Computer Generated Drawings vs Physical Ball Drawings vs GAT

G.A.T. Engine general discussion
Post Reply
Jmichael
Newbie
Newbie
Posts: 3
Joined: Thu Jul 10, 2014 9:01 pm

Computer Generated Drawings vs Physical Ball Drawings vs GAT

Post by Jmichael » Mon Jul 21, 2014 12:55 pm

Hello LA...
In my State, every drawing is done by the lottery with their random generator. I was wondering (I searched the forums and couldn't find much on this) if you have done any comparisons using GAT with lotteries that draw with random generators and those that use the physical ball method ??

and one off-topic question.. when I back test and enter in the unknown drawing numbers into the "compare" function to scan down through the various GATs' in the Panorama view to see hits in the "predicted" numbers, I understand the yellow highlighted numbers are the ones that are -/+ 1 number from the actual hit...so my question is how is this feature utilized to the benefit of the player ??, as I see it, respectfully I say..."if it's 1 number off, it might as well be 10 numbers off !", I can't see what good knowing this info can do to help me pick a better GAT ! Maybe I'm missing something that is very apparent :)

Thanks LA..., I enjoy the GAT program, and learning about it.

John

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

Re: Computer Generated Drawings vs Physical Ball Drawings vs

Post by lottoarchitect » Mon Jul 21, 2014 5:13 pm

Hi John, generally speaking GAT will produce lower hits on random computer generated numbers compared to a physical machine. The reason is, simply put, RNG numbers generation lacks those physical factors that contribute to the reduced randomness of the system. GAT is better used and delivers higher hits on barrel machines because of that. It is not impossible to get good hits even if an RNG system is analysed; if the RNG used is of lower quality (it can't produce completely random behavior) it is possible for GAT to identify elements in there to take advantage of. Truth is however, only real barrel machines have those properties more or less that contribute to the reduced randomness of the system and make more apparent its extraction. I can't say for any RNG system however if it can have elements that contribute to the system having reduced randomness behavior.
About the yellow colouring, experience has shown that a GAT can be very close to predicting the shape of a draw, thus quite often you'll see a GAT predicting some correct numbers and the others will be mostly one-offs. This particular observation lead to the introduction of the yellow colouring, so to make it easier to distinguish those GATs that come so close. There isn't any direct way to utilize that information, other than to note particular GATs do come so close often. The simplest strategy for that sort of information is to find a GAT with the bare minimum of numbers picked, which demonstrates quite often that 1-off achievement, then expand the predicted numbers to a set containing all the +-1 too and form a larger pool to play. So, if your initial GAT predicts 5 numbers, expanding them can produce a 10-15 number set. Given this 1-off occurrence, it is possible for this set to contain even all of the correct numbers. It is just another way to look and use the predicted numbers. You can easily ignore that yellow colouring as well and concentrate solely on GATs that deliver exact hits.

Jmichael
Newbie
Newbie
Posts: 3
Joined: Thu Jul 10, 2014 9:01 pm

Re: Computer Generated Drawings vs Physical Ball Drawings vs

Post by Jmichael » Mon Jul 21, 2014 8:27 pm

Ok, got it...
Well, I suppose though if a certain GAT looks good in the past because the program has shown that it can pick, for example, 4 or 5 hits fairly regularly over the course of the 100 draws displayed on the graph, RNG or not...., it should continue to do so in the near future; at least that's my take on it. If I'm wrong on this, please let me know. :)

Thanks Anastasios for your time and all of the info !
John

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

Re: Computer Generated Drawings vs Physical Ball Drawings vs

Post by lottoarchitect » Mon Jul 21, 2014 9:30 pm

It doesn't work like that, you try to equal a RNG system to a physical device. They are really far apart to the properties they hold. A physical device has some properties not found in an RNG system (in the general manner); the most important is the initial conditions. These properties are the main reason for reduced randomness in a system. RNG systems generally do not have initial conditions the same - not sure if we can actually define initial conditions here.
Now, let's consider a simple example, a sequence of numbers which really looks random to our eyes, thus by normal observation we have no clue what is the next value in that sequence. It is possible to make an equation (a time series function) that has as input a known part of that sequence and if there is really a tight correlation among that sequence, to properly come close to the next value - this function predicts the next value. If the same sequence comes from a real RNG system however, whatever the outcome of the function we have computed is irrelevant really - we can't predict a true RNG system. In the middle, there is the reduced randomness behavior where we have short-term dynamics affecting the results. Such is a device like a real barrel made to draw lottery numbers; the primary reason for reduced randomness is the way the balls enter the barrel. It does happen over a short period of draws some conditions to remain relatively stable and this is what GAT tries to take advantage of to formulate a prediction.
Now, what makes a huge difference here to the performance illustrated by any GAT for the "validity" in the future unknown draws is IF the system actually exhibits reduced randomness behavior. Consider what GAT displays to you at the graph as the outcome of the time-series function we computed above (what GAT does isn't a time-series analysis, I use the time-series as an example to understand this). What I mean by that is, a GAT can display some good performance over that tested range. If the analysed data is in fact a completely random system (like a real RNG mechanism), there is absolutely no reason to expect a similar performance at the future, similarly to a time-series function that indeed produces an irrelevant prediction even if it can accurately predict all the initial data used. It is not a failure of the system itself; it is just impossible to predict pure real random events. However, a lottery draw in a barrel experiences reduced randomness behavior thus in this case what the graph illustrates is an overall idea of what can be expected at the future unknown draws too. Generally, a RNG system lacks many (if not all) of the concealed properties of draws performed in a barrel that contribute to the reduced randomness of that system. A lesser quality RNG system may also have some properties that can be found by GAT and in that case you may experience some good hits at the (yet) unkown future draws. But a RNG draw vs a draw made in a barrel are completely different entities with completely different properties.
Finally, I have read over the years many attempts to analyse draws using RNG systems and make conclusions about lottery games. What all these people FAIL to understand is that an RNG system and a lottery made in a barrel are completely irrelevant entities; they have nothing in common. Bottom line is, GAT can be used for draws in a barrel obviously. Not much to be expected however in draws made by RNG systems; if the RNG used is of bad quality it is possible for GAT to find there something to use but in the general case a lower hits-wise quality should be expected maybe close to what natural odds suggest. Thus it makes huge difference if the draws are made in a barrel vs RNG.

Jmichael
Newbie
Newbie
Posts: 3
Joined: Thu Jul 10, 2014 9:01 pm

Re: Computer Generated Drawings vs Physical Ball Drawings vs

Post by Jmichael » Mon Jul 21, 2014 9:55 pm

LOL, well,Ouch then.... LA... that's really depressing news, as all of the lotto drawings held in my state are definitely RNG. The only type of drawing available that is "barrel drawn" are the multi-state lotteries (such as Powerball, MegaMillions ) that use the barrel method, but their odds are so high, I'm not sure how much GAT can help with those...., but I suppose I can run a bit of back-testing and see what GAT can do with them. :shock: :cry:

Again, many thanks for your time and detailed explanation.

Post Reply

Who is online

Users browsing this forum: No registered users and 27 guests