Should We Take Into Consideration Lottery Pretest Drawings?

I. Fair Shake: Pretest Lottery Draws
II. Lottery Test Drawings Do NOT Affect Gambling Strategies
III. Lottery Test Draws Can Affect Number Patterns
IV. Inclusion or Exclusion of Test Drawings Keep Lottery Strategies Intact
V. I Want ALL Lottery Draws, Including Fake/Test Ones
VI. Lottery Strategies Are Always the Same, Despite Test Draws

I. Fair Shake: Pretest Lottery Draws

Hello my fellow lottery analysts!

Please read this page (especially starting with page 3) concerning auditing procedures during the California Super Lotto draw. Search for california superlotto automated draw.

My question is: Should we take pretest draws into consideration?

After all, just because they are not televised, they still happen. How would the data file look like if we introduced two random draws in it before we input the real draw that resulted? (It would be great to possess 2 G-tech machines to simulate these lotto draws). Alas! If not, we may resort to code.

I would simply run a simple random routine with the TIMER$ set at some time that I think they would be conducting their pretest draws. I would introduce maybe a routine to produce numbers each nanosecond for 5 seconds, then output 2 draws as the pretest combinations. Why do I want the time to be approximately the same? Because as the universe spins around, where we are at and when events take place may be influenced as such — the time and place we are in.

At the same time, would I introduce a sleep statement between the generation of each number to simulate the gravity-lotto machine putting the balls down the chute? (As we see in the real draws, there is a tiny lag to consider between when the balls come out.)

Ion, I observed that you do have a sleep statement in the RANDOM.BAS program, yet I don't think it is optimized for the actual time that the lag happens in the real draws. It may. I don't know. In the real draw, the balls are always ordered before being released, and two paddles shuffle them around. How would we simulate the paddles action in code? What about the interaction of the balls between each other (friction)?

Now, since my SuperLotto data file is already in existence as published on the lottery site, I would have to introduce these random numbers in it and check to see if it shows any interesting stats. If I already have 500 combinations, I would run the random routine as described above, and after the 5 second shuffle, I would instead output 1000 combinations — that I would later take and insert 2 at a time between the lines of my real draw data file. Then check the stats.

In the real draw, the audit report states that the machine possesses a microprocessor which is set [as far as in the amount of time the balls are shuffled before being released]...which time may be chosen by a person at the lottery, and which may be different from state to state (depending on each lottery's decision).

Then, who releases the balls? The machine or the operator? If it's the machine, that may be better for us. It it's the operator, any second lag could mean an interference in our results. My opinion...

That might be an easy answer if we start a computer timer at the time of the real draw and mark the start and occurrence of each ball, for a few draws and see how the time fares out. Of course, we might get close, but not in millisecond time; our reaction is far less optimal. Maybe a video capture of the televised draw, with a piece of software to analyze it might be more of a solution to the timing factor.

One might say — why the big fanfare about THIS TIME!? Well, like I said — I believe in the forces of the universe and each millisecond of it depending where we are at, or where anything is, matters to some extent.

So, if I simulate random draws by using the computer time at the time I am running my program, why not use a TIME that I am more inclined to use — like the time when I think they may be conducting the pretest draws, or for that matter, the real draw (which is known and can be simulated down between observed instances)

Camelia Sacui (a.k.a. mindtab)

II. Lottery Test Drawings Do NOT Affect Gambling Strategies

Like in the ol' days, fata mea! It's like going back in time to 2001, I guess, when you started posting on my old bulletin board (Saliu.com/bbs/). That message board is still there… I mean, the posts that I converted to Web pages… posting is no longer available…

You touch a very interesting aspect of lottery, randomness in general. I talked about it sporadically at my site. My lottery software ignores the test (fake) lottery drawings conducted by the commissions prior to the real (official) drawing.

In the end, it doesn't really matter in the lottery if they skip drawings. Condition: As long as the test drawings are fairly random. That is, the commissions perform a numbers of tests, normally 3, or 4, or 5 fake draws. They don't look at the tests to analyze them based on some criteria (such as avoiding combinations that would lead to very large payouts, especially in pick games). If the lottery commissions do not care at all about the test results, I accept that such drawings are fairly random. Most importantly, they must never change the drawing set to be official. They must follow a rule like this one:

"We shall conduct four test drawings today. The draw following the fourth test shall be the official lottery drawing."

Ideally, the random events must be serial, not parallel. I am at the same roulette table. I don't care about what happened at any other roulette table. I only record the spins, as they occur, at this table. I bet based on my systems and the spins I recorded. The casino dealer might as well say: "We conduct two test spins before the official spin." I don't care to record the test spins. The only requirement: Fairness in spinning the wheel. It's not acceptable for the casino to conduct another test just because a certain roulette number was drawn.

This is a curious coincidence. I had a nice experience at the horse track. I turned time back more than a dozen years. I had to delay a trip and other personal businesses. So, I went to an off-track wagering facility yesterday. Just for fun, not to strike it big. I applied again my very first horseracing betting system I first created in the year of grace 1995. The system was inspired by my pick-3 lottery theory. I had started to write a new page for my website. I took a short break, and then I checked this message board. I saw this thread!

My recent experience in horse racing proved to me that the sequence of random results has a strong influence over betting. I recorded results of horse races (trifectas especially) as they occurred, sequentially. As I said, I had a nice evening! See section I of the horse racing betting systems page at my site.

Again, just briefly now: It is safe to ignore the test drawings of the lottery commissions. The lottery results, with or without the test/fake draws, follow one and only one law: Fundamental Formula of Gambling (FFG). The law does not apply if only and only the test drawings are manipulated. But randomness is based on the exclusion of manipulation. That's why it is safe to randomly add results from different lotto games to your SIMulated lotto file. But I wouldn't add such data to my data file of real lotto drawings.

Folks, you heard it right from the horse's mouth!

Bafta and best of luck to all!

Ion Saliu,
Rider At-Large
Perfect record for my horse racing software at Belmont Stakes 2001

III. Lottery Test Draws Can Affect Number Patterns

Yah, time back to 2001! Seems like yesterday; maybe just like the horse-racing day you had. Google caches lots of pages, so they may still be out there.

I get what you are saying. Any new draw has the same fair shake as any other — from what we understand what randomness means. I wonder how that would change our ideas if we suddenly found out that we live in a quantum world, and nothing is really as we understand it. What about the butterfly effect? What about golden ratios, and Fibonacci, oh, if I only understood half of it.

But I like patterns, and I like to think they can be twisted out of their spiral of randomness toward our favor. So, we put different weights on things, and hope they do so. I think I went off the path, ha-ha. It is rather that we don't do anything but try to figure out which way the spiral is going, and we bend accordingly into their favor in order to comprehend and absorb a drop of understanding from it.

I don't want to get off tangent here, merely a metaphorical thought of sorts. But to me, any idea is fine, as long as it can be tried, regardless of reason. Like they say in Romanian: “Incercarea moarte n-are!”

Again, I respect your thought, and I understand fully what you are telling me.

Camelia Sacui

IV. Inclusion Or Exclusion of Test Drawings Keep Lottery Strategies Intact

The inclusion or the exclusion of test drawings from a data file has NO effect in applying theory of probability. The condition: The entire process must be conducted randomly. That is, the operators have no intervention whatsoever in the process.

Here is what I did several years ago. There was a filter in my pick-3 software named TOT. Actually, it was about the straight pick3 set. The filter checked when the same straight set hit in the past drawings. For example, if the last pick-3 drawing was 1,2,3 the TOT filter showed a value like 693; that is, 123 hit last time 694 draws back. The skip or gap between hits was 693 drawings. (Actually, 693 is the median skip for that type of pick lottery filter, TOT.)

The TOT filter checked the past drawings one by one. I also checked the past drawings from seven to seven, as if to mimic checking only the same day of the week. I just added a STEP statement in my software. If checking each and every drawing, the default STEP is 1 (I don't need to write it in the software). If I check the past drawings from seven to seven, then the STEP is 7:

FOR IndexCount = 1 TO AllDrawings STEP 7

The results for the TOT filter were very similar, with STEP 1 and STEP 7. The FFG median was the same: around 692 or 693. The only difference was in timing. For example, last TOT with STEP 1 was 0, while last TOT with STEP 7 was 193.

I also made the STEP 5 and STEP 10. The result was the same median skip. The only difference was in timing. So, strategies based on TOT with various STEPs varied only in timing: They hit at different times in the data file. The strategies had the same frequency and the same median skip.

So, if they conduct 3 test drawings, it's like generating the filter reports (W files) with STEP 3; if they conduct 4 test drawings, it's like generating the filter reports (W files) with STEP 4; etc. The strategies hit with the same frequency, regardless of STEP. But the strategies will hit at different points in the data file.

That's why the change in the drawing machines has no influence, except for timing. Actually, I urge them to change the lottery drawing machines frequently. The machines are mechanical, therefore prone to wear and tear. That creates biases in favor of some numbers or patterns. I haven't noticed that in lottery drawings. But I did notice a strong bias at a roulette table, in a casino in Atlantic City. There was clearly wear and tear in the wheel. The bias was way outside the standard deviation! One female dealer was substituted by a male dealer: It didn't make any difference.

You know what is the most daunting hurdle in lottery strategies? PATIENCE! I know it firsthand. (I also have the excuse of spending a lot of time in developing theories and software.)

Just stick with a reasonable number of lottery strategies. Check them in the past (with the 'Check Strategy' functions). Preferable to play a strategy if the last skip (current skip) is under its FFG median. Also, if the strategy skipped a large number of drawings; yes, there is such a thing as an overdue strategy. There will be some long skips; but the majority of the skips are there and they are shorter!

Parpaluck

V. Fair Shake: I Want ALL Lottery Draws, Including Fake/Test Ones

Ion,

I am probably too headstrong, or stubborn in ideas...and I don't mean any disrespect toward you or your thoughts whatsoever. Here I go again.

Now, if I have my data file with the real draw (thus am dismissing the pretest draws as being real) and decide to take out one draw. It shouldn't matter, right(?), since each draw is independent by its own randomness, whether it's real or a pretest. Ok, well, I decide to take out the next one, or, why not all? I can make a new data file with just random numbers. But then I'm not looking at what has been actually drawn.

What was actually drawn? The televised draw?

Maybe I am 'stuck' in the area of the randomness of draws. In this case, I reckon, I wasn't necessarily talking about how random each draw is, although it looks like it.

I am talking about a series of events. A draw==an event. How each event has effect on the next event, even though each event is individual and has its own 'random' outcome (the discussion at hand).

It is just as much saying that if a chicken lays 3 eggs and I took one and ate it, then I can say the chicken only laid 1 egg. Now each egg may taste different, it was laid at different times and has no effect on any other egg being laid. But yet, and still, 3 eggs were laid that day and I only think it's 1, because I only ate one of them.

So one can say now — so what? There were 3 eggs, you ate 1, and what? Well now I can go back and delete my whole data file and start from scratch. Just kidding. Well, if I ate 3 eggs, I would have tasted 3 different eggs and I would be fuller. My perception might be better that day of them eggs. I won't talk about this anymore. Probably better.

The day is nice, and I am juggling numbers here, the wind outside not as furious as yesterday has been. It is mellower and a bit chilly, but sunny between the clouds that want to rain — but won't.

Quote:

"For example, if the last pick-3 drawing was 1,2,3 the TOT filter showed a value like 693; that is, 123 hit last time 694 draws back. The skip or gap between hits was 693 drawings. (Actually, 693 is the median skip for that type of filter, TOT.)"

Your amazing FFG formula, axiomatic one, seems to follow dates (in the median sense)!? Or do dates follow the FFG? Looking at your step 'to mimic day of the week'. Maybe it's different in different years. What if we assigned days of the weeks a number, as it relates to the calendar across the years, such as:

jan-oct =0=sunday
may =1=monday
aug =2=tuesday
feb-mar-nov=3=wednesday
june =4=thursday
sept-dec=5=friday
apr-jul=6=saturday

The above is from a reference scale I used in a program to get the day of the week input. Mind you, back then I didn't know what the internet was (1995), was young and dumb (hmm) and I figured it out by numerous scribblings and calculations --from a handout on the Perpetual Calendar given to us by our QBasic teacher (the project was Julian Date Calendar).

Why did I mention this? From your mention of the different step in the TOT filter. Although different figures in TOT, the result may be similar in correlation to a predetermined set (such as I have above)

Quote:

"The strategies hit with the same frequency, regardless of STEP. But the strategies will hit at different points in the data file."

The step as a timing factor with the frequency staying the same is great and better news. Now maybe what is left is putting those steps in step with each other.

Yet, we are probably looking at it in different ways (and maybe the same)

I do keep up with the lotto strategies lately. It has been a while though, due to lack of any money. I kept up with things for a while, then not being able to play was like going to Frys (computer related store) and slobbering over the things I wanted and couldn't buy. So I didn't go there anymore.

Giggling in pain, I guess, right now. Could have done a lot of things, and observed much more.

Camelia Sacui

VI. Lottery Strategies Are Always the Same, Despite Test Draws

It's not about being stubborn, axiomatico! It is about a healthy form of debating based on critical thinking.

Now, if I have my data file with the real draw (thus am dismissing the pretest draws as being real) and decide to take out one draw. It shouldn't matter, right(?), since each draw is independent by its own randomness, whether it's real or a pretest. OK, well, I decide to take out the next one, or, why not all? I can make a new data file with just random numbers. But then I'm not looking at what has been actually drawn.

What was actually drawn? The televised draw?

YES! You can check any strategy against any data file. That includes data files that consist entirely of random combinations (such as SIM files). The strategy will show the same frequency and the same FFG median regardless of data file. The only difference will be in the string of 'strategy hits'. For the sake of simplifying only. The same strategy will show 'hits' strings like these for data1 and data2:

3, 6, 2, 1, 1…
2, 1, 3, 5, 6…

The skip median will be the same for longer strings.

You need to check for data files of real drawings because of the last (most recent) values in the hit strings. In the examples above, I would prefer data2 if the skip median was 2. Data1 shows 3 and 6 as the latest skips; the skip can go higher than 3.

I used to keep records for PA and MD lotteries. I weighted the strategies on the two state lotteries based on the most recent hits for a strategy. Sometimes I would play the strategy in both states; other times, I wouldn't play the strategy in either state.

I had that similarity with horseracing. I thought I would write about it yesterday. I don't know why this type of coincidence happens. Usually, this place goes quiet for days, even weeks. Now, when I wanted to write something, I feel like I have to respond to posts. But this my last post before I finish that horseracing stuff.

So, I recorded the results of all races from the previous day. I would go to equibase.com. I had no choice but record all the results by horse track. It is mission impossible to write the results in sequential order. TIME CONSUMING AND NERVE WRECKING!

I record data for several days. I get thousands of real results. I also have a SIM data file. I only look for races with at least 9 horses. I care only about trifectas that have not come out in a long time. All trifectas will come out in a reasonable time. But I only care to play trifectas that pay big. So, long shots have to be part of the trifectas I play.

This type of betting does not take into consideration the latest results in any particular horse track. I play my trifectas in any horse race of the day in U.S. and Canada. Sometimes it takes several days until this strategy hits a huge trifecta. In other situations, this strategy hits multiple high-paying trifectas in one day. The trifectas used to be much bigger before I wrote about this strategy in newsgroups. This horseracing betting method requires computer analysis and a sizable bankroll (to withstand losing days). The payouts can be big, however. I heard a gentleman on the radio, after I made this strategy public, who proved he was making an average profit of $100,000 a week. The profits are lower now because a number of horseracing bettors play that way.

So, this method is almost like your situation of disregarding the real drawings in a particular lottery.

The method I used the other day was actually my first horseracing betting system (applied first in 1995). I precisely record the results, as they occur, at three horse tracks (determined by the TV monitors in the facility). I will eliminate horse numbers in the same position from the recent results of the three tracks I… track. Actually, last time I could only track two tracks. It was Wednesday, and things are much slower these days.

Race #9 at Belmont. The result at the previous race was 7-2-3. #7 was a favorite in this race #9. I came up with two boxed trifectas: 1-4-9 and 5-7-9. The result was 4-7-1. Had I wheeled the 5 unique numbers (1,4,5,7,9), I would have played 10 boxed trifectas. 4-7-1 paid over $1000.

I still used that data plus data for the only available track remaining (Indiana). I had to wait a couple of races for more results. When waiting, it is a good idea to play the three longest shots in a race. When you hit, you get some money back. You can add to the bankroll. Anyway, the first race at Indiana (that I didn't play) had the trifecta paying over $3000.

I attracted some attention, because I only wrote down numbers. I told someone that I don't care about the horse name, or the jockey. For race #4, I came up with three boxed trifectas: 2-6-9, 7-8-9 and 6-7-11. The latter hit — after an objection delay! It wasn't big — a few hundred bucks. 2-6-9 would have been huge… it occurred for a short time in the middle of the race! I left for home immediately.

In this second betting method, the real results at a particular race do count. I want to win the same day. The strategy is fine-tuned. It is also the method to play the lottery. Track just one state lottery, or two or three — but no more.

The first horse-racing method is better suited for professional gamblers. They have a good bankroll and they dedicate all their time to horse betting. You get the results and analyze them with the computer in the morning. In the afternoon and evening, you go to a track and play diligently and patiently across all available horse tracks.

Ion Saliu

• Computerized Lottery Drawings: A Reason for Concern?
• Skip Systems Software: Lottery, Powerball, Mega Millions, Horse Bet.
• Random Numbers Generator, Algorithms, Randomize, Software Source Code.
• Randomness, Degree of Randomness, Degree of Certainty, True Random Numbers Generators.

Back to Forums Index        Socrates Home  Search