History Sums Filter

[Falcon]

Hi Stan,

I notice the History Sums filter does not have an appearance in the Statistical Filter filter set.

Could this be included please.

Thanks

falcon

[stan]

That's not possible. The Statistical Filter can operate only with stats the yield a fixed set of values which is not the case for History sum values.

[Falcon]

Hi Stan,

I have to say I don't comprehend your response here re the inclusion of History Sums in the Statistical Filter List.

I see that History Sum Root is included in the Statistical Filter list - is History Sum Root not a derivative of History Sums?

regards

falcon

[stan]

Sum root has a well known value range 1 to 9 so it's easy to use it in Statistical filter. History sums don't have any fixed range so it doesn't fit into the Statistical filter concept.

[tdnl46w]

I would also like to see the History Sums filter added to the Statistical Filter list(and other statistics lists). It would open up a lot of new possible strategies in complex filters/custom analyzers(etc.). I understand that number intervals/history sums have no fixed values or ranges, but I hope that one of the options below would not be to difficult to implement.

1. Set a Min/Max sum value limit from the current winning draws sums from each level. This would be similar to the feature, "Lookup minimum and maximum possible sum values.".
2. Set a fixed maximum sum value limit(s) similar to the Number Differences statistics table which has a maximum value of 50+.
3. Implement an option to let the user decide on the Min/Max sum value limits. This would be similar to what we can do now with Segments. Many multi-core computers in recent years should easily be able to handle high values or cover the most important values.

Thanks.

[D1Magnet]

Hi I also think it would be a good idea to allow the history sums filter as a statistical filter, providing a user can set a min and max range to its parameters, or use the values from lookup min and max values.

Thanks

[stan]

There are already analyzers for negative/positive differences which basically provide the same value. Is that not good enough?

[tdnl46w]

There are already analyzers for negative/positive differences which basically provide the same value. Is that not good enough?

The positive/negative differences are similar to the #2 option from my last reply, so it should work. I will have do some more filter/testing.