Why I can not find the "result wizard" in GAT Engine 2.3b?

G.A.T. Engine general discussion
Post Reply
bluemoon
Advanced
Advanced
Posts: 50
Joined: Sat May 05, 2018 3:47 pm

Why I can not find the "result wizard" in GAT Engine 2.3b?

Post by bluemoon » Fri Nov 23, 2018 3:36 pm

Hi, Mr. lottoarchitect:

I have a question through reading your User Guide in page 14:

"......Now we are to predict for draw#14, but of course we cannot now compute the hits as this draw at this stage is unknown. Thus the prediction for draw #14 resides across all GAT tables so it is therefore necessary to explore all of these tables for the best prediction. Fortunately a results wizard is used in this task as it would be virtually impossible to analyze by hand 1m or so GAT tables.¡±

Could you tell me what is the "result wizard"? Why I can not find it in GAT Engine 2.3b? Thanks.

User avatar
lottoarchitect
Site Admin
Posts: 1635
Joined: Tue Jan 15, 2008 5:03 pm
Location: Greece
Contact:

Re: Why I can not find the "result wizard" in GAT Engine 2.3b?

Post by lottoarchitect » Sat Nov 24, 2018 3:41 am

"Result wizard" is a metaphor for the panorama display.

bluemoon
Advanced
Advanced
Posts: 50
Joined: Sat May 05, 2018 3:47 pm

Re: Why I can not find the "result wizard" in GAT Engine 2.3b?

Post by bluemoon » Sun Nov 25, 2018 9:15 am

I understood. thanks.
What's "Total real draws to be tested for statistical hits evaluation" & "Total statistical data to be used for analysis"? Could you give me an example to explain their difference?
To get a good prediction, how can make an ideal setting for them?
Could you give me some links about their discussions in the forum? Thanks so much for your help.

User avatar
lottoarchitect
Site Admin
Posts: 1635
Joined: Tue Jan 15, 2008 5:03 pm
Location: Greece
Contact:

Re: Why I can not find the "result wizard" in GAT Engine 2.3b?

Post by lottoarchitect » Sun Nov 25, 2018 4:02 pm

The help file describes these parameters.

In order to make decisions for the generated GAT tables, the system produces hit performance statistics of the generated tables.

"Total real draws to be tested" defines how many real draws will be used to generate those statistics. The default is 100 which means the last 100 history draws will be used and each generated GAT table will be compared against them to produce the hit statistics. The data/prediction generated by each GAT table is not affected by this parameter but this affects which GATs will eventually end up in the panorama. For example, setting this parameter to 500, only the most long-term steady GAT performers that produced the most hits over the course of those 500 history draws will be proposed in the panorama. Avoid low settings to this parameter, like lower than 50 because this will hinder the ability to see the real ups and downs of each GAT (hot/cold cycles). Also a very high setting will propose long-term good GATs but this potentially may hinder the ability to find the GATs that are currently "more hot". The default 100 is a balance between these extremes really, suitable for every scan analysis.

'Stat.data" is the crucial parameter that instructs how many history draws will be used for each attempted prediction. It contains the loose definition of "the longest cycle in the data", so a lower setting may not be able to identify a complete cycle. Lotteries contain very small cycle patterns, in the length of a few draws mostly, practically undetectable by traditional observing of data or using statistics. The default setting of 20 is not large enough so to produce an averaged outcome and also not low enough so not to "miss" those small "patterns". This parameter along with the start point affect how the GATs are generated.

Also if you make a search for "tested draws" and "stat. data" you'll find several posts discussing different aspects of these parameters. For example
http://forums.anastasios-tampakis.net/v ... 744&p=3298
http://forums.anastasios-tampakis.net/v ... 725&p=3212

bluemoon
Advanced
Advanced
Posts: 50
Joined: Sat May 05, 2018 3:47 pm

Re: Why I can not find the "result wizard" in GAT Engine 2.3b?

Post by bluemoon » Mon Nov 26, 2018 8:09 am

Very clear! I've got it, thank you so much!

What's happening if the settings are below:
The tested draws=20; The stata. data=20 or
The tested draws=20; The stata data=100??

After a long scan, I got the Ratio% of (3-15)=100% when
the settings=15 Draws/15 Data. But the predictions of
Top IDs in the 3rd column were not a 3-hit yet (even 0 hit). Why? Is this reason from the lower settings (15/15)
or others?

User avatar
lottoarchitect
Site Admin
Posts: 1635
Joined: Tue Jan 15, 2008 5:03 pm
Location: Greece
Contact:

Re: Why I can not find the "result wizard" in GAT Engine 2.3b?

Post by lottoarchitect » Mon Nov 26, 2018 6:30 pm

That's exactly the reason you should not use small tested draws values. I recommend no values lower than 50. You have to be able and observe hot/cold cycles in any GAT. Over a range of 20 draws, outstanding hits will emerge but as any GAT is destined to also enter a cold cycle, this results in the artificial higher hit outcome. These GATs did indeed produce even 100% a 3+ hit production (with tested draws=15 as in your example). So, the question now is, will it do that again for the 16 consecutive draw? This is already an "overproduction" of hits. Use the default 100 or a minimum of 50 to get more accurate estimations of the hit performance.
What's happening if the settings are below:
The tested draws=20; The stata. data=20 or
The tested draws=20; The stata data=100??
Tested draws=20 is very low.

Post Reply

Who is online

Users browsing this forum: No registered users and 15 guests