
22: 4,194,303 7,663,775 57,339,053,061
23: 8,388,607 15,753,316 176,795,413,608
24: 16,777,215 32,381,819 545,119,191,962
25: 33,554,431 66,562,629 1,680,784,175,221
26: 67,108,863 136,823,183 5,182,417,873,600
27: 134,217,727 281,247,655
28: 268,435,455 578,120,181
29: 536,870,911 1,188,358,151
30: 1,073,741,823 2,442,736,201
How many trials do we need, MATHEMATICALLY, to find a sequence of AT LEAST length X?
X Corner SngStr Split Single
2: 84 151 341 1,368
3: 790 1,875 6,330 50,652
4: 7,319 23,136 117,134 1,874,160
5: 67,717 285,365 2,166,997 69,343,956
6: 626,397 3,519,513 40,089,474 2,565,726,408
7: 5,794,180 43,407,350 741,655,289 94,931,877,132
8: 53,596,182 535,357,330 13,720,622,865 3,512,479,453,920
9: 495,764,692 6,602,740,424 253,831,523,036
10: 4,585,823,413 81,433,798,580 4,695,883,176,188
11: 42,418,866,580 1,004,350,182,499
12: 392,374,515,880
13: 3,629,464,271,903
... and the opposite:
How long sequences can we expect, MATHEMATICALLY, in X trials?
X 50/50 "Even" Cl/Dz DblStr Corner SngStr Split Single
1,000 9 9 6 3 3 2 2 1
5,000 12 11 7 4 3 3 2 2
10K 13 12 8 5 4 3 3 2
50K 15 15 9 5 4 4 3 2
100K 16 15 10 6 5 4 3 3
500K 18 18 11 7 5 5 4 3
1M 19 19 12 7 6 5 4 3
5M 22 21 13 8 6 6 5 4
10M 23 22 14 8 7 6 5 4
12M 23 22 14 8 7 6 5 4
15M 23 22 14 9 7 6 5 4
20M 24 23 14 9 7 6 5 4
25M 24 23 15 9 7 6 5 4
50M 25 24 15 9 7 7 6 4
100M 26 25 16 10 8 7 6 5
1000M 29 28 18 11 9 8 7 5
DOUBLE ZERO TABLES
How many trials do we need, MATHEMATICALLY, to find a sequence of AT LEAST length X?
X 50/50 "Even" Cl/Dz DblStr
2: 3 3 9 39
3: 7 8 30 253
4: 15 18 99 1,607
5: 31 40 317 10,188
6: 63 87 1,007 64,533
7: 127 185 3,192 408,719
8: 255 393 10,110 2,588,562
9: 511 831 32,018 16,394,232
10: 1,023 1,757 101,395 103,830,144
11: 2,047 3,711 321,088 657,590,920
12: 4,095 7,835 1,016,781 4,164,742,499
13: 8,191 16,543 3,219,811 26,376,702,500
14: 16,383 34,925 10,196,071 167,052,449,175
15: 32,767 73,732 32,287,561 1,057,998,844,780
16: 65,535 155,659 102,243,946 6,700,659,350,280
17: 131,071 328,614 323,772,499
18: 262,143 693,743 1,025,279,584
19: 524,287 1,464,571 3,246,718,687
20: 1,048,575 3,091,873 10,281,275,844
21: 2,097,151 6,527,289 32,557,373,510
22: 4,194,303 13,779,833 103,098,349,450
23: 8,388,607 29,090,761 326,478,106,596
24: 16,777,215 61,413,830 1,033,847,337,557
25: 33,554,431 129,651,422 3,273,849,902,267
26: 67,108,863 273,708,558
27: 134,217,727 577,829,180
28: 268,435,455 1,219,861,604
29: 536,870,911 2,575,263,388
30: 1,073,741,823 5,436,667,154
How many trials do we need, MATHEMATICALLY, to find a sequence of AT LEAST length X?
X Corner SngStr Split Single
2: 89 159 360 1,443
3: 856 2,031 6,858 54,871
4: 8,144 25,741 130,320 2,085,135
5: 77,377 326,069 2,476,098 79,235,167
6: 735,090 4,130,227 47,045,880 3,010,936,383
7: 6,983,371 52,316,223 893,871,738 114,415,582,591
8: 66,342,042 662,672,173 16,983,563,040 4,347,792,138,495
9: 630,249,408 8,393,847,545 322,687,697,778
10: 5,987,369,391 106,322,068,924 6,131,066,257,800
11: 56,880,009,226 1,346,746,206,390
12: 540,360,087,661
13: 5,133,420,832,794
... and the opposite:
How long sequences can we expect, MATHEMATICALLY, in X trials?
X 50/50 "Even" Cl/Dz DblStr Corner SngStr Split Single
1,000 9 9 5 3 3 2 2 1
5,000 12 11 7 4 3 3 2 2
10K 13 12 7 4 4 3 3 2
50K 15 14 9 5 4 4 3 2
100K 16 15 9 6 5 4 3 3
500K 18 17 11 7 5 5 4 3
1M 19 18 11 7 6 5 4 3
5M 22 20 13 8 6 6 5 4
10M 23 21 13 8 7 6 5 4
12M 23 21 14 8 7 6 5 4
15M 23 22 14 8 7 6 5 4
20M 24 22 14 9 7 6 5 4
25M 24 22 14 9 7 6 5 4
50M 25 23 15 9 7 6 6 4
100M 26 24 15 9 8 7 6 5
1000M 29 27 17 11 9 8 7 5
If you've got a calculator... WHY spend time doing boring multimillion spins tests?
The end results are the same... Aren't they?
Regards,
KFS
Nice work KFS. :)
One thing though, do you think this would be better off in the reference section? It's bound to come in useful in the future, and if left here it will be buried...
if i'm looking for a sequence of 5....i look at greater than 4 ??
thanks
Hi Bliss, kompressor, and all,
@ Bliss:
Yes, you are right. I wasn't sure...
So I followed the instruction: "If unsure; post it here..."
@ kompressor:
Yes, you have to be looking for greater than four.
Five is greater than four, right? ;)
How shall I explain this?
I wish I was a teacher...
If you have 35 roulette-results there should (mathematically and in average) be at least one sequence there of length 5.
But we cannot say for certain that it is broken by the opposite or a zero at that exact length.
The sequence MAY be 6 or 7 long.
We can only say "for certain" (math certainty, that is ;)) that there will be at least five of the same.
(Mathematically and in average...)
The thing is, that if you look at enough 35-spins samples you will - sooner or later - find one sample that contains 35 of the same...
(I wrote sooner or later ;) - maybe you have to check a "few" billions of samples before you find it. But it will be there...)
Anywayzzz... I understand what you are saying...
So I have - as you may see - changed the headings and indices to what I hope will be more clear.
Best regards,
KFS
good lord kon-fu...
FULL TRIO (TERA TNT)
"...All Even-money bets balance out closely after +/- 100 spins. (The Law of averages)
The average is 55/45. Nine out of ten times the discrepancy between two even-money bets will be
less the 20 i.e. 60/40. YOU WOULD WIN NINE OUT OF TEN TIMES WITH BASICPLAY IF
THERE WERE NO TABLE LIMITS. BUT, the Casinos also need to be profitable. This is why
we developed AdvPlay1 and 2. These two systems combined with BasicPlay will keep your
stakes low ( under the table limit) and bring your count back to one as quick as possible ...."
FULLTRIO.pdf (nolinks://nolinks.freeadultstuff.us/gamble/302systems/FULLTRIO.pdf)
any statistics endorse that ??
Hi kompressor,
No.
;)
As I have said before; nothing will really "even out" other than in percent.
And the larger the sample the smaller the difference is ... in percent.
But a small percentage of a large sum is also a large number.
I mean that 5% of 100 is 5 while 0.05% of 100,000 is 50 - ten times more while the percentage was 1/100th.
And so: The larger the sample, the wider is the spread.
IN NUMBERS.
Or, in the down-to-earth term: IN UNITS.
And regarding 100-spins sequences that is talked about...
I ran 1,000,000 such sequences and counted how many "High" numbers (>18) there were in each.
This was for single-zero...
I stopped at some "levels" and output the sums that far.
And this is the result:
[table=,]
High, 10, 50, 100, 500, 1000, 5000, 10K, 50K, 100K, 500K, 1M
21, , , , , , , , , , , 1
22, , , , , , , , , , ,
23, , , , , , , , , , ,
24, , , , , , , , , , ,
25, , , , , , , , , , 1, 1
26, , , , , , , , , , ,
27, , , , , , , , , , , 5
28, , , , , , , , 1, 1, 4, 9
29, , , , , , , , 4, 6, 22, 36
30, , , , , , 1, 1, 4, 6, 33, 82
31, , , , , , 1, 3, 10, 16, 67, 129
32, , , , , , 3, 5, 20, 34, 146, 300
33, , , , , , 1, 6, 26, 59, 286, 594
34, , , , , 1, 2, 6, 46, 97, 512, 1038
35, , , , , 1, 9, 18, 92, 183, 944, 1860
36, , 1, 1, 1, 2, 12, 26, 144, 294, 1513, 3081
37, , , , 1, 5, 28, 56, 252, 516, 2614, 5270
38, , , 1, 2, 6, 48, 89, 422, 816, 4116, 8259
39, , 1, 1, 5, 11, 78, 137, 651, 1291, 6404, 12666
40, , , 3, 16, 23, 95, 177, 835, 1696, 8781, 17637
41, 1, 1, 5, 16, 33, 144, 256, 1267, 2525, 12636, 24978
42, , 1, 2, 12, 32, 188, 385, 1714, 3331, 16490, 33167
43, , 1, 2, 17, 35, 185, 407, 2090, 4167, 21005, 42058
44, , 2, 4, 21, 43, 244, 504, 2618, 5215, 25913, 52038
45, 1, 4, 7, 40, 77, 330, 618, 3030, 6105, 30657, 61085
46, 2, 8, 13, 38, 71, 326, 655, 3414, 6865, 34632, 69218
47, , 4, 6, 36, 67, 374, 746, 3845, 7580, 37919, 75505
48, 1, 3, 8, 44, 82, 407, 770, 3956, 8033, 39607, 78955
49, 1, 1, 6, 33, 77, 416, 785, 3902, 7889, 39892, 79800
50, , 4, 6, 35, 67, 365, 732, 3793, 7680, 38556, 76966
51, , 3, 5, 32, 60, 350, 711, 3507, 7078, 35357, 70939
52, 1, 3, 9, 40, 69, 312, 642, 3204, 6362, 31904, 63554
53, , 2, 3, 33, 68, 288, 589, 2855, 5586, 27585, 55099
54, , 3, 6, 21, 52, 233, 455, 2235, 4505, 22323, 44762
55, 1, 1, 3, 19, 38, 161, 355, 1765, 3478, 17555, 35581
56, , 1, 3, 11, 30, 131, 294, 1378, 2791, 13628, 27162
57, , 1, 1, 9, 18, 90, 189, 998, 1999, 10003, 20011
58, 2, 4, 4, 11, 20, 80, 154, 724, 1422, 6810, 13780
59, , 1, 1, 5, 6, 41, 90, 485, 956, 4719, 9519
60, , , , 1, 2, 28, 60, 286, 575, 3013, 6162
61, , , , , 1, 12, 32, 162, 363, 1860, 3765
62, , , , 1, 2, 8, 18, 117, 225, 1139, 2260
63, , , , , 1, 5, 17, 81, 122, 636, 1305
64, , , , , , 2, 7, 35, 69, 344, 660
65, , , , , , 1, 2, 16, 33, 204, 387
66, , , , , , , 1, 9, 15, 91, 170
67, , , , , , , 1, 4, 8, 45, 76
68, , , , , , 1, 1, 3, 4, 16, 40
69, , , , , , , , , 3, 11, 18
70, , , , , , , , , , 4, 7
71, , , , , , , , , 1, 2, 3
72, , , , , , , , , , 1, 1
73, , , , , , , , , , , 1
[/table]
As you can see, the more sequences we investigate (the larger the sample), the wider is the spread of Highs/hits - in NUMBERS.
Please note that there was for example as few as 36 Highs/hits in one sample of only 50, which shows that also low-probability events can happen very early in the chain of events.
Also in your personal chain...
(Take care ;))
Regards,
KFS
Quote from: Kon-Fu-Sed on January 19, 2009, 03:27:22 PM
I happened to see that someone had posted statistics from a multimillion spins investigation on "another" board.
Hi mate that "someone" (poit!) generously gave thumbs-up for us to host such analysis here:
nolinks://vlsroulette.com/reference-area/12-million-rng-spin-stats-shared-by-poit/ (nolinks://vlsroulette.com/reference-area/12-million-rng-spin-stats-shared-by-poit/)
:thumbsup:
Hi Victor,
I saw that and I think that's excellent as anyone now can compare my calculated figures with a true statistics investigation.
His stats include so much more than my figures, but what that data shows is that everything else also conform to math - like the SD, for example.
Check it out - there is no surprise at all in there.
This is a math game when you treat number-sequences as number-sequences and the wheel/dealer as an RNG.
That's why we can simply use a calculator instead of running billions of spins to collect stats.
Poit's stats and my calculations show that conclusively, IM(ns ;))HO.
KFS
Regarding to "even out in the long run"...
Suddenly I thought like this:
Suppose that during 3700 spins one "even" chance hits exactly as expected: 18/37.
That means it hit 1800 times.
Now, suppose we expect it to "even out" during the next 3700 spins.
So we expect it to hit a total of 3700 times during the total of 7400 spins - that's to even out, right?
Of those 3700 hits we already had 1800 in the first 3700 spins so we expect 1900 hits during the next 3700 spins.
But then the "even" chance has to hit 19/37...
Is there ANY reason why it should?
And it doesn't matter if you think it's going to even out during the next 37 or 37 billions spins: It still has to hit BETTER than 18/37.
Even if it hit AT THE EXPECTED 18/37 before.
Now, think of the necessary hit-rate to even out, if it had hit BELOW 18/37...
The truth is that an "even bet" (or ANY bet) will only "even out" to its NORMAL hit-rate of 18/37 (or whatever) in the long run.
And for the "even" bets, that missing "half/37" is an ever growing number in the long run:
0.5/37 of 3700 = 50 below half
0.5/37 of 37M = 500,000 below half...
To some, this seems to mean to "even out". In the long run.
Just my $0.02 - just a thought of how to explain it...
Never mind.
KFS
Hi Kon Fu Sed,
First of all thank you for all your hard work.
Now I'm trying to understand your stats.
What do these extracts from your tables below mean ?
Is it for instance that we need a maximum of 37 spins for a doublestreet to repeat twice?28 spins for a cl/dz to repeat twice? 1368 for a straight up to repeat? That a cl/dz will repeat 4 times within 89 spins?
And what are these long sequences of 10K and 50K? Is it that within 1000 there will be at least 6 repeats of a cl/dz?
Thanks
Purple
SINGLE ZERO TABLES
How many trials do we need, MATHEMATICALLY, to find a sequence of AT LEAST length X?
X 50/50 "Even" Cl/Dz DblStr
2: 3 3 8 37
3: 7 7 28 233
4: 15 16 89 1,445
5: 31 35 277 8,916
and
How many trials do we need, MATHEMATICALLY, to find a sequence of AT LEAST length X?
X Corner SngStr Split Single
2: 84 151 341 1,368
3: 790 1,875 6,330 50,652
4: 7,319 23,136 117,134 1,874,160
... and the opposite:
How long sequences can we expect, MATHEMATICALLY, in X trials?
X 50/50 "Even" Cl/Dz DblStr Corner SngStr Split Single
1,000 9 9 6 3 3 2 2 1
5,000 12 11 7 4 3 3 2 2
10K 13 12 8 5 4 3 3 2
50K 15 15 9 5 4 4 3 2
Hi purple,
Yes.
;)
Or maybe not. :D
Quote
we need a maximum of 37 spins
Not really maximum...
Mathematically that's what is needed to find a repeating 6/37 bet.
Quote
Is it that within 1000 there will be at least 6 repeats of a cl/dz?
In 1,000 trials you can expect (mathematically) to find at least one repeat-sequence of a column/dozen (a 12/37 bet) that is at least six long.
You can compare my results (the
12M line above) to the stats posted here:
nolinks://vlsroulette.com/reference-area/12-million-rng-spin-stats-shared-by-poit/ (nolinks://vlsroulette.com/reference-area/12-million-rng-spin-stats-shared-by-poit/)
Stats results:
Evens repeats: 22, 22, 22, 21, 21, 23. My calculator says [highlight]22[/highlight]
Dozens/Columns repeats: 18, 14, 14, 14, 13, 16. My calculator says [highlight]14[/highlight]
Double Street repeats: 8, 8, 8, 9, 8, 9, 8, 9, 8, 10, 9. My calculator says [highlight]8[/highlight]
Corner repeats: 7, 7, 7, 7, 6, 7, 7, 7, 7, 7, 6, 7, 7, 8, 7, 7, 7, 6, 7, 7, 8, 7. My calculator says [highlight]7[/highlight]
Single Street repeats: 7, 6, 6, 7, 6, 6, 7, 6, 6, 6, 6, 6. My calculator says [highlight]6[/highlight]
Splits repeats: 1 repeats 4 times, 39 repeats 5 times and 20 repeats 6 times. My calculator says [highlight]5[/highlight]
Single numbers: 29 repeats 4 times and 8 repeats 5 times. My calculator says [highlight]4[/highlight]
IMHO, a calculator is useful sometimes.
(One doesn't need to work so hard to get the figures ;))
KFS
Hi I'm still trying to get it.
Stats results:
Evens repeats: 22, 22, 22, 21, 21, 23. My calculator says 22
Are you saying that in 1000 spins there will be about 22 alterations for every even chance (BR, HL, OE), ie 132 alterations in total?) I don't think you're saying that there will be one instance of an even chance repeating 22 times are you?
Dozens/Columns repeats: 18, 14, 14, 14, 13, 16. My calculator says 14
In 1000 spins On average each column repeated itself at least 13 times, or one instance when a col/doz repeated itself 13 times in a row?Double Street repeats: 8, 8, 8, 9, 8, 9, 8, 9, 8, 10, 9. My calculator says 8
In 1000 spins each street repeated itself at least 8 times, or was there one instance when a street repeated itself 8 times in a row?[/i]
etc for
Corner repeats: 7, 7, 7, 7, 6, 7, 7, 7, 7, 7, 6, 7, 7, 8, 7, 7, 7, 6, 7, 7, 8, 7. My calculator says 7
Single Street repeats: 7, 6, 6, 7, 6, 6, 7, 6, 6, 6, 6, 6. My calculator says 6
Splits repeats: 1 repeats 4 times, 39 repeats 5 times and 20 repeats 6 times.
My calculator says 5 Please explain this
Single numbers: 29 repeats 4 times and 8 repeats 5 times. My calculator says 4
Please explain this
Hi purple,
The repeat-values I showed - for example:
Quote
[highlight]Evens repeats: 22, 22, 22, 21, 21, 23.[/highlight] My calculator says 22
were copied from the
12-million spins statistics that I gave you the link to: .
nolinks://vlsroulette.com/reference-area/12-million-rng-spin-stats-shared-by-poit/ (nolinks://vlsroulette.com/reference-area/12-million-rng-spin-stats-shared-by-poit/)
So they were found in
12 million spins.
The six values are for Low, High, Red, Black, Odd and Even, respectively.
I copied the repeat-values for col/doz, dbl streets, corners, single streets, splits and singles from the same 12 million spins stats.
My calculator values (same example):
Quote
Evens repeats: 22, 22, 22, 21, 21, 23. [highlight]My calculator says 22[/highlight]
were taken from the table "How long sequences can we expect, MATHEMATICALLY, in X trials?"
This is the X =
12M (12,000,000) line in that table (single zero):
X 50/50 "Even" Cl/Dz DblStr Corner SngStr Split Single
[highlight]12M[/highlight] 23 [highlight]22[/highlight] 14 8 7 6 5 4
Showing that in 12 million spins, a 22-repeater of an "even" (18/37) bet will be found (mathematically).
And the rest of the "calculator values" are also found at that line.
I hope this clears things up for you.
KFS
HI Kon-Fu-Sed
Now that we know that all chances even out , we should take advantage of this , shouldnt we?
here is a example of how i think we could take advantage of the even out rule...
so lets say a dozen didnt showed for at least 13 spins and when that dozen shows after the 13th spin we should bet on it , because its rather unlikely that that dozen will go now back to sleep for another long trail because of the even out rule.. isnt it?
or how often do you see such an extreme event exmpl on doz .: 12122211121221123212122111222112
waht do you think?
cheerz