Page 2 of 4

Re: Wheeling

Posted: Thu Sep 27, 2012 2:50 pm
Thanks again LA

Sorry - I have difficulty understanding these k's, l's and t's etc. Can you put this in an understandable format. I think you are saying the 1st block should be 0-k1 (4?) I will eventually get my head around this.

Regards - Frank

Re: Wheeling

Posted: Thu Sep 27, 2012 9:14 pm
by lottoarchitect
k is the block size of the covering, so if each block constructed has e.g. 6 numbers, k=6.

Re: Wheeling

Posted: Fri Sep 28, 2012 6:12 am
Thank you LA

I will go over these options and see if I get the desired effect.


Re: Wheeling

Posted: Thu Oct 04, 2012 7:39 am
Hi Sooz

Just to let know I tried your suggestion using the positions tab and it worked out quite well.
Played 20 (v) (5 winning numbers were in the 20)
Played 100 blocks
Played 6 Positions

Result: 5 three winning numbers and 3 four winning numbers!
I ran WG for 12 hours (over night) and even though the result was good (my opinion) I was concerned to see that that in some positions the percentage covering was still quite low? I tried to attach a copy however was unable so listed the results below:
Non-Guide Filters fit: 46 (45 - 45.00%)
Sum 91 [90 - 90.00%]
Position 1-96[98-98%]
Position 2-83[87-87%]
Position 3-70[70-70%]
Position 4-55[56-56%]
Position 5-48[48-48%]
Position 6-98[99-99%]
As can be seen Poistions 3 - 5 are low?
Should I have run WG longer and if yes how much longer.

Thanks for the advice which I will try again and then try the other suggestions made.
Regards - Frank

Re: Wheeling

Posted: Thu Oct 04, 2012 8:30 am
by lottoarchitect
when building with non-guide filters, make sure you have the "proportional" option enabled at the optimization tab.
For build techniques refer to viewtopic.php?f=19&t=427.
Generally, an overnight run benefits mostly from lower settings at both left & right bias sliders. So if e.g. in normal build you would pick 0 & 0, set the sliders at half that range towards the left side of each slider (i.e. in this example -500 & -500). If the settings are e.g. - 400 & -200, set them to -700 & -600 etc. That is for an overnight run where you are not in front of WG to manually adjust the sliders in steps as per the guide.


Re: Wheeling

Posted: Thu Oct 04, 2012 8:54 am
Thanks LA

Something new to look at which will hopefully improve the performance.

Refards - Frank

Re: Wheeling

Posted: Thu Oct 04, 2012 1:53 pm
by Sooz
Hey Frank, congratulations on your result! Did you choose the 20 numbers using GAT?
LottoArchitect I didn't know about about setting the bias sliders differently for a longer overnight run. Thank you so much for that!
Frank it also might help, for when you are at the screen, to just click once on Pause, and then immediately on Resume, because the program starts from a different spot when it resumes. That often gives me a quick improvement.

Re: Wheeling

Posted: Thu Oct 04, 2012 2:35 pm
Thanks Sooz

I must admit I was quite exited considering this was my first attempt. Although three's and four's aren't the jackpot it would have definitely covered my cost plus some profit!

I didn't use GAT as I am still trying to get used to WG. I have my own manual system I use for filtering which works quite effectively but every now and then lets me down badly. With this particular case I was fortunate to end up with 20 numbers as most of the time it is anything between 28 - 35 numbers.

Thanks for the advice on the pause button - I will certainly try that. Do I only do that once or do I do that lets say every hour?

Regards- Frank

Re: Wheeling

Posted: Thu Oct 04, 2012 2:40 pm
by Sooz
Frank your percentage covering does seem low - but of course it all depends on what you're asking WG to do. Here are the position filters I use with WG. I'm not convinced they are the best, they're probably too lenient, but they do give blocks which have numbers that are distributed more regularly - without being too regular.

position 1 1-20
position 3 3-39
position 4 11-47
position 6 25-49

I don't bother with position 2 because it has to fall between 1 and 3 anyway. that is, position 2 has to fall between 2 and 39. Similarly my position 5 has to fall between 11 and 48. Hope that makes sense.

Still, your win is quite impressive. Makes me wonder if I should increase my "v" and play a few more tickets. :-)


Re: Wheeling

Posted: Thu Oct 04, 2012 4:40 pm
by lottoarchitect
Sooz wrote:...
LottoArchitect I didn't know about about setting the bias sliders differently for a longer overnight run. Thank you so much for that!
The idea is really that during an overnight run the covering will be at a better (coverage-wise) position and generally in that case we usually have to lower the bias controls to make tighter the search and continue improvements. Setting this up that way, it may have a harder time to initially find improvements but it will have more correct settings to deal with optimization when it bypass the initial optimization, so it has more time (overnight) to work with more proper bias settings. Unless if the covering is a bizarre one, this will work ok. I work on a fully automated mechanism that fully adjusts the bias no matter what is the condition of the coverage. The user will simply have to enable the bias automation and let WG do its thing.