Development - G.A.T. Engine 2.3

Any latest news will be posted here...
User avatar
lottoarchitect
Site Admin
Posts: 1457
Joined: Tue Jan 15, 2008 5:03 pm
Location: Greece
Contact:

Re: Development - G.A.T. Engine 2.3

Post by lottoarchitect » Fri Apr 12, 2013 7:03 am

The graph always shows the prediction results up to the last draw in the history, that is tested draws + run factor. However in order to maintain a stable size equal to the total data displayed when running at run factor 0, it eliminates the earliest hit points. Therefore, if at run factor 0 we show 100 points (this is the total tested draws), no matter what the run factor is we still show 100 points. That way we can directly compare statistics during the whole run factor range usage. So, if we are at run factor 5, what you see at the graph is the last 95 dots of the original graph (with run factor 0) plus the 5 new dots computed due to run factor. That is a designed behavior.

baalhabait
Advanced
Advanced
Posts: 80
Joined: Wed Apr 11, 2012 9:58 pm

Re: Development - G.A.T. Engine 2.3

Post by baalhabait » Fri Apr 12, 2013 10:54 am

Can you explain why this behaviour cannot be designed to show the 105 points? why it must eliminate the earliest hit points?
For example, i have statistics for 20 tested draws and since then 75 draws was added to the database, which means the run factor is 75 and the graph will show the last 20 hit points only.
Why dropping the 55 hit points before the last 20? there's a reasonable chance there is something there that is worth a view same like with viewing the last 20...
The graph should show all the 75 hit points within the run factor - not just the last 20.

User avatar
lottoarchitect
Site Admin
Posts: 1457
Joined: Tue Jan 15, 2008 5:03 pm
Location: Greece
Contact:

Re: Development - G.A.T. Engine 2.3

Post by lottoarchitect » Fri Apr 12, 2013 11:23 am

Three main reasons:
1. Upon deciding on a GAT for subsequent use, we maintain that equal-sized relation no matter the size of run factor. Thus if you decided on a GAT with 20 tested draws, you should still judge the performance based on that tested draws size regardless of run factor. Older data become obsolete in that sense and they are eliminated but we still monitor the performance of that GAT table with RF>0 to an equal amount of data which include the latest new data computed.
2. The statistics can be directly compared to the initial selection when we had run factor 0. As new GATs come and go with the various run factors, you can directly compare the performance as if using run factor 0 to make a selection in first place. This complements case 1 above.
3. Simplicity. If you select 100 tested draws, why should I display e.g. 150 tested results?

In short, this approach was picked to maintain the comparability between the runs with various run factor settings.

Post Reply

Who is online

Users browsing this forum: No registered users and 2 guests