This software may be useful for all who like to
create and test systems, or it could be used as a standalone
spin generator (either RNG or recorded spins on a file).
Of course every system creator hopes that their system is
going to make a profit, but if it does, are the results
"statistically significant"? Basically, what it means for a
result to be statistically significant is that it is unlikely
to have occurred by chance. In other words, there is something
more going on than mere randomness. You could argue that it
doesn't matter whether the result is significant in this sense
because the bottom line is : does the system make a profit?
Sure, but it's easy to be misled into thinking that the
system has real merit if it makes a profit in the short term,
when in reality you've just been lucky. This is especially so
if you're using some kind of progression which recovers
losses. If using a progression, a system could do rather well
(profit-wise) in circumstances where flat-betting would have
returned an overall loss. The benchmark of a "good" system is
that it returns a profit flat-betting over time,
and in order to achieve this you need an "edge"; ie the
system's results (in terms of bet selection) will be
statistically significant. The program tells you whether this
is the case or not.
There are three ways to enter spins.
- Click on the numbers in the layout on the left. This would be the preferred way if you're getting the spins from an online casino in real time.
- In the "spin options" area on the right you have the option of selecting RNG or spins from a file. To use RNG select the RNG radio button and then click the "Get A Spin" button to get a spin. Note that clicking this won't automatically add the spin to the marquee (you will have to additionally click the relevant number in the layout to do that).
- To get numbers one at a time from a spin file, select the
"Spin File" radio button in the spin options area and then
click the "Open File" button. When the file has been loaded
one spin at a time will be displayed when you click the "Get
A Spin" button. As with the RNG option, you will need to
click the corresponding number in the layout to add the
number to the marquee. The spin file should have one number
per line (if it's not in the correct format you may get a
warning that the file is corrupted).
When you have the result of a bet, ensure you have set how many numbers were bet by clicking the combo-box at the top right and selecting the number of numbers (between 1 and 35). On starting up the program the label "NTB" (Numbers To Bet) is shown in the combo-box. Then click the "W" (win) or "L" (loss) button depending on the outcome. The significance level will be updated on the first win and every bet thereafter. The initial bankroll is zero and assumes that one unit is staked on each of the numbers selected.
When you close the program you will have the opportunity to save the results. This file can be reopened when you start a new session by clicking the "Results File" button. The final results from the previous session will be loaded and you can continue as though the session was uninterrupted. So you could have multiple systems being tested at any one time, with a separate results file for each. The results file consists of two columns of data : the wins/losses and the sequence of significance probabilities. You could import it into a spreadsheet to see (for example) if there is any trend in the significance, or to do some other analysis.
Note that you can only save results when closing the program, so if you want to start testing a new system halfway through a session and want to save the current results you will need to close the software and restart it. If you want to start a new session but don't need to save the results, just click the "Clear" button to reset everything.
The meaning of "Significance"
The significance level is a probability; it is the probability
that you would have got the results you actually have (ie, the
number of wins versus losses) given that the system had no merit
at all. In other words, the assumption is that the bet selection
is no better than betting randomly. If you think about it,
you'll realize that a low probability is more
significant than a high one under these circumstances, because
it is evidence that the result is not just due to
chance. Conventionally, the threshold below which the result is
deemed significant is 5%, meaning that the probability is 1 in
20 that the result would have occurred by chance. If the trend
of the significance level is towards zero over time it indicates
a genuine edge. Significance levels less than 0.05 are shown in
green, otherwise they are shown in red.
A word about "sample size" and the interpretation of significance. The more bets you make which continue to show significant results, the less likely it is that randomness is playing a part. This doesn't mean you have to make thousands of bets, but on the other hand, you might be betting just 1 number which happens to hit twice in the first 10 spins, and obviously this doesn't mean you have winning system! Generally speaking, if you're playing few numbers you will need more data than if you're playing many. If in doubt, get more data.