Occurrence Treshold

[PadawanLotto]

What does the simulation occurrence threshold settings do?

[Jerzy]

Padawan,

The occurrence threshold is a parameter based on the proportion of the occurrence of winning difference to the occurrence of the most frequent difference in the level (the blue bar in the simulation histogram). To enter the occurrence threshold, go to Tools --> Options --> History --> Estimates. This parameter applies to all levels. The winning history differences should be able to cross this threshold and the 1st prize ticket should be present in the package after filtering.

To find the right occurrence threshold for your game, you should calculate actual threshold for the past 10 - 20 drawings and keep calculating it for next drawings to get enough data.

Begin with calculating the occurence threshold for the latest draw.
First prepare the list of the latest winning differences, then hide the last draw result in the Winning Numbers --> Edit, and next run the simulation and analyze the histograms.
Calculate the threshold for the latest winning history differences which are located peripherally in the simulation histogram. Do it separately for the negative and positive differences.
No need to calculate anything for the differences located centrally.

For example:

The latest most negative difference was -73 and it was in the level 30. It had the occurrence of 518 and the most frequent difference (the blue one) had the occurrence of 24,785.
518/24785 * 100% = 2.08%. I round it down to 2.00%. It is the passing threshold for the level 30.

The latest most positive difference was 24 and it was in the level 26. It had the occurrence of 12,623 and the most frequent difference had the occurrence of 29,332.
12623/29332 * 100% = 43.03% . I round it down to 43% (the passing threshold for the level 26).

Probably the min occurrence threshold for the latest draw, in my example, will be 2.00% and the max threshold 43.00% , both values applicable to all levels, but we need to confirm it by doing more calculations.

Calculate thresholds separately for negative and positive differences. From the passing thresholds for negative difference select the lowest value. It will be min occurrence threshold for the latest draw. Similarly select the lowest value from the group of positive differences. It will be the max occurrence threshold for the latest draw. Usually the threshold for negative differences is lower than for positive differences, but not always.

With correct occurrence threshold entered centrally for all levels, you also need to enter the range of differences that cross this threshold into each level separately. Open My Estimates window, in the lower part of it in the Simulation [summary] you will see the min and max difference and a range of differences that cross the threshold. Click on the box, which has a small arrow in it, to enter this value to My Estimates and next transfer all entered values to the filter.

This is a very fine method of filtering, more effective than the ordinary method, because all levels are filtered proportionally.

[PadawanLotto]

Thanks jerzy, I will have to work with the info and see what it can do. The first test covered the first 9 levels but the last 2 were wrong, like I said I'll have to work with it some more.

Thanks again,

Padawan

[PadawanLotto]

So what you are saying is start out with the highest positive diff and the lowest negative diff to set the base line.

Calculate thresholds separately for negative and positive differences. From the passing thresholds for negative difference select the lowest value. It will be min occurrence threshold for the latest draw. Similarly select the lowest value from the group of positive differences. It will be the max occurrence threshold for the latest draw. Usually the threshold for negative differences is lower than for positive differences, but not always.

With correct occurrence threshold entered centrally for all levels, you also need to enter the range of differences that cross this threshold into each level separately. Open My Estimates window, in the lower part of it in the Simulation [summary] you will see the min and max difference and a range of differences that cross the threshold. Click on the box, which has a small arrow in it, to enter this value to My Estimates and next transfer all entered values to the filter.

Here you are saying to tweak and change the threshold for each passing level until the threshold setting allow all levels to pass?

[Jerzy]

Padawan,

There is only one winning difference in each level and it could be in the negative or positive part of the simulation histogram. Consequently, in each level there is only one passing threshold for that difference. The passing threshold for the difference in any level applies only to that particular level and it is not the same as the Occurrence Threshold which has the min and max value and applies to all levels.

To identify the min and max value of the Occurrence Threshold, which would be correct for all levels, we need to know the passing thresholds in other levels (not necesserily in all levels) for differences located in the most negative and the most positive parts of the simulation histogram. The differences located centrally, close to the blue bar in the histogram, play no role in finding the occurrence treshold.

Find a group of levels with lowest passing thresholds for the most negative differences. The lowest passing threshold in this group would become the min value of the OT. And next do the same for the most positive differences. The lowest passing threshold in the positive differences group would become the max value of the OT.

Calculate the OT for each of the latest draws and keep doing it for the future draws.
Begin with the latest draw:

1. Print a list of the latest winning differences.
2. Hide the latest draw result in the Winning Numbers.
3. Run the simulation for the previous draw which after hidding the real latest result became the new latest.
4. See the position of the true latest winning differences (from your list) in the simulation histogram
5. Calculate the passing threshold for the most negative and most positive differences (two groups of the passing thresholds, one for the most negative and one for the most positive differences).

After calculating the OT and entering its min and max values into Estimates in Options, run My Estimates and enter the ranges of differences that should cross the OT. It can be done manually only and for each level separately. Transfer My Estimates ranges to the Sum/Diffs filter and filter the package (Accept).
Afer filtering, uncover the last drawing result (that was hidden) and check the winnings.

I know it is very complicated but the concept of the OT is very interesting because of its effectiveness.

[PadawanLotto]

Ok, not to bad but it still leaves a few level incorrect and a boat load of combinations.
For the Rolling cash 5 I load a full package and the min/max diffs in my estimates. I then subtract 5 from each of the max diffs and set the min diffs to -37. Still leaves a lot of tickets but all levels are passing at this point.

I'm going to work more on the OT thing and see what I can come up with.

[Jerzy]

Padawan,

See also:
viewtopic.php?f=18&t=1703#p16769

[PadawanLotto]

Ok, I found that you don't have to run a simulation and search the histogram to find the differences occurrences. Just install a full package and look in the package analysis pos/neg differences. Click the occurrences to descending and the top occurrence will be at the top of the list. Of course you can also look at the latest column for the latest hits.

[ryker]

Read the above and the other forum, which refers to build 110513. I'm running build 110618.
I do not see a "Estimates" section in Options->History. Was this removed? Or changed to the "what if?"

[Jerzy]

ryker,

The build that you are running is an ordinary EL version and Occurrence Treshold is not available in it.

To use the Occurrence Threshold download and install version 5DEV[110614] which is the development build or beta version.

However, to avoid problems, do not open both versions at the same time. Both versions use the same data of past drawings.

[ryker]

Thank you