Filter priority

[relowe]

Hi LA,
I think I have my observation of what goes on within WG correct.
I notice when doing a run and there are non guide filters active the program appears to want to achieve 100% in these non guide filters at the expense of what is happening in main coverage. ie say a SUM guide filter is running at say 99% and main coverage also at 99% when an improvement is found within the SUM filter it may then achieve 100% but when this occurs the main coverage optimization level might drop back to 98%. Ie it appears that the filter achievements are overriding the main coverage achievements. I think I would like to see this the other way around. ie the main objective would be the achievement of say 100% at the main coverage level and then for any non guide filters which are active they would have to do their thing within the improvements generated at main coverage; consequently I would find the non achievement of 100% for the non guide filters as acceptable if 100% has been achieved at main coverage.
Might this be a consideration for the next Version?

relowe

[lottoarchitect]

Hi relowe,

The most common setting when filters are involved is all filters/coverage have equal priority and the improvement criteria is set to proportional. However, if you want this particular approach you suggest, you can emulate this behavior by setting the coverage engine's priority to 100% and the non-guide filters to a very small value (i.e. 1-10%). Also set the improvement criteria mode to absolute. Using these settings, 99,999% of the time whenever an improvement of the coverage occurs, it will almost certainly override any improvement of the filters. On the other end, if an improvement occurs at the filters side, it has to be quite large to be accepted, given a detioration of the coverage occurs. Finally, if the coverage remains the same, the smallest filter improvement will be accepted. Hope this covers your needs.

cheers
lottoarchitect

[relowe]

Hi LA,

Thanks for your response, and yes your suggested method of setting priorities of filters and main coverage does indeed satisfy my concern of filter results seeming to override main coverage. My thinking on this and which I should have included in my initial post, was that as there is already an operating condition within WG where filters must operate to the parameters specified for them ie when they are in Guide Mode then in effect Main Coverage has to respond as best it can after the filters have been satisfied. Thus perhaps when Guide Mode is not invoked then Main Coverage might be by default the primary area where improvements are to be won and the filters to be secondary. But as you mentioned this can be satisfied by appropriate prioritizing of the various filters and main coverage.

Thanks for the discussion

relowe

[lottoarchitect]

Hi relowe,

I understand how you come to this idea. However, when we set filters to guide mode, we basically instruct the engine to produce a set of all possible candidate blocks for the covering (actually we define internally a mother covering with the filters in guide mode), before any actual optimization begin. After that initialization step, the filters are not involved in the optimization process, since they will be always to 100% fit. Therefore, the only real engine operating underneath is that of the coverage, so it is on its own and not competing against any filters in guide mode. The opposite, as per your suggestion would mean we should have the possibility to set the coverage engine in guide mode (if such a thing could be possible) so to initialize in advance a covering with a certain coverage (obviously the maximum possible) and then the coverage engine would cease operation during optimization and only the non-guide filters should be evaluated. By that means, yes the filters would be operating only within the constraints set by the coverage engine. Actually, what I described here is a new optimization mode I have plans to add in WG. You'll be able to import a highly optimized covering (coverage wised) and without affecting its actual coverage properties, the engine would try to fulfill the non-guide filters. Since we cannot alter anymore the actual blocks, what we can alter is the placement of the numbers themselves, what I call the "numbers rearrangement" feature and is indeed missing from WG. In this mode, the coverage engines have to be disabled. For example you could import an 100% covering and find the best arrangement that can fulfill the filters. In current WG operation, the engines would decide to reduce the coverage a bit so to let more filters qualify, always based on the priorities defined.

cheers
lottoarchitect

[relowe]

Thanks LA,

I am certainly happy with the resource that WG now brings to the user, and I even thought I might have got ahead of you there at one stage with my suggestion above but as usual you come back with even more for this product. It is already by far the best wheeler available and you intend to make it better. Excellent.

Best wishes and good lottery luck to all for the New Year.

relowe