Re: Roulette WurfWeiten Kesselgucken Forschung von Selzer-McKenzie

On 9 Nov., 13:49, selzer_mcken...@xxxxxxxxxxxx wrote:
Roulette Wheel WurfWeiten
SelMcKenzie Selzer-McKenzie AlphaCode Research
Mister Selzer-McKenzie comment this Research project auf WurfWeiten
and Kesselgucken
2. Background
It is understood that a roulette wheel can become "biased", either
through wear of the rim or tilting of the wheel, so that the ball will
fall towards the centre at a reasonably predictable fixed area of the
stationary part of the wheel (known as the "drop zone").
Any wheel will exhibit this "bias" if sufficiently tilted, but the
necessary angle of tilt to cause this is not known. As SelMcKenzie
already owns calibrated digital tilt measuring devices (inclinometers)
the following proposal was made to test for the effects of tilt on a
wheel.
• Adjust the level of the wheel until there is no "bias" (i.e.. there
is no regular "drop zone").
• Measure the angle of tilt, if any, of the wheel in this position, to
determine whether simply
levelling the wheel accurately would result in a lack of bias.
• Tilt the wheel from this position, in at least four different
directions, until regular "drop
zones" are produced.
• Consider whether a spirit level is sufficiently accurate to
determine whether an installed
wheel will be biased, if simply levelling is found to be sufficient.
3. Method and results
A modern roulette wheel, in excellent condition, was received from TCS
Huxley. The roulette wheel was mounted onto a "tilt table" which is
normally used for tilt testing of weighing instruments. The table
allowed tilting of the wheel in any direction, up to a maximum angle
of 1.35°.
The proposed method outlined above was modified slightly such that the
tilt table (roulette wheel) was set so as to be "level" (N.B. small
angles of tilt could be measured on parts of the table as shown in
Annex 1). The outer rim of the roulette wheel was then divided into 8
equal segments (see Annex 1). 176 spins of the ball were conducted,
with a record made of which segment the ball fell into on each spin.
The ball was introduced to the wheel at the same point (segment 5) for
each spin on this test and for all subsequent tests. Table 1 shows the
results of this initial test.
Table 1 - Wheel "level"

Segment Number of times selected
1 23
2 18
3 24
4 22
5 22
6 28
7 18
8 21
By applying chi-square analysis to the results (X2 = 3.36 with 7
degrees of freedom) it was determined that the probability that the
measurement results were down to pure chance (i.e. random) was
approximately 85 %.
The tilt table was then tilted to give a maximum tilt of 1.35° in one
plane (right to left when viewing the wheel The test was repeated,
with Table 2 showing the results of this test.

Table 2 - ■Wheel tilted 1.35° "right to left"
Segment Number of times selected
1 40
2 37
3 11
4 8
5 10
6 10
7 22
8 38
Chi-square analysis of these results (X2 = 64.1 with 7 degrees of
freedom) indicates a probability of less than 0.1% that these results
are down to chance, i.e. there is a greater than 99.9% probability
that the results are due to a "bias" on the table.
The angle of tilt was then reduced, to 0.5°, 0.25° and then 0.1°.
Again, the tests were repeated at each of these angles of tilt. A chi-
squire analysis was made of each set of results. These results are
shown in Table 3.
Table 3 -Wheel tilted 0.5°, 0.25° and 0.1° "right to left"

Segment Number of times selected

0.5° 0.25° 0.1°
1 34 39 31
2 48 30 31
3 21 28 28
4 9 6 17
5 16 11 10
6 7 10 19
7 13 14 22
8 28 38 18
Chi-square (7 degrees of freedom) X2 = 62.2 X2 = 55.9 X2 = 17.8
Probability of "bias" > 99.9% > 99.9% 98%
The table was then tilted to an angle of 0.1° in the other three
planes. Only 80 spins were conducted for each of these tests. The
results are shown in Table 4.
Table 4 - Wheel tilted 0.1° in three other planes
Segment Number of times selected

Left to right Front to back Back to front
1 7 6 7
2 9 6 2
3 5 20 4
4 8 18 15
5 21 16 14
6 13 7 8
7 9 3 17
8 8 4 13
Chi-square (7 degrees of freedom) X2 = 17.4 X2 = 32.6 X2 = 21.2
Probability of "bias" 98% > 99.9% 99.5%
A check of a typical standard spirit level (obtained from a DIY store)
showed that a 0.1° angle equates to the bubble just touching the first
indicating line on the bubble window as shown in
4. Summary of results
The results indicate that when the wheel was "level" (it should be
noted that a certain amount of tilt was discernable on some parts of
the tilt table using the digital inclinometer) a figure of 85%
probability that the results were down to chance (i.e. random) was
achieved. It is possible that this figure could be improved to the
more generally accepted 95% level by making very fine adjustments to
the tilt table to ensure that it was perfectly level. However, it was
decided not to "chase perfection" as these initial results indicated a
good benchmark with which to compare the results of the tilt testing.
Tilting the wheel to 1.35° led to a greater than 99.9% probability
that the results were the result of a "bias" on the wheel. Reducing
the angle of tilt reduced the probability figure, but even at an angle
of only 0.1°, a probability of 98% that the wheel was "biased" was
still obtained. The results in the other three planes gave repeatable
results, with the lowest probability still in the region of 98%. An
analysis of the peak segment numbers also indicates that the "bias"
appears to move in relation to the plane in which the table is tilted.
5. Conclusions
It appears from the results of testing that an angle of as little as
0.1° (the first line on the bubble window of a standard spirit level)
will introduce a "bias" into the wheel. It can also be seen that the
"bias" will vary according to the plane in which the wheel is tilted.
A tilt of only 0.1° could be easily introduced if the wheel is not
mounted correctly, if the mounting and/or the table is not stable over
time, or by leaning on the table on which the wheel is mounted if it
is not sufficiently rigid. It therefore appears essential from these
initial findings that particular attention should be given to the
tables in the casinos onto/into which the roulette wheels are mounted.
It is suggested that calibrated digital inclinometers be used when
mounting the wheels (although it can be seen that a standard spirit
level can indicate, quite clearly, a tilt of 0.1°). The tables should
be sufficiently rigid and stable over time and not be liable to
tilting by someone leaning on the table. Stability over time could be
ensured by regular checks of the level and adjustment if necessary.
As it can be seen how easily a "bias" can be introduced, the next two
stages of the project can now be undertaken. These will begin upon
receipt of the roulette prediction
device and software from Selzer-McKenzie.

Roulette Wheel WurfWeiten Research
SelMcKenzie Selzer-McKenzie AlphaCode Research
Background
A number of individuals, most notably Selzer-McKenzie of SelMcKenzie
Technologies, claim to have developed roulette prediction devices that
will provide the user with an advantage over the casino when gambling
on roulette. Selzer-McKenzie claims that his prediction device can
give a return of 20 % on turnover over a period of time on a biased
wheel. This relates to a win once in every six spins when betting on
the predicted number and neighbours (predicted number and the two
numbers either side). SelMcKenzie claims a better return, even on a
wheel with no tilt.
Selzer-Mckenzie was contacted with a view to him loaning SelMcKenzie
one of his prediction devices (software running on a modified iPAQ)
for the purposes of this evaluation exercise.
The following proposal was made to determine the effectiveness of the
prediction device.
• Adjust the level of the wheel to introduce a "bias" (i.e. a regular
"drop zone") and use the
prediction device to attempt to predict the outcome (section 3.2 of
proposal).
• Adjust the wheel so that there is no discernable "bias" and use the
prediction device to
attempt to predict the outcome (section 3.3 of proposal).
N.B. during this stage of the project SelMckenzie contacted
SelMcKenzie requesting that we "certify" his device. I responded to
him that we were unable to endorse or "certify" devices of this kind.
He also claims to have developed a device that can be fitted to a
roulette wheel to negate the effect of prediction devices and other
prediction methods. I indicated to Mark that we could be willing to
investigate the effectiveness of this particular device as it would be
of interest to the gaming industry.
3. Prediction device (Method of operation and calibration)
The prediction device is a modified iPAQ running specially written
prediction software. The iPAQ has been modified to allow the
connection of a simple pushbutton switch which is used to operate the
prediction software. The prediction software has two modes; a
calibration mode to calibrate the device against the wheel, and a
normal operation mode for predicting the number. The normal operation
mode also permits recording of the actual number selected on
completion of each spin. This is used by another piece of software
which produces a histogram to identify the optimum offset value (to
account for "scatter"). Normally, this software will automatically
correct the prediction software to take account of any variation in
the offset value. However, this automatic correction was disabled for
test purposes so that the correction could be done manually.
A video clip of a roulette spin was provided by Selzer-McKenzie. The
purpose of this was to enable me to practice using the device against
the video clip to see if I could get the prediction device to predict
the number against a known outcome, i.e. the ball dropping from the
rim at the same point and coming to rest in the same slot every spin.
After some practice I was able to get the device to predict a number
such that a win would have been obtained
(using the predicted number plus neighbours betting approach) on
practically every practice spin. This proved that I was able to
operate the device correctly, and that the device itself was
functioning correctly. Selzer-McKenzie supplied the device calibrated
against this video clip. However, I repeated this test on a couple of
occasions after re-calibrating the device myself and I was able to
achieve the same consistent performance.
In practice, the device operates in the following manner. For
calibration, only the ball is "clocked". The ball is spun, and each
time it passes beneath the reference point (drop zone) the switch is
pressed. The device will ignore all revolutions faster than 1200
milliseconds. When the correct revolution is captured the device emits
a high pitched "beep". The switch is then pressed when the ball hits
the wheel so that the device can calculate the "time to drop".
With the device calibrated it is then possible to attempt predicting.
The "zero" on the wheel is clocked when it passes the reference point
for the first time and is then clocked again when it passes the
reference point for the second time. Clocking of the "zero" enables
the speed of the wheel to be calculated.
With the speed of the wheel calculated, the ball is then clocked as it
passes the reference point. The ball is clocked each time it passes
the reference point until the ball speed falls within the capture
envelope (in this case 1200 - 1400 ms). When this occurs, the device
predicts the number. Any revolutions greater than 1400 ms result in an
"error" message.
The device says "now" when it estimates that the ball should drop from
the rim into the rotor. This gives an indication of how well the
device has been calibrated and how well that particular spin was
clocked. If the device does not say "now" when the ball drops it is
likely that the wrong revolution of the ball was captured, e.g. the
revolution captured was on the extreme limits of the 1200 - 1400 ms
acceptance envelope.
On completion of the spin it is then possible to enter the actual
number that was obtained to enable the optimum offset to be determined
(caused by the "scatter" of the ball as it hits the rotor). When an
offset has been determined, this is compensated for by changing the
clocking point accordingly, e.g. if the offset is calculated to be
plus 5 then the reference point is moved forward 5 slots in the
direction of travel of the ball. The "zero" of the wheel and the ball
are both clocked at this new reference point.
4. Test method and results (section 3.2)
The TCS Huxley wheel was tilted by 1.35° in one plane to create a
"bias" (known drop zone). Over 50 spins the ball hit the vertical
diamonds in segments 3 and 4 (see Annex 1) 49 times out of the 50
spins (24 hits for the diamond in segment 3 and 25 hits for diamond in
segment 4). The horizontal diamond between these two vertical diamonds
was therefore selected as the reference point for "clocking".
The device was calibrated against the wheel using the procedure
described above. The calibration had to be repeated a number of times
to achieve a sufficient level of confidence that the bal was dropping
from the rim when the device said "now", and also that the predicted
number was below the ball when it dropped. It should be noted that
this situation is not expected to occur on every spin. However, it is
expected that it should occur "more often than not". After some trial
and error this situation was obtained on 38 out of 50 spins.
It should be noted that the calibration of the device had to be
repeated through the course of the testing due to changes in the ball/
wheel performance, i.e. on some days the ball would complete more
revolutions following the 1200 - 1400 captured revolution than on
other days. This need to change the calibration led to a large number
of repeat tests being performed, hence the delay in completion of the
project.
With the device calibrated, the "scatter" for this particular wheel
had to be determined. The "scatter" is the amount of variation in the
actual slot that the ball comes to rest in, compared with the slot
beneath the ball when it drops from the rim.
Further spins were conducted to analyse the difference between the
number under the ball when it dropped from the rim and the number
(slot) that the ball falls into. Over the course of 25 spins, the
average difference ("scatter") was plus 9 (although the range of
scatter was in the region of 25 slots).
Test1
This was performed by clocking at the original reference point (i.e.
no offset). The actual number was entered into the offset software
following each prediction. On completion of 50 spins the offset
software indicated that an offset of plus 5 (5 slots in the direction
of travel of the ball) would have resulted in 14 wins out of 50
(number plus neighbours). Without the offset, i.e. by comparing the
predicted number to the actual number, 4 wins were obtained.
Test 2
This was also performed by clocking at the original reference point
(i.e. no offset). The actual number was entered into the offset
software following each prediction. On completion of 50 spins the
offset software indicated that an offset of plus 9 (9 slots in the
direction of travel of the ball) would have resulted in 12 wins out of
50 (number plus neighbours). An offset of plus
5 (the figure from Test 1) would have resulted in 6 wins out of 50
(number plus neighbours).
Without the offset, i.e. by comparing the predicted number to the
actual number, 7 wins were
obtained.
Test 3
Based on the results of the previous two tests, Selzer-McKenzie
suggested that an offset of 5 slots be used when clocking the "zero"
and the bal. On completion of 50 spins the offset software indicated
that an additional offset of plus 5 (5 slots in the direction of
travel of the ball) would have resulted in 15 wins out of 50 (number
plus neighbours). This is effectively an offset of 10 slots (the
original offset of 5 slots plus this additional offset of 5 slots).
However, without the additional offset, i.e. by comparing the
predicted number to the actual number, 3 wins were obtained.
Test 4
Based on the results of the previous test, an offset of 10 slots was
used when clocking the "zero" and the ball. On completion of 50 spins
the offset software indicated that an additional offset of plus 31 (31
slots forward in the direction of travel of the ball) or,
alternatively, minus
6 (6 slots in the opposite direction of travel) would have resulted in
14 wins out of 50 (number
plus neighbours). This is effectively an offset of 4 slots (the
original offset of 10 slots minus
this proposed offset of 6 slots). However, without the additional
offset, i.e. by comparing the
predicted number to the actual number, 4 wins were obtained.
Test 5
Based on the results of the previous test, an offset of 5 slots was
used when clocking the "zero" and the ball. On completion of 50 spins
the offset software indicated that an additional offset of plus 4 (4
slots forward in the direction of travel of the ball) would have
resulted in 13 wins out of 50 (number plus neighbours). This is
effectively an offset of 9 slots (the original offset of 5 slots plus
this additional offset of 4 slots). However, without the additional
offset, i.e. by comparing the predicted number to the actual number, 8
wins were obtained.
Test 6
The results of the previous tests indicated that the offset was
varying with each test and that "peak chasing" was occurring. This
test was therefore performed by clocking at the original reference
point to check the performance of the device. No offset was used when
clocking the "zero" and the ball as per tests 1 and 2. On completion
of 50 spins the offset software indicated that an offset of plus 14
(14 slots forward in the direction of travel of the ball) would have
resulted in 14 wins out of 50 (number plus neighbours). However,
without the additional offset, i.e. by comparing the predicted number
to the actual number, 7 wins were obtained.
Test 7
Based on the results of the previous test, an offset of 10 slots was
used when clocking the "zero" and the ball. On completion of 50 spins
the offset software indicated that an additional offset of plus 32 (32
slots forward in the direction of travel of the ball) or,
alternatively, minus 5 (5 slots in the opposite direction of travel)
would have resulted in 14 wins out of 50 (number plus neighbours).
This is effectively an offset of 5 slots (the original offset of 10
slots minus this proposed offset of 5 slots). However, without the
additional offset, i.e. by comparing the predicted number to the
actual number, 8 wins were obtained.
Test 8
This test was performed with an offset of 8 slots (approximately
midway in the range of offset values obtained) when clocking the
"zero" and the ball. On completion of 50 spins the offset software
indicated an offset of zero. 13 wins out of 50 (number plus
neighbours) were obtained.
5. Test method and results (section 3.3)
The tilt table was levelled in order to try and eliminate the bias on
the wheel. 200 spins were performed and the segment number in which
the ball fell from the rim was recorded. The results obtained are
shown in Table 1.

Segment Number of spins
1 25
2 26
3 20
4 22
5 22
6 23
7 20
8 40
By applying chi-square analysis to these results it can be seen that
there is a 90% chance of a bias on the wheel (due to the peak on
segment 8). A slight tilt was therefore applied (lifting the wheel
under segment 4) in an effort to eliminate the peak at segment 8. A
further 160 spins were performed and the segment number in which the
ball fell from the rim was recorded. The results obtained are shown in
Table 2.

Segment Number of spins
1 16
2 21
3 18
4 18
5 20
6 15
7 25
8 26
By applying chi-square analysis to these results it can be seen that
there is now a 40% chance of a bias on the wheel. There is still a
noticeable peak at segments 7 and 8, but it was felt that the wheel
was providing a more random performance for this stage of the testing.
As a reference point for clocking, the horizontal diamond midway
between the vertical diamonds in segments 7 and 8 was used. From the
results of the previous set of tests, an offset of 8 from this
reference point was selected as the clocking point for the wheel
"zero" and the ball. This clocking point was used for five sets of
tests (50 spins in each set).
Test 9
On completion of 50 spins the offset software indicated that an
additional offset of plus 15 (15 slots forward in the direction of
travel of the ball) would have resulted in 11 wins out of 50 (number
plus neighbours). However, without this calculated offset, i.e. by
comparing the predicted number to the actual number, 8 wins were
obtained.
Test 10
On completion of 50 spins the offset software indicated that an
additional offset of plus 3 (3 slots forward in the direction of
travel of the ball) would have resulted in 12 wins out of 50 (number
plus neighbours). However, without this calculated offset, i.e. by
comparing the predicted number to the actual number, 6 wins were
obtained.
Test 11
On completion of 50 spins the offset software indicated that an
additional offset of plus 30 (30 slots forward in the direction of
travel of the ball) or, alternatively, minus 7 (7 slots in the
opposite direction of travel) would have resulted in 12 wins out of 50
(number plus neighbours). However, without this calculated offset,
i.e. by comparing the predicted number to the actual number, 5 wins
were obtained.
Test 12
On completion of 50 spins the offset software indicated that an
additional offset of plus 11 (11 slots forward in the direction of
travel of the ball) would have resulted in 9 wins out of 50 (number
plus neighbours). However, without this calculated offset, i.e. by
comparing the predicted number to the actual number, 8 wins were
obtained.
Test 13
On completion of 50 spins the offset software indicated that an
additional offset of plus 20 (20 slots forward in the direction of
travel of the ball) or, alternatively, minus 17 (17 slots in the
opposite direction of travel) would have resulted in 10 wins out of 50
(number plus neighbours). However, without this calculated offset,
i.e. by comparing the predicted number to the actual number, 6 wins
were obtained.
6. Summary of test results
6.1. The test results for stage 3.2 of the project are summarised in
Table 3.

Selected offset (slots) Wins with this offset Offset software analysis
(slots) Total calculated offset (slots) Wins with calculated offset
0 4 Plus 5 Plus 5 14
0 7 Plus 9 Plus 9 12
Plus 5 3 Plus 5 Plus 10 15
Plus 10 4 Minus 6 Plus 4 14
Plus 5 8 Plus 4 Plus 9 13
0 7 Plus 14 Plus 14 14
Plus 10 8 Plus 5 Plus 5 14
Plus 8 13 0 Plus 8 13
Avg = 6.75 Avg = 13.6
For each spin, 5 units would be bet (predicted number plus
neighbours). For the predicted number with the selected offset, an
average of 6.75 wins per 50 spins would result in a loss of 7 units as
shown below.
- (250 + 6.75) + (6.75 x 35) = - 7 units
For the predicted number with the calculated offset, an average of
13.6 wins per 50 spins would result in a gain of almost 240 units as
shown below.
- (250 + 13.6) + (13.6 x 35) = 239.6 units (95.8 % on turnover)
6.2. The test results for stage 3.3 of the project are summarised in
Table 4.

Selected offset (slots) Wins with this offset Offset software analysis
(slots) Total calculated offset (slots) Wins with calculated offset
Plus 8 8 Plus 15 Plus 23 11
Plus 8 6 Plus 3 Plus 11 12
Plus 8 5 Plus 30 (minus 7) Plus 1 12
Plus 8 8 Plus 11 Plus 19 9
Plus 8 6 Plus 20 (minus 17) Minus 9 10
Avg = 6.6 Avg = 10.8
For the predicted number with the selected offset, an average of 6.6
wins per 50 spins would result in a loss of 12 units as shown below.
- (250 + 6.6) + (6.6 x 35) = - 12.4 units
For the predicted number with the calculated offset, an average of
10.8 wins per 50 spins would result in a gain of almost 139 units as
shown below.
- (250 + 10.8) + (10.8 x 35) = 138.8 units (55.5 % on turnover)
6.3. Ten series of 50 random numbers were generated to compare the
performance of the device against randomly selected numbers. Each
series of 50 random numbers were compared against the actual numbers
obtained during each of the tests. On average, 6.4 wins per 50 spins
would have been achieved using the random number plus neighbours
betting approach. For the sets of random numbers, the maximum number
of wins per 50 spins was 11, and the minimum number of wins per 50
spins was 3.
For the randomly generated numbers, an average of 6.4 wins per 50
spins would result in a loss of almost 20 units as shown below.
- (250 + 6.4) + (6.4 x 35) = -19.6 units
7. Conclusions
As expected, and as shown in stage 3.1 of the project, tilting the
table introduces a definite bias ("drop zone") onto the wheel. This
made determining the reference point for clocking the "zero" on the
wheel and the ball during each spin very easy during stage 3.2 of the
project.>From the testing conducted for stage 3.2, the results appear to

indicate that the prediction device could predict the number such that
the peak number of wins would occur within a relatively narrow band of
10 slots (the calculated optimum offset varied from 4 slots to 14
slots). If the calculated offset was used then an average of 13.6 wins
per 50 spins could be obtained; a return of almost 100 % on turnover.
However, there was an element of "chasing the peak" involved with the
testing, as on only one occasion (test 8) did the selected offset
coincide with the peak number of wins. Further tests would need to be
conducted to determine how consistently an offset of 8 would give this
number of wins.
Without adjusting the offset to coincide with the peak, which
obviously can't be done in practice as it is "after the event", it can
be seen that the performance of the device was only marginally better,
on average, than the randomly generated numbers. It should be noted,
however, that for three of these tests no offset at all was used when
clocking the "zero" and the ball. The probable reason for the peak
moving within the 10 slot band is most likely the result of "scatter"
of the ball when it drops into the rotor. It would appear that the
wheel supplied is not suitable for predicting as the scatter is too
random.
With the bias introduced into the wheel I could "feel" when the spin
was potentially going to produce a "win". On these occasions, when the
predicted number was called it was possible to observe the ball and
predicted number arriving at the drop zone simultaneously. "Scatter"
would ultimately determine if a "win" was achieved on these occasions,
but I got a real sense that the device had provided a good prediction.
I could also "feel" when the spin was unlikely to produce a "win" as
there was no cross-over of the number and ball at the drop zone. This
was invariably caused by catching the wrong revolution (i.e. on the
extreme of the 1200 -1400 ms capture envelope) of the ball such that
the ball would drop from the rim when the predicted number was on the
opposite side of the wheel. It should be noted that wins were
sometimes obtained in this situation due to the random nature of the
scatter pattern.

I would therefore expect that with a more consistent "scatter", a
prediction device of this nature could work very well in this
particular test situation. However, testing of the device took place
in laboratory conditions with a very heavily biased wheel. I was able
to stand immediately over the table which made clocking of the "zero"
and the ball easy. This would not be the case in an actual casino
where clocking would have to be done subversively. In addition, in
practice it is unlikely that such a heavily biased wheel would ever be
found.
To support this, stage 3.3 of the project shows that when the bias is
reduced (N.B. I was unable to completely eliminate the bias) the range
of the calculated offset varied enormously from minus 17 to plus 15
(almost the whole wheel). In addition, even when the calculated offset
was used, the number of wins per 50 spins reduced from an average of
13.6 on the biased wheel to 10.8. It was noticeable that the histogram
produced by the offset software was a smoother profile during these
tests when compared with the profile of the histogram for the biased
wheel tests. Again, without adjusting the offset to coincide with the
peak, the average number of wins per 50 spins (6.6 wins) is comparable
with the randomly generated numbers.
Therefore, it can be seen that on a wheel with a definite bias ("drop
zone") and a manageable scatter, a prediction device of this nature,
when operated by a "skilled" roulette player, could obtain an
advantage when used in a casino. Selzer-McKenzie claims that he can
achieve an advantage of 20 % on turnover over a period of time on a
wheel that exhibits these characteristics.
However, the testing at stage 3.3 shows that the effectiveness of such
a device could be negated (minimised) by eliminating bias from the
wheel as far as possible, i.e. ensuring that the wheel is levelled
correctly and is not worn. In addition, wheels having a sufficiently
random scatter pattern, i.e. using a light ball with shallow pockets,
would also reduce the effectiveness of the device. This is
demonstrated by the results at stage 3.2 which show that even with a
significant bias on the wheel, the scatter appeared to produce a
constantly changing offset (although with fairly tight limits), i.e.
the wheel supplied was not suitable for predicting due to the scatter
being too random.
A simple way to prevent these devices from having any effect at all
would be to call "no more bets" as soon as the ball is spun.
Alternatively, as the predicted number is not known until shortly
before the ball falls from the rim, the practice of calling "number
plus neighbours" (where the croupier then places the necessary bet)
could be stopped. This would make it difficult for the person using
the device to place the bet on the predicted number and the neighbours
(5 numbers in total)
WurfWeiten Research
SelMcKenzie Selzer-McKenzie AlphaCode Research

Heiratsangebote, Jobangebote, Geschenke und vieles mehr
Werte Gambling- unde Bettingtrader,
zunächst mal: für die Präsente werde ich noch mit den Schenkern
persönlich Kontakt aufnehmen aber ich muss doch schon sagen, es
überrascht mich sehr. Nicht nur die Geschenkpräsente, die hier
ankamen. Mit einer vollbeladenen Sackkarre hat der Paketbote die
Sachen hier hereintragen müssen. Woher kennen Sie eigentlich die
Daten, das wundert mich.
Aber ich bitte Sie trotzdem, zukünftig oder nächstes Jahr nichts mehr
zu veranlassen, sondern das Geld vielleicht weiner Spende für eine
Hilfsorganisation zukommen zu lassen, falls Sie wirklich was zu
schenken beabsichtigen sollten. Ich will nicht undankbar sein, aber
eigentlich habe ich alles und weiss ehrlich gesagt momentan garnicht,
mit wem und wo ich die vielen Sektpullen usw. verzehren soll.
Natürlich, und das rechne ich mir als Wissenschaftler jetz mal selber
hoch an, Sie haben druch meine Aktienprognosen und Wettprogniosen
viel Geld verdient, aber denkebn Sie daran, Sie haben auch immer Ihr
eigenes Geld investieren müssen und auch ich bin vor einer
Fehlprognose der Aktrien- und Wettmärkte nicht geschützt.
Wenn ich mir überlege, ich selbstr habe eigentlich so gut wie nichts
verdient, weil ich als armer Schlucker eben nur mit 5-Markstücken
traden und wetten kann und selbst bei 100% Gewinn ist das dann immer
mager.
Aber ich will da noch was sagen, es ehrt einen, als nicht mehr der
Jüngste kommen hier reihenweise Heiratsangebote und aber auch gute
Jobangebote an. Sie wissen, auf Videpplattformen habe ich eigene
Musikstücke aber auch eigene Autotests veröffentlicht. Was glauben
Sie, wieviel Liebesbriefe hier von Teenes e angkommen. Als Altrocker
ist man offenbar auch mit Falten noch beliebt , oder vielleicht denken
die Kids vor Begeisterung der Music ich sehe noch so aus wie in den
70ern.
Aber auch Jobangebote kommen hier an, teilweise nicht schlecht. Das
Internet wird offenbar auf den Vidoeplattformen sehr intensiv
gelesen, und ich habe bei einigen Musicstücken aber auch Autotests
teilweise bis zu 40.000 Clicke täglich, laut dem Anzeiger.
Und da will ich auch gleich nmoch die Fragen hier beantworten. Die
Fahrzeuge bei den Autotests werden mir natürlich zur Verdfügung
gestellt und ich brauche keinen Pfennig dafür zu zahlen. Ich bin ja
garkein Techniker, sondern Mediziner, und habe das aus Hobby gemacht.
Aber das hat dafür geführt, dass hier schon mehrfach die
Marketingabteilungen der Autohersteller angerufen und mir einen Wagen
zur Probe angeboten haben, obwohl meine Test manchmal auch sehr
kritisch waren.
Sehen Sie diese japanische Studie, da ruft aus Nakasaki der Hersteller
an, sagt, am Airport München liegt ein Freiflug nach Tokyio bereit,
Hotel wird auch bezahlt und ich gleich los, nur um die Studie dann 1
Tag in Toky herumzufahren und meinen Videotest im Internet
abzuliefern. Aber das machen die Autohersteller mit den
Autojournalisten von Presse und Rundfunk überalle, die bekommen auch
alles bezahlt und ich darf mich als Nobody jetzt dazugehörig zählen.
Es macht Spass, auf Kosten anderer in der Welt herumzufliegen, ein
test zu machen, nur weil man sich mal einst als Hobbytester empfohlen
hat. Wie gesagt,. Nochmals besten Dank für die vielen Präsente, das
tut auch gut, weil man weiss, wie beliebt man doch im Internet sein
muss, denn ich kenne Sie alle ja garnicht.
Selzer-McKenzie

.

Loading