download build 595

[stan]

the stats tables and charts are extended in this new build. but it's not finished yet, some things may not work properly.

(the pairs and triplets tables will be available in the next build)

[stan]

* added table and chart for number groups stats
* added charts for ticket positions and first/last digit positions
* the stats page ui is a bit reorganized
* fixed some bugs

(the pairs and triplets will be in the next build )

[stan]

* added pairs table
* added triplets table
* added latest draws to number matrix and inverted matrix that i removed in previous builds


as you will no doubt notice the triplets table is not very good, it takes very long to sort the table by any column. i'll have to come up with something better.

[stan]

the latest build is 623 now - there was a bug in the triplets table

[stan]

* fixed the sorting problem of large stats tables (triplets) - a sudden flash of inspiration

[stan]

* added a new successive triplets table
* 'neighbours' moved to 'successive numbers' page
* fixed some minor bugs

[hexagonvex]

Looking good Stan.

For the number groups statistics (you know me and my number groups ), would there be a way to have an overview tab which would show the following:

winning numbers coming from 1 group: X times (col.1) Y%(col.2)

for example for the german 6/49 lottery the overview would look like this

from 1 group 0 mal 0%
from 2 groups 15 mal 0,34%
from 3 groups 510 mal 11,67%
from 4 groups 1929 mal 44,15%
from 5 groups 1649 mal 37,74%
from 6 groups 266 mal 6,09%

this is a rather interesting one for gameplay. Player can decide if he wants to check that numbers fall into 3 to 5 or 3 to 6 groups in the number groups filter.


I got another question...maybe this one is for Mr. Hakl...?!

With the Best tickets filter...I think you mentioned that this filter was for maximizing the winning of minor prices. I used it on a package containing 100.000 tickets. I asked for the 10 best tickets of a pick 5 game.

The best tickets filter came up with something like this (it's just an example, I don't remember the numbers now):

2,10,12,23,45
2,10,12,23,38
2,10,12,25,45
4,10,12,38,48...

I don't know if the combination 2,10,12,23 should appear 2 times in the result set especially if my input file contains 100.000 tickets.

Maybe an option for the maximum overlap of numbers (like in the package generator) would be an idea. I know this is going to slow down the filter....but for MAXIMIZING the winning of minor prices I wouldn't play 4 out of 5 numbers twice, would I?

Greetings

[stan]

[quote=hexagonvex]
For the number groups statistics (you know me and my number groups ), would there be a way to have an overview tab which would show the following:

winning numbers coming from 1 group: X times (col.1) Y%(col.2)

for example for the german 6/49 lottery the overview would look like this

from 1 group 0 mal 0%
from 2 groups 15 mal 0,34%
from 3 groups 510 mal 11,67%
from 4 groups 1929 mal 44,15%
from 5 groups 1649 mal 37,74%
from 6 groups 266 mal 6,09%

this is a rather interesting one for gameplay. Player can decide if he wants to check that numbers fall into 3 to 5 or 3 to 6 groups in the number groups filter.
[/quote]
yes, this is no problem (i was somehow expecting a feature request like this )

I got another question...maybe this one is for Mr. Hakl...?!

With the Best tickets filter...I think you mentioned that this filter was for maximizing the winning of minor prices. I used it on a package containing 100.000 tickets. I asked for the 10 best tickets of a pick 5 game.

The best tickets filter came up with something like this (it's just an example, I don't remember the numbers now):

2,10,12,23,45
2,10,12,23,38
2,10,12,25,45
4,10,12,38,48...

I don't know if the combination 2,10,12,23 should appear 2 times in the result set especially if my input file contains 100.000 tickets.

Maybe an option for the maximum overlap of numbers (like in the package generator) would be an idea. I know this is going to slow down the filter....but for MAXIMIZING the winning of minor prices I wouldn't play 4 out of 5 numbers twice, would I?

well, that's how the filter works. my guess is that your package before processing had a lot of "2,10,12,23" combinations and the filter just reflects that.
besides you can always use 'match package' filter first and then apply the best tickets.

[hexagonvex]

Hi Stan,

You start to know me well :-p

But for the best tickets filter....I'm not sure but I don't quite agree on that one. I'll double check what you said over the weekend...with before and after...and with the stats page as well....and I might get back to you on that one.

Because even if I have many 2,10,12,23 combinations for example in a pick 50 game there are only 50 cobinations for the 5th number. So if the package has 100.000 tickets I still wouldn't see why this combination would come up so many times. To me the best tickets filter would have to be more biased to maximize minor winnings....that's what it's supposed to do.

Assume you play 2,10,12,23,45. Then you filter out with Match numbers filter Match5, Match4, Match3....there will be no more than a few thousand...less than 10000 probably filtered out...so what about the other 90.000....there will be some minor prizes in it.... It's just some thinking out loud that I do here....don't get me wrong please

Greetings

P.S. What do you mean by applying "Match Package Filter" first....which package(s)?

[stan]

the best tickets filter select the first xx tickets that have the most matches with the rest of the package (highest numbers in the 'match summary' table from visual package page) - no less no more.
another solution might be to select 1000 best tickets (it won't slow the filter down) and then filter the results for matching combinations.

but do try checking your package contents before and after applying the best tickets filter. i'm interested to find out if i'm right on this one or not.

[stan]

[quote=hexagonvex]
P.S. What do you mean by applying "Match Package Filter" first....which package(s)?
[/quote]
i meant filtering out combinations that match each other in four or more numbers first and then applying the best tickets filter.

[stan]

* new group count table in statistics
* the (long) list of installed lotteries in the main menu is broken up into smaller submenus

[Bobes]

Hi all,

let me provide some comments to your problem.

A) In fact the Best ticket filter can filter out groups of tickets. It happens when the starting package is too large or when the first selection was not accurate enough (you chose intervals close to Min/Max margins). It can be resolved in many ways. Try to use subsequently any other parameter. It is recommended to use any parameter that apply in general (O/E, H/L, SUM BASE etc.). The best seems to be SUM ROOT because of almost uniform distirbution.
Then you can select up to three tickets from such group. I feel that the Match Package filter recommended by Stan is not suitable because it prefers ticket's first occurrence. Naturally you can use the Package Statistics Page, which provides you an extensive set of many supporting parameters. I recommend to use SUM ROOT due to the fact that when my estimation is not correct i can still keep high chance to win higher prize (usually match in 5) while the package contents reduces by factor of 9. Apparently when the large starting package (50k of tickets) comprises the Jackpot winning numbers (match in 6 = 1x, match in 5 = 179x, match in 4 = 2158..) then each of the reduced package (for example filter for SUM ROOT=3, while the correct SUM ROOT value = 4) will contain almost the same count of tickets matching in 5.
B) Usually such group of tickets is in one location on the screen after "Best 50 tickets" selection (Visual Package - tickets table with tickets) thus you can see columns creating continuous graphic on the screen. So you can highlight these (for example 4) columns and click on the "Total Match" button to see appropriate table. When you have more such groups it is either due to large package (not too strong condition for selection) or due to uncorrect margins. It is clear that only one ticket can be the Jackpot winning ticket thus finally you should chose up to three tickets from each group to keep your chances at least for small wins.
C) For those who prefer their gut feelings in decision making process it is recommended to use function "Use Package tickets" in bottom part of the WN History Page. Then they can simulate tickets from package step by step using all available support provided by the Differences History Tables, Graphs etc. Then they can decide based on their visually supported feelings, however this procedure could be too subjective. In fact also this procedure may comprise features which can be almost generally applied. For example removal of all tickets comprising all 6 (6 is an example; you can chose 5 or even 4 which are more stringent conditions) numbers from the bottom part of each table. You can also take into account that 4 numbers from one row is rather rare case etc.. In addition you can use the StDev graph (Graphs, Deviation, last row in dialog box and Table of sums, StDev Diffs column). When this StDev should be large it means either that you should favor rather 3 to 5 columns = +/- 10 than usually apllied condition 4 to 6=+/- 10 or more than one large deviation (WN History filter) can be expected for the next draw. Thus when you started with selection for +/-10 in at least 4 to 7 you should carry out this selection again rather for at least 3 to 5.
D) In case you get more than one group of tickets in your Best xx tickets selection the best procedure could be to take all numbers from these groups and generate new full wheel from such limited set (reduced pool of all available numbers) with subsequent filtering using the same conditions for filtration. You can also use Table "Simulation" (WN HIstory Page, click on the ">" button in SImulation Tickets field and chose "Package") to simulate full wheel. Then you can use values in the column AVG as the starting vaules for MIN, MAX margins in the WN History filter. If you guess the reduced set (created for example from 25 to 32 numbers) in correct way (it comprises all 6 numbers) and the rest is from other groups after the selection of the Best xx (it contains all 6 numbers) tickets in the first selection then you can use margins from +/-5 to +/-7. Usually from 8 to 11 columns match these conditions even if you perform new selection for full template for all possible numbers (49 in play 6/49) (recommended). Then you get smaller package where you can select the Best 10 for example (in order to avoid groups of similar tickets).
E) It is also possible to divide large package to smaller parts (use the SUM ROOT filter). Then select for example the Best 30 from each group. Gather all groups with the Best 30 to one package and carry out the Best 10 selection again. When you realize for example the Best 50 tickets selection from large package and subsequently the best 10 selection again your package can be strongly distorted. Dividing the large package to small parts you avoid this strong distortion (however not fully).
F) In fact the main reason in implementing the Best Ticket Filter was to prevent omission of something in large packages focusing on one set only. It can happen in especially large packages resulting from full wheels when numbers are spread through whole set (from 1 to 49) in a near uniform way. In such cases the package is sorted thus you can easy miss something important.

[hexagonvex]

Mr Hakl,

I didn't understand everything you said....at least some I did understand. I will have to re-read this a few more times. I bow down before your statistical knowledge and your thinking. Gee u really got brains!

Stan,

I've double checked what you said by a trial and error session. the filter is working exactly as you said....so I bow down before you as well.

I was too much concentrated on the words "maximize minor winnings" from the package. From there I made a logical misconception about the way it would work and tried to trace it back :rolleyes:

When I play I usually have a starting set of 265.000 ticket out of 2,1 million tickets. then I start filtering

Low/High
Panel
Successive numbers
Sum
Number groups (u know I like that one Stan...)
1st digit of position 1
winning numbers filter
WN Hist filter (segments)
WN Hist filter (sums)

The parameters I choose follow a rather "careful" or defensive elimination strategy because there is always the "random factor". So I end up with 82.000 and less tickets to play. Which is much but that number will decrease over time -> this is due to my strategy

I haven't managed to have 5 of 5 in the package. This is also due to strategy (depends on pairs) but managed to have some 4 of 5 in package....but haven't managed to hit . That's the random factor :finger: hehe

but I have time and patience

Still your software is really the best around. After being on a strict loosing streak for half a year now I manage to hit "something" more or less regularly....always the smallest prizes but still more regularly than ever before

Greetings

[stan]

[quote=hexagonvex]
I didn't understand everything you said....at least some I did understand. I will have to re-read this a few more times.
[/quote]
well, his english isn't best but he's got years of experience. so don't be afraid to ask

I've double checked what you said by a trial and error session. the filter is working exactly as you said....so I bow down before you as well.

no need to bow, we're all on the same ship heading towards the jackpot island