How Many Drawings in a Row?

Reply #14

LottoBoner — Fri, 15 Nov 2013 03:29:27 GMT

It depends on the variable you are tracking AND the game you are playing.

Reply #13

Lottonomics — Fri, 15 Nov 2013 03:24:29 GMT

1. I know the event is random. All of the methods I test depend on the event being random first and foremost.

2. The something could be any desired outcome you are persuing. For example, you started a thread with a formula to generate number groups where you said this outcome gives constant 3/5 for lotto games. In your case, the desired outcome of your formula would be 3/5 for lotto games. It's a tangible outcome that you can back test, and keep track of the success going forward.

3. With... [ More ]

Reply #12

adobea78 — Fri, 15 Nov 2013 00:43:37 GMT

1. The event is RANDOM

2.'How many drawings in a row would SOMETHING', explain SOMETHING

3. Since the event is random, consistency is a bit of a stretch for any system.

4.Focus more on Prize RATIO (I consider 30% hit ratio positive, odds is not probability).

5.Large picks may mean high hit ratio, but is not cost effective.

Reply #11

Lottonomics — Thu, 14 Nov 2013 23:43:22 GMT

Well, it's time to try out the new idea. It's for California Super Lotto Plus, but can be adapted to any game really. It has passed my testing guidelines. A successful 25 consecutive draws in back testing, an additional 10 arbitrarily chosen back tested draws, and 10 consecutive forward tested draws. So 45 succesful trials in total.

You would think with all of those successful trials it will eventually be successful at least once more, right? At least that is my hope. I only need one. We will... [ More ]

Reply #10

Lottonomics — Mon, 21 Oct 2013 00:14:57 GMT

Dangit, I had typed up a lengthy more detailed explanation of some things, and I accidentally hit the back button and it got lost. Sorry guys Maybe it was a sign that I should keep my mouth shut :P Thanks for everyone's comments and suggestions though!

P.S I'm not playing the new mega millions.

Reply #9

str8ca$hhomie — Sun, 20 Oct 2013 22:00:30 GMT

For what it's worth and this is just my conclusion is that NO system's player will be able to get an accurate reading on the behavorial patterns of the new MEGA MILLIONS lottery numbers until each and every single number 1-75 has had a hit at least once.

You could be talking one (1) year; two (2) years; or perhaps maybe even more for this event to happen.

Reply #8

phileight — Sun, 20 Oct 2013 09:12:11 GMT

Lottonomics

over the years I have come up with a couple of systems that had super hit rates in back testing but fell on thier face when played. in both cases when I started digging into why, I found that in my backtesting, I was unintentionally including the next draw in the data being tested.

easy test for this is to look at suggested #'s for the next draw, then after entering the next draw, look at suggested #'s for that draw again, they should be the same. good hunting

Reply #7

LottoMetro — Sun, 20 Oct 2013 03:55:18 GMT

Looks like you've got your answer, but I'll just add my comments.

The way to test would be to determine the deviation of empirical occurences versus theoretical occurences. I've been doing this with my state's jackpot games. For instance, if a certain group of numbers (i.e. even) occurs X% of the time theoretically (based on total available combinations), then empirically (in real life) the % of occurrences should be very close. If the deviation is substantial, say a group occurs twice the pe... [ More ]

Reply #6

Lottonomics — Sun, 20 Oct 2013 01:42:11 GMT

Thanks! Good helpful info. I think I know what you mean by can be made more predictable thru tracking/charting. Some of my components I like to keep track of specifically for the days when they don't hit. Because of the nature of how the component behaves, I know it's a pretty much guaranteed success for the next draw.

You've been very helpful. Now to pick a new game, and back to the drawing board for me

Reply #5

grwurston — Sat, 19 Oct 2013 23:50:56 GMT

I only play the p3 p4 so when I do a backtest, I'll start with a month to see if it shows a lttle promise. Then I will go to 3 or 4

months just to get a longer term view and to make sure the first month wasn't just a short term hot or cold streak. Whether I

go to 6 months or a year depends on how ambituous I feel about it. lol But I think 3-6 should give you a pretty good idea

over a longer time frame. This will also give you a good idea of it's performance, ie how often it hits. To see... [ More ]

Reply #4

Lottonomics — Sat, 19 Oct 2013 23:24:47 GMT

Thanks a bunch! This is the type of answer I was hoping for. When you say over time and a certain % in backtests, what time frame and how many backtests? Over the course of the entire game history, 100 draws? Less? Somewhere in between?

Also, when you do find something that seems to work consistently in backtests, have you ever come across a situation where it immediately stops working the moment you start using it?

Basically, here's the story. I had been frantically trying to finish some... [ More ]

Reply #3

grwurston — Sat, 19 Oct 2013 20:25:53 GMT

Good question. I think that would depend a lot on the individual's expectations for their particular system or method and

how much success they have had in the past. If an individual hardly ever hits anything, and they come up with something

that hits say, 35% of the time, they may consider that to be very good, whereas someone that has something that hits 50%

may not thinks it's so good.

Personally, I am always looking for something that hits consistently over time. I am not interes... [ More ]

Reply #2

Lottonomics — Sat, 19 Oct 2013 07:58:57 GMT

Hmm, I guess I should have been a little more specific. By something I didn't mean a specific ball/number, more like an event for prediction (I don't know what else to call it).

Example: A group of numbers for elimination. To ensure consistency, the method of number selection stays the same throughout the testing phase (but the actual numbers produced by the method vary from draw to draw). The prediction event of importance would be elimination. If none of the numbers hit, it is a success, if... [ More ]

Reply #1

LottoMetro — Sat, 19 Oct 2013 06:00:33 GMT

That would depend on how many balls are in the pool and how many are drawn. I use statistical tests to determine significance, and you would be surprised how many times a number can repeat/out before it is considered non-random. It's usually a large number.

Example:

1000 draws for 1 of 10 (when I say 1 of 10, I mean i.e. the first position in Pick 3)

If a ball appears =140 times or = 60 times it is pretty close to significant. Any number of times in between and it's normal!

Note that... [ More ]

How Many Drawings in a Row?

Lottonomics — Sat, 19 Oct 2013 05:44:48 GMT

A question for those of you that like to come up with your own systems. How many drawings in a row would something have to occur before you would start to think it might not be a fluke/coincidence