A GUT-situation tested

[Kon-Fu-Sed]

Hi All,

winkel wrote in a reply to a request I made:

"because if 15-15 appears by this row:

20-10
19-11
18-12
17-13
16-14
15-15

I dont bet

if it appears
15-20
15-19
15-18
15-17
15-16
15-15

I bet. and that makes the difference to the statistical expectation of hitting one of the 15 numbers I bet."

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

So, of course I couldn't resist...
I investigated this in a ONE HUNDRED MILLION SESSIONS study...


Why 100,000,000 sessions? Are you nuts?
No one will EVER play one hundred million sessions!

Well...
I wouldn't trust a 250-bets test - would you?

I want tests to be as FAIR and UN-BIASED as possible. (Don't we all?)
And the only way to be FAIR and UN-BIASED is to run AS MANY SESSIONS AS POSSIBLE. In this case 100M is - in REALITY - nothing, as a session may be up to 50 spins long.

Do you want to have at least a small probability that at least one session starts with "32-18-25-2-13-29"?
Or maybe with "6-0-15-30-8-11"?
Or with "27-31-19-10-5-32"?
Or YOU select any 6-numbers sequence.

You should know that even 100 MILLION sessions will far from secure that ANY of those are used to start a session!

FAR from it:
As six numbers can be combined in 37^6 different ways, the probability for each one of the above examples is 1 / 2,565,726,409.
Therefore, the probability for each one of those three is LESS THAN 1/25 to occur as the start of one session of 100 millions!

Less than 1/25 for only six numbers...
In 100,000,000 sessions.

Just 10 results make 4,808,584,372,417,849 combinations - that's like 182,975,052,222 YEARS spinning every 2nd minute 24/365...
To find ONE unique 10-numbers combination...

And 50 results that we are dealing with here give a sum that is 78 digits long.

That's a lot of sessions...

Do you understand why a test of only some 250 bets, in reality is useless? (See the PS)
Maybe the most common session-results (hits/misses) are under-represented in such a ridiculously small sample?
Or maybe the most obscure session-results are over-represented?
Do you want that? Do you KNOW that that is not the case?

As a tester you must at least TRY to secure that such things don't happen.
You have to TRY to make a FAIR and UN-BIASED test.


Admittedly; using only 100M sessions is a compromise - the below two tests was a ten-hours run...

But I hope the 100M sessions are enough to show the trend anyway...
250 bets for sure does not.

/KFS

[Kon-Fu-Sed]

There is also ANOTHER reason to make 100M sessions:

I asked winkel for a RATHER FREQUENT situation.


@winkel:
As a reply, you show a situation that is found in 32,157 sessions out of 100,000,000... 0.032157 % of the sessions...
Not even 1/30 of one %...

Or in other words: It occurs approximately once in 3,109 sessions.

Are you serious? Is this a RATHER FREQUENT situation?

*** How many times have YOU seen this situation in live play, as you claim that you bet when you see it?
*** How many sessions have you really played?
*** How many times did this situation occur in your 10 million spins test?

Let me guess that last one: 50 spins per session = 200,000 session gives an average of approximately SIXTY-FOUR TIMES!


How did you come to the conclusion that this situation is a favourable one as you wrote:
"and that makes the difference to the statistical expectation of hitting one of the 15 numbers I bet."

By the results of a 64-bets test???


Anyway, such a low frequency made it use-less to test on my Wiesbaden spins as I don't even have 2,000 days.
Even if we ran like 6 sessions each day, we would only find something like FOUR of the suggested situations. In total. Maybe not even that...
And that is data for nearly SIX YEARS!
(The same for my Hamburg spins: I haven't got enough of them either.)

Four times in six years...
And this is claimed to be rather frequent...

I have to say I'm really disappointed. 


SO the tests were run exclusively on RNG numbers.
But as you claim that the method beats such numbers also, it should be OK. Right?

Are you sure there isn't ANY situation that is at least a BIT more frequent than 1 / 3,000 sessions in average?
Kon-Fu-Sed

[Kon-Fu-Sed]

THE FIRST TEST:

Testing Col "=0" vs Col "=1"
20-10 going to 15-15

Quoting winkel:
"because if 15-15 appears by this row:
20-10
19-11
18-12
17-13
16-14
15-15
I dont bet"

RESULTS:

Started sessions:               100,000,000
"=0" col gone <20 = Exit:        84,045,871 (and/or "=1" col not = 10, before/at spin 50)
No cond met in 50 spins = Exit:           4 (f ex col "=0" not reached 20)

Reached 20-10 bef/at spin 50:    15,954,125 (May continue after 50 spins if necessary)

20-10 Spin Became 20- 9 = Exit:   4,310,171
           Still  20-10 = Exit:   3,021,875
           Became 19-11 = Cont:   8,622,079

19-11 Spin Became 19-10 = Exit:   2,564,502
           Still  19-11 = Exit:   1,629,572
           Became 18-12 = Cont:   4,428,005

18-12 Spin Became 18-11 = Exit:   1,435,185
           Still  18-12 = Exit:     836,454
           Became 17-13 = Cont:   2,156,366

17-13 Spin Became 17-12 = Exit:     756,529
           Still  17-13 = Exit:     408,521
           Became 16-14 = Cont:     991,316

16-14 Spin Became 16-13 = Exit:     374,582
           Still  16-14 = Exit:     188,426
           Became 15-15 = Cont:     428,308

15-15  TOTAL to start bet:          428,308

15-15 HIT at the 1st bet = Exit:    173,881 ==> Math: From 172,675 to 174,602 hits (Ave: 173,638.378 hits, Result = +0.140% from ave)
15-15 Lost the 1st bet:             254,427

Still 15-15 after the loss:          81,234
15-15 HIT at the 2nd bet = Exit:     33,107 ==> Math: From 32,513 to 33,352 hits (Ave: 32,932.702 hits, Result = +0.529% from ave)
15-15 Lost the 2nd bet = Exit:       48,127

15-15 Bec 15-14 after the loss:     173,193
15-14 HIT (the 2nd bet) = Exit:      69,787 ==> Math: From 69,601 to 70,826 hits (Ave: 70,213.378 hits, Result = -0.607% from ave)
15-14 Lost (the 2nd bet) = Exit:    103,406

Sum of ALL 15-numbers bets:         682,735
Sum of ALL hits:                    276,775 ==> Math: From 275,568 to 278,001 hits (Ave: 276,784.459 hits, Result = -0.003% from ave)

TOTAL bet units (15 x 682,735):  10,241,025
TOTAL won units (36 x 276,775):   9,963,900
ACTUAL LOST units:                  277,125
ACTUAL LOST units in % of bet u:    -2.706% <=== Quite close to the math predicted -2.703%, don't you think?


("Ave" = "Average number of hits" = 15 / 37 x bets)
(The range of "Math" hits is the "Ave" value minus and plus NOT breaking the +-3SD barrier)

[Kon-Fu-Sed]

Note:
For this second test we used exactly the same RNG spin-sequence as for the first test.

THE SECOND TEST:

Testing Col "=0" vs Col "=1"
15-20 going to 15-15

Quoting winkel:
"if it appears
15-20
15-19
15-18
15-17
15-16
15-15
I bet. and that makes the difference to the statistical expectation of hitting one of the 15 numbers I bet."

RESULTS:

Started sessions:               100,000,000
"=0" col gone <15 = Exit:        98,379,821 (and/or "=1" col not = 20, before/at spin 50)
No cond met in 50 spins = Exit:     433,082 (f ex col "=0" not reached 15)

Reached 15-20 bef/at spin 50:     1,187,097 (May continue after 50 spins if necessary)

15-20 Spin Became 14-21 = Exit:     483,356
           Still  15-20 = Exit:      63,375
           Became 15-19 = Cont:     640,366

15-19 Spin Became 14-19 = Exit:     259,703
           Still  15-19 = Exit:      52,530
           Became 15-18 = Cont:     328,133

15-18 Spin Became 14-18 = Exit:     133,187
           Still  15-18 = Exit:      35,111
           Became 15-17 = Cont:     159,835

15-17 Spin Became 14-17 = Exit:      64,468
           Still  15-17 = Exit:      22,033
           Became 15-16 = Cont:      73,334

15-16 Spin Became 14-16 = Exit:      29,469
           Still  15-16 = Exit:      11,708
           Became 15-15 = Cont:      32,157

15-15  TOTAL to start bet:           32,157

15-15 HIT at the 1st bet = Exit:     12,935 ==> Math: From 12,773 to 13,300 hits (Ave: 13,036.621 hits, Result = -0.780% from ave)
15-15 Lost the 1st bet:              19,222

Still 15-15 after loss = Bet:         6,237
15-15 HIT at the 2nd bet = Exit:      2,539 ==> Math: From 2,413 to 2,644 hits (Ave: 2,528.514 hits, Result = -0.415% from ave)
15-15 Lost the 2nd bet = Exit:        3,698

15-15 Bec 15-14 aft loss = Bet:      12,985
15-14 HIT (the 2nd bet) = Exit:       5,215 ==> Math: From 5,097 to 5,432 hits (Ave: 5,264.189 hits, Result = -0.934% from ave)
15-14 Lost (the 2nd bet) = Exit:      7,770

Sum of ALL 15-numbers bets:          51,379
Sum of ALL hits:                     20,689 ==> Math: From 20,496 to 21,163 hits (Ave: 20,829.324 hits, Result = -0.674% from ave)

TOTAL bet units (15 x 51,379):      770,685
TOTAL won units (36 x 20,689):      744,804
ACTUAL LOST units:                   25,881
ACTUAL LOST units in % of bet u:    -3.358% <=== Quite close to the math predicted -2.703%, don't you think?


("Ave" = "Average number of hits" = 15 / 37 x bets)
(The range of "Math" hits is the "Ave" value minus and plus NOT breaking the +-3SD barrier)

NOTE:
As the number of bets in this test is considerably lower compared to the first test (less than 10%), the results in this test are not as statistically secured.
To reach the same level of security, at least 1,000,000,000 sessions should be done.
That is, however, more than 60 hours computer-time...
(And the results would, with a high probability, end quite similar to the first test)

[Kon-Fu-Sed]

winkel wrote:
"- there is no use to test 1 single combination for crossings"


Now, as seen above, the situation winkel suggested does not, ON ITS OWN, give a better probability to hit than math says.
As winkel knows, I guess based on what he wrote above.

Then there HAS to be ANOTHER situation that gives a LOT higher probability! But then; why didn't winkel show us THAT ONE?
The one that will, in co-operation with the suggested one, give a better result when the suggested situation loses...
And maybe vice versa...


It simply cannot be another one that ON ITS OWN gives a NEGATIVE result!
That is just not possible.
There HAS to be one or more that are POSITIVE to balance the NEGATIVE one.
(That's really not hard to understand, is it?)


Let me use an EXAMPLE to illustrate this - a game that is not roulette (because roulette is more complex) but is not far from:

But first:
Suppose we have two betting opportunities: A and B.
Suppose that when opportunity A loses, the B opportunity is GUARANTEED to come.
(This is way more than winkel can guarantee)

Now...
Suppose that for A you bet 9u and you get 19u back on a win. So A gives +10u on a win and -9u on a loss.
Suppose A wins 4 trials out of 9, in average, in a fair test.
So A is a losing opportunity in average: +10x4 -9x5 = +40 -45 = -5. For every 9 trials.
Expressed in % of total units bet, it gives an average loss of 6.17%.

Suppose that for B you bet 8u and you get 20u back on a win. So B gives  +12u on a win and -8u on a loss.
Suppose B wins 4 trials out of 9, in average, in a fair test.
(NOTE: In reality, as we are betting less numbers compared to A, we would have a lesser chance than A to win but this is not reality anyway, so...)
So B is a GUARANTEED WINNING opportunity in average: +12x4 -8x5 = +48 -40 = +8. For every 9 trials.
Expressed in % of total units bet, it gives an average profit of 10%.

Way above the A loss.


So!
We have a guaranteed (in the long run) loser of 6.17% of bet units.
And we have a guaranteed (in the long run) WINNER of 10% of bet units.
And the guaranteed winner is also GUARANTEED to kick in when the A opportunity is lost.


Can the +10% opportunity balance the -6.17% in the long run?


As both A and B have 9 as a common denominator the possible combinations are 9 x 9 = 81.

A wins in 4/9 and we stop. This is 36 cases of 81. (4/9 x 9/9 = 36/81)
As the profit for each case is 10u, the total profit for these 36 cases is 360u.

A also loses in 5/9 and B kicks in and wins in 4/9 giving 5 x 4 = 20 cases out of the 81. (5/9 x 4/9 = 20/81)
A loses 9 units but B wins 12 so the net profit is 3u.
As the profit for each case is 3u, the total profit for these 20 cases is 60u.

A also loses in 5/9 and B kicks in but loses 5/9 giving 5 x 5 = 25 cases out of the 81. (5/9 x 5/9 = 25/81)
A loses 9 units and B loses 8 so the total loss is 17u.
As the loss for each case is 17, the total loss for these 25 cases is 425u.

[CASES REALITY CHECK: 36 + 20 + 25 cases = 81 cases = correct]


TOTAL: +360 +60 -425 = -5.

A LOSS of five units... And the B-bet was GUARANTEED to come when A lost and was GUARANTEED to win in the long run.!

winkel can NEVER show ANY betting opportunity that will be GUARANTEED to come when another loses AND is GUARANTEED to give a positive result.
Can you, winkel?
If not, how can you claim that the betting opportunities will balance to positive?


The very simple truth is:
You will need ONE OR MORE GUARANTEED HEAVY POSITIVE opportunities to balance a negative one.


If positive results have a hard time doing it: How can NEGATIVE results possibly balance each-others LOSSES to become POSITIVE?


Proof beyond any doubt: Have a look at my and my wife's bank-accounts!



If it WAS possible: Why didn't winkel show what to bet if the tested situation should fail?
As it did.
As he KNEW it would ("- there is no use to test...").


If you bet 15 numbers - how can they possibly hit better than a 15-numbers bet?
Kon-Fu-Sed

[Kon-Fu-Sed]

My conclusions, taking all the above into consideration:

I don't know how winkel came to the conclusion that this method can beat 10 million RNG spins.
The probability for that to happen is so small, he SHOULD have done another test to verify the first result.
If it's still the same result he SHOULD look for programming errors or a faulty RNG.
And the matter would have been solved in a snap, probably.

The G.U.T. method is, in reality, a variation of the usual "Gambler's Fallacy" methods.
(And an un-necessary complicated one, IMO...)

The bottom line for all such methods is: "Wait for something to happen because then something else is due to happen."

Heard it before?


"Gambler's Fallacy" methods has been tried for centuries to be proven correct but unfortunately no one has ever come up with the math evidence or a FAIR replicable test.
Not even winkel.

Only the opposite has been shown.
For centuries...


Too bad, really.
KFS





PS:
WHILE the above test was running, I studied the live-spins tests winkel has performed using Wiesbaden spins, in order to learn the rules for this "clinical" way of finding betting opportunities.
I wanted to make such "clinical" tests myself but on way larger samples.

When I saw the above results I found no incentive to continue...

But as I had the data already, I checked how many bets there were:
Some 250 bets...

That is not a TEST to conclude anything from.

The day I bet money using a method that's been tested in public for only 250 bets, well, that day will never come, I hope.

Now, I have used enough space, time and resources on this...
And I have shown my conclusions.

No more will be said by me.
KFS

[winkel]

Hi KFS,

nicely done. And you´ve produced exactly what I´ve predicted.

The simple remaining question: "Why does it win?"

If you would be so kind to do the following:

take any crossing you like doesn´t matter which one.
produce a grph/diagram

What you will see there will be long streaks of the lines going up/going down/going sideways

So now you will see that this kind of graph will be produced by every other possible crossing-situation.

let´s go to the betting situation then:
I´m betting 15-15
I´m betting 13-13
I´m betting 9-9

to lose: all three crossings have to be in a downfall
to win: all three crossings have to be in a upward

and:
to win: 2 have to win

and:
to recover: 1 has to win: (9-9)
to recover: 2 have to win (9-9;13-13)(13-13;15-15)(9-9;15-15)

And again: to be confronted with all bet-crossings losing at the same time is more rare than the other recover or win-situations.
To be sure of that, all crossings on their own have to end up at -2,7%.

btw: I didn´t claim the tests in Testing-zone the ultimate proof. They were just for explanation of the game and to proof it is possible to recover.

But I wouldn´t have called this a "Holy-Grail" if I wasn´t sure enough it is.

br
winkel

ps:

Then there HAS to be ANOTHER situation that gives a LOT higher probability! But then; why didn't winkel show us THAT ONE?

As you know I said before: Because there is no single one producing this. At a time 15-15 is doing this job at another time 12-11 is doing this job, changing daily.

[toby]

KFS, It took a real serious study to tell everyone the truth.

Gambler´s Fallacy fits Winkel´s system.

One hundred millions is enough to conclude.

KFS, you might save a plenty of forumers time if you fullproof certain systems.

Thank you very much.

[winkel]

one example:

this is the diagram of 12-13-12 for march 2007 RNG.



as you see:
a big [highlight]long[/highlight] losing of the F (>1)
a big long winning streak of "0" (R)
a first winning than sideway trend of "1" (N)

That means: If we had bet constantly 0vs>1 we would have been positive after about 200 bet-situations

and as a counter example of 13-12-12

[Kon-Fu-Sed]

winkel,

If you set up the rules, clear in every way, for a replicable test that you know will end positive after, say, 100,000,000 sessions we will do it.

You DO want a 3rd party confirming your findings, don't you?
That way it will be proved beyond any doubt that a math game has been beaten by math.

But until then I will - as there is no proof of the opposite - call your method a GF method.
Because the proof that it's a GF method is the usual:
"Record results until you see something (maybe a 15-15 situation) because then something (like a 14-16 situation) will happen".

Maybe so, but you can NEVER with a better-than-math-says probability predict WHEN it will happen.

Conclusion: This is a (ridiculously complicated) GAMBLER'S FALLACY LOSING METHOD.


Until I see the rules...
KFS

[Kon-Fu-Sed]

@toby,

Unfortunately not everyone will accept a truth.
There are numerous examples of messengers becoming a target...


KFS

[ernesto]

KFS!

What was the situation which occurs approximately once in 3,109 sessions?
And why was it interesting? It's a winner trigger?

ernesto

[winkel]

winkel,

If you set up the rules, clear in every way, for a replicable test that you know will end positive after, say, 100,000,000 sessions we will do it.

Until I see the rules...
KFS

Just take the rules of the clinical Test

bet any crossing <18 numbers
stopp at sesion result >+39
stopp at spin 50 with any result
start with spin 1
restart wtih spin 2
if there are 2 crossings at the same time bet that one with the highest count of numbers (eg. 13 12 12 = bet 13)

that´s it

You DO want a 3rd party confirming your findings, don't you?

I´m not in the need of this, it´s just for the gallery

[highlight]that a math game has been beaten by math.[/highlight]
This is your fallacy. G.U.T got nothing to do with math

You just act like I´d sell this for a lot of money.
I don´t press anyone to play
I do just believe in it, if that is so annyoing you don´t have to cope with it.

br
winkel

[TwoCatSam]

winkel

bet any crossing <18 numbers
stopp at sesion result >+39
stopp at spin 50 with any result
start with spin 1
restart wtih spin 2
if there are 2 crossings at the same time bet that one with the highest count of numbers (eg. 13 12 12 = bet 13)

Since I can't jump back one number, I have to use the Track 2 and jump back 20.  This should not change the outcome of any clinical study I do, should it?

Sam

[Kon-Fu-Sed]

@ernesto:
See test #2. That was the situation winkel suggested to be tested.

@winkel:
Everything is clear except which columns to use?
=0 Yes
>0 ?
=1 Yes
>1 ?
=2 Yes
>2 ?
=3 Yes
>3 ?
...

And of course it's math...
You SUBTRACT and ADD to your columns without taking anything but the winning numbers into consideration.
Hence: A math method.
What else could it be?


Regards,
KFS

[Kon-Fu-Sed]

-