Algorithm discussion

Discussion on strategies that can be developed within Lotto Architect
Post Reply
draughtsman
Site Admin
Posts: 112
Joined: Thu Jan 17, 2008 4:22 am
Location: Melbourne, Australia

Algorithm discussion

Post by draughtsman » Sun Jan 27, 2008 6:01 am

TRANSFERED FROM OLD FORUM

Hi there at Lotto Architect,

I see five pre-installed Algorithms for use in conjunction with rejection filters and was wondering about their best application, for I notice when I prepare my filtering project 'Remove out of bounds' algorithm never seems to exceed rejections created under 'Remove not found' algorithm. To achieve maximum filtering I have therefore adopted the use of this latter algorithm as my main filter approach, but is there some commonality between the two. And I have to say I have not been able to use Remove Values most Found' with much success. Is there some preferred Draw filter setup to get the best out of this one? Also any comments you can make in respect of the other two algorithms 'Remove if delay less than 2 or 10' would also be appreciated.

Thanks in advance

relowe

Hi relowe and welcome on the forums!

These 5 algorithms are predefined and they cover the most common uses on rejection filters. By all means, you can create much more detailed analysis by setting up new algorithms. I'll explain the most common usage of those predefined algorithms here.

"Remove out of bounds" vs "Remove values not found": The first one scans the data and determines the minimum & maximum values detected. Then it removes any values out of these limits. The latter scans the data and removes any values not found in the data. Obviously, the latter algorithm will reject more values including all rejected values from the first algorithm. If your history draw is small, then probably many of the in-between values will not appear but they still have a good chance to occur. In this case, the latter algorithm will remove them but not the first algorithm=more accurate rejection. When you have a lot of history draws, then both algorithms will essentially remove the same values. Still they are the 1st step in your rejection filters process.

"Remove most found" algorithm: This algorithm might look strange at first glance. Why someone wish to remove the most dominant values? This algorithm best works with a small selection of history draws (eg select last 5/10 draws) and a filter that has a large range of possible values - eg sums. It is more or less the same idea of a value that appears many times in a sort period of time; obviously it should not appear again to compete with classic probability. Take care with this algorithm as it will remove the most dominant values and if these are many, then all these values will be rejected. Perhaps it is better to establish a new set of algorithms based on "Occurrence" where you have determined what hit occurrence is considered enough for several filters and set this as the algorithm's filtering parameter.

"Remove if delay less than x" is a more complex process. "Less than 2" performs stronger rejection and therefore may be more inaccurate. "Less than 10" should be used in such cases. It is very accurate but removes fewer tickets. Feel free to practice with other values too.

Of course you can setup your own algorithms. These predefined ones are a basic guideline and the possibilities are not restricted to those included. There is also the "Difference" algorithm (not included as a predefined one) where you can try and explore its rejection & success rates too.

Post Reply

Who is online

Users browsing this forum: No registered users and 7 guests