Wheels and Selection methods

Discussion on strategies that can be developed within Lotto Architect
Post Reply
frog_15537
Casual
Casual
Posts: 15
Joined: Mon Jan 28, 2008 5:27 pm
Location: New Jersey

Wheels and Selection methods

Post by frog_15537 » Sat Apr 05, 2008 4:37 pm

Good afternoon LA and all

I have been playing around with the program, in particular the wheels portion. I have come upon something that I find confusing and would like to ask about it. Sorry if this rambles but I would like to make sure I make my question clear.

I have 2606 drawings listed in the database used by LA and I load them all and use them all when I open the program. I then click on the “numerical statistics” just to see the columns and then close it. Next I open the wheels tab.

Now I click on the “database tab” and set M=4 and P=4 and also check the full/abbreviated box. When I click on the “display wheels list” it brings up a list of wheels to be used for selecting tickets.

In this case I chose the wheel with 12 balls by left clicking on that line. Right clicking on that line puts the results into the “details” tab. Clicking on the “details” tab brings up a screen that allows access to the “Hot/cold/Due system”. That allows access to the small screen for the selection methods.

There are six selection methods and clicking on each method allows access to a set of 12 numbers, which are highlighted to the left. The six methods are as follows

1. Most Frequent Numbers
2. Least Frequent Numbers
3. Balanced Frequency
4. Combined (Most Bias)
5. Combined (Least Bias)
6. Combined (Balanced bias)

When I compare the numbers for 1 and 4, 2 and 5, and 3 and 6, highlighted on the left they are the same numbers for each set. By that I mean that selection method 1 and 4 give identical numbers to each other.

Okay, the questions are:
1. Is it supposed to be that way?
2. If it is supposed to be that way, is something happening in the actual suggested tickets that would make them different for each selection method? I just checked and the tickets are the same.
3. Or have I just gone deeper into what is happening than I need to be?

I hope this is clear. In my mind at the moment it seems the second set are just duplicating the first set. Thanks again for the all the work on the program and I’m looking forward to the newer version when it’s ready.

Bill

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

Re: Wheels and Selection methods

Post by lottoarchitect » Sun Apr 06, 2008 3:35 pm

Hi frog_15537,

every question is welcome when it aims to make the usage of a system more clear. Your questions have more or less been answered to this thread
http://www.forumforfree.com/forums/inde ... owtopic=60.

By reading this thread, to your questions
1. Same numbers for both 1&4,2&5,3&6 means that no selection conflicts occurred (as explained in the thread above). This is normal.
2. If you refer to the tickets generated by the application of the wheel you selected, no there isn't any difference since HCD aims to pick the best numbers (even if they are the same picked by the various methods) and not to re-organize them for the selected wheel.
3. Indeed, what you need to know about using HCD is to make reports and see what results obtained by altering the hot/cold/statistical size and algorithms. If you see satisfactory results in any of the settings you use, stick with them. Right-click on the generated report and there you'll see a menu which copies all the parameters and their hit ratio to the notepad so you can later observe and pick the best settings.

cheers
lottoarchitect

frog_15537
Casual
Casual
Posts: 15
Joined: Mon Jan 28, 2008 5:27 pm
Location: New Jersey

Re: Wheels and Selection methods

Post by frog_15537 » Tue Apr 08, 2008 1:13 pm

Thanks LA for the answer to my question and the reference to an earlier posting. I've read it once and will read it again as I go along. There is another question that i have on the HCD system.

I have two databases -- one containing all the draws (about 2600) and one containing only 100 draws. Regardless of which database I use I get the same numbers for the wheels. I even get the same numbers when I use the 2600 database and change the "test range" from 100 to 1000.

I would think that the numbers should vary at least some when you change the size of the database if they are based on the frequency of the numbers in the database. Where is my thinking going wrong?

Thanks in advance

Bill

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

Re: Wheels and Selection methods

Post by lottoarchitect » Tue Apr 08, 2008 7:01 pm

Well frog_15537,

the test range affects the internal statistical data process and is not directly related to the history draws we use for the HCD analysis. The amount of draws used by HCD is only affected by the draw filter selection. In my tests, changing the test range from 100 to 1000 usually produces a result which differs by 1-2 numbers,occasionally more, but it is possible not to have any changes at all. Actually this is good because it suggests that the algorithm produces an almost steady result beyond a given test range. However, if you set this value too low (e.g. 1-10 etc) you'll definitely observe changes to the suggested numbers. To give you an idea of what this "test range " represents, think of it as a variance modulator; the higher the value the more resistant to sudden changes the result will be (or in other words like being averaged to the most "promising" results). This allows the applied algorithm to "remember" the behaviour of numbers longer and make decisions based on what HCD observes. Generally speaking, if your lottery has some sort of obvious pattern, setting this value to the highest possible it will generally produce a more steady outcome, however this will not produce very high hits (will occur but more infrequently) compared to the case of a "local time lottery behaviour that suddenly changes every few draws" that can be trapped by using a much smaller test range for the statistical internal analysis i.e. 10 where you mainly can observe in the reports e.g. very good hits and very bad hits as well. Also the algorithm selection uses this test range to produce its internal data for analysis. By altering all the 4 parameters - hot/cold/test range/algorithm - (and the draw filter), the aim is to produce the desired hits in the report and stick with these parameters you have observed at least for the draws you expect HCD to produce the desirable hits again. I really don't think I can give much more detail on this subject without diving into technical details of the algorithms (and I don't really want to post that sort of information :roll: ).
Thus the test range indicates how quickly the applied algorithm will adapt to the changing trends of your lottery, however if it is too low, there isn't much room for a more accurate statistical analysis. I feel the best values to use are in the range of 1 (extreme case) to 100, however if you seek steadiness to the hits produced (i.e. for use in a number group as a negative filter and, why not, as part of a serotic wheel which uses such groups), use a very high test range.

cheers
lottoarchitect

Post Reply

Who is online

Users browsing this forum: No registered users and 8 guests