Making Estimates

[PadawanLotto]

I'm assuming that it is possible to make sum estimates that will filter a full package down to 50 or less combinations but still not be within the correct ranges to produce a jackpot win.

[stan]

yes, that's correct. you can even filter the whole package down to one ticket...

[PadawanLotto]

Using the Ohio Rolling Cash 5's last 40 drawings, I was trying to come up with a good way to make estimates for the WN History filter. I went to the WN History page and the Summary tab, updated the table and opened Differences Only, I then clicked on Export and saved as a .txt. I placed the Differences table into MS Excel and using conditional format I highlighted the +/- values in different colors. The ranges I used were 0 thru 25 and -1 thru -25 and found that 16 out of the last 40 drawings pasted all 11 columns. I stepped up the range to 0 thru 26 and -1 thru -26 and found that 18 out of the last 40 drawings pasted all 11 columns. I stepped it up one more time to 0 thru 27 and -1 thru -27 and found the 20 out of the 40 last drawings pasted all 11 columns.

I understand that using this kind of range would leave a whole lot of combinations in the package but it's a start. I haven't had time to work with or test these ranges with the WN History filter yet.

[maroc]

Hi PadawanLotto,

This how i do my win estimates.In MS Xls ,i export the wnhisto difference only.Then i test the most commun difference for each column.I com up with this combination that work neary 75 % of drawing.(this is for keno morocco):
Clmn 0 = min -10 max +0 those are what i add to the recent Sum
Clmn -1=min -7 max +3
Clmn -2=min -6 max +4
Clmn -3=min -3 max +7
Clmn -4=min -6 max +4
Clmn -5=min -3 max +7
Clmn -6=min -3 max +7
Clmn -7=min+3 max +13
Clmn -8=min +9 max +19
Clmn -9=min +9 max +19
Clmn -10=min +10 max+20.
if i have left with a lot of tickets to play i use the at least 1 Clmn must pass from 11,or 2 colmn or 3 until ...

[Bobes]

Tools for WN History Filter – Brief Recommendations

The WN History Filter is a tool for finding a set of tickets that includes the Jackpot winning one. The heart of the matter is to use the correct tools (lower and upper ranges) in a correct way to determine at most 11 different intervals for the next expected sum values. When the intervals are set correctly you should obtain reasonably small set of tickets out of the entire wheel of all possible combinations.

Starting Set

The starting set of tickets is limited only by the computer capabilities:

• There are no problems with 6/49 lotteries, the computer can handle the full list of combinations easily.
• In case of Keno lotteries the number of all possible combinations is usually huge (significantly exceeding computer’s capabilities – both storage and processing speed) There you can either filter wheels created from a restricted count of numbers or you can create a wheel from the whole pool of numbers while skipping every Nth combination.


Filter Nature

The filter evaluates the frequency of occurrence of individual numbers in the lottery history up to their 11th occurrence back in the history and it provides assessment of intervals between particular occurrences. Therefore the independent variable is the fixed occurrence while the time interval is a dependent variable in contrary to other filters where the time is independent. The time intervals between individual occurrences back in the lottery history will differ.

Why Just 11 Occurrences Back in the History

This value is chosen as a practical compromise resulting from the consideration of the lowest suitable value from the point of view of statistical significance (if appropriate) and the most suitable choice from the point of view of errors introduced in the estimation of ranges. Therefore this filter is difficult to apply in lotteries without sufficiently long history of winning numbers. Of course you can use several first occurrences only, however if this is the case your estimation must be extremely correct.


How to Determine the Ranges Correctly

The most frequently asked question is which ranges are to be determined, why just +/-10 ranges are to be used for all 11 columns. Naturally these ranges don’t have to be defined just in this way. To correctly understand what ranges are to be used it is necessary to keep in mind the following important principles:


• The combination you are looking for projects itself into the charts and tables in 11 different ways. Therefore there are 11 different points of view when making your estimates.
• The smaller the range size (the difference between upper range value and lower range value) the less tickets will pass the filter
• The ranges don’t have to have the same size for each column (i.e. the expected sum value).
• The difficulty of estimation is not equal for all columns; some of them may be easier to estimate than others.
• It’s better to make your estimates lower than the actual future sum value because having your estimates bigger than the actual value leads to more tickets left after filtering.
• If you estimate an increase for the next sum value you can expected that even a small variation in the upper range margin will lead to a large increase in the count of tickets that will pass the filter.
• It is safe to keep the lower range margin as low as possible as it won’t increase the count of filtered tickets that much. For example if the current sum value is X and you’re expecting a decrease in this column you can set your range estimates to e.g. X-5 to X-60. Even though the range is 55 which is much more than the usually recommended 20 (+10/-10) the amount of filtered tickets won’t change that much because most (approximately 85%) combinations will fall into range X-5 to X-20.
• In a way this filter acts as a tool providing alternative solutions because the filtered combinations can be antagonistic to a certain degree. You should keep this in mind when statistically analyzing the filtered combinations.
• It is suitable to filter package using “sureâ€￾ estimates as a first step only. For example if you are fairly confident that the sum value will decrease in e.g. four columns, it may be useful to filter the combinations according to only these 4 columns first. Then you can use statistics and other tools (Best Tickets filter for example) to find the most frequent triplets and/or the 50 best tickets. This way you can get more accurate hints for further estimations.
• Applying too small ranges may filter out all combinations. In such a case there’s mostly no reason to change the lower range margin, try (minor) increase in the upper range margin instead.


We’re preparing more detailed recommendations and instructions that we’ll post here in our own discussion board during the 1st half of May .

[Red Devil]

I carried out some similar tests to PadawanLotto using the uKLotto results. Using =25 to -25 from each columns latest value and filtering all 11 columns, this left approximately 3,000,000 tickets and when doing a similar check in Excel +25 to -25 would have given 31 jackpots from the summary list of 200.

Similar test using values of +20 to -20 yielded 912,000 tickets and a possible 8 jackpots from the 200 lines in the summary list.

Values of +10 to -10 left 2976 tickets but checking in Excel no jackpots from the summary list.

However, these were values taken from each sum value in the summary list, I should imagine that +10 or -10 from another value would leave a different amount of tickets when filtered on all 11 columns, it obviously won't be the same number of tickets remaining each week.

I read Hakl's post and done some more testing and it is definitely true that a bigger negative bias leaves less tickets than a positive one of the same size.

Gone midnight now, going to bed to think some more :zzz:

[PadawanLotto]

[quote=Red Devil:1146093667]
I carried out some similar tests to PadawanLotto using the uKLotto results. Using =25 to -25 from each columns latest value and filtering all 11 columns, this left approximately 3,000,000 tickets and when doing a similar check in Excel +25 to -25 would have given 31 jackpots from the summary list of 200.

Similar test using values of +20 to -20 yielded 912,000 tickets and a possible 8 jackpots from the 200 lines in the summary list.

Values of +10 to -10 left 2976 tickets but checking in Excel no jackpots from the summary list.

[/quote]

The tests that I did using Excel to find the most common highs and lows were for a starting point only. I am using the Ohio Rolling Cash 5 (5/39)for a learning tool. The average Max up value is +34 but, +34 is an unlikely value as that would be a repeat of all 5 of the last numbers drawn. The Max negetive value would be all 5 of the numbers that have gone the longest period of time without being drawn and that's very unlikely also. Based on the past occurrence history -27 to +27 was the most common occurrence among the 11 historical occurrences.

The tools that are offered in Epert Lotto are great to one who knows how to use them or if you know what you are looking for. In the History Differences the only column that I even understand is the 0 column. After spending years of looking at lottery analysis in one presective and then looking at it in a different prespective as in Expert Lotto takes some time to get use to.

The Simulation page in the WN History is great but only about 70% accurate and trying to determine which predictions of increase/decrease are correct is not easy. Yes there are some columns that are easier than others to determine which direction the sum value will go but others are not so easy. There are times when there is no change in the sum value up or down and the value stays the same in the next drawing.

[stan]

[quote=PadawanLotto]
There are times when there is no change in the sum value up or down and the value stays the same in the next drawing.
[/quote]
yes, we're experiencing the same problems with our lottery - almost as if the lottery commission was looking at the same charts to know which numbers to draw (not)

[Red Devil]

Stan,

Like PadawanLotto, I too cannot understand the History DifferencesTable. "History to Show 0" yes that's OK but I cannot see where the figures come from on any of the following (-1 to -10). The numbers produced do not follow what I would expect, am I correct in assuming that the (-1) table for instance should show the skips from the previous draw as if the current draw hasn't been drawn yet.

Also can you explain the difference between "Draw Simulation" & the Simulate Button. Is it that the simuation operates on all 11 columns and the Simulate Button only operates on the current "History to Show".

I like the Filter Function on this page to see from which areas of a skip table the winning numbers come from, very useful when either small or large skips are due.

Rgds

Red Devil

[Simon69]

Hi,

Here a tool I developed last week showing some Differences statistics.

I hope it wil be useful for making differences estimates.

Here the link : http://home.tiscali.nl/lsystems/winsum.html

Step 1 (Creating Diffrences(only) export file : Select the WN History page. Next You select the Summary Table subpage. Select in the dropdown box Diffrences only and press the Update Button. Press the Export button and save the "differences only" output.

Step 2 (Importing the output in WinSum.exe) : Go to the import results page of the tool and press the Import Result Button. Select the export file. The result will be loaded in the grid. Finally press at the Copy Results button and the results will be saved and shown at the WinSum Dif. Results page....

I guess the rest is clear enough....


Later this month ehh May, I will add columns statitstics to the program...


Good luck,

Simon69

[maroc]

Hi Simon 69,

Great tool Thank's for sharing.
Would be nice if you make the window sizable.To fit the creen

[Red Devil]

Simon69,

I share Maroc's sentiment, thanks for sharing. I have downloaded and while I find the Results table useful and the fact that it can be imported into Excel, I am a bit confused with the statistics window? Not sure what it indicates.

Rgds

Red Devil

[Simon69]

Hi Maroc,

Next version the window will be resizeable.

Hi Red Devil,

The statistics page ;

Hit : How many times the diffence (range) was hit Bruto and Netto.
(Netto = if difference(range) hit more than 1 column, this will be counted as 1 hit)
Skip : tmes the diffrence(range) didn't hit
Skip avg : Average skip for the diffrenc(range)
Skip pattern : Previous skip in chronil. order. Left is most recent
Columns Hit pattern : In how many columns the diffrence(range) was fallen.

Regards,

Simon69.


Ps. Next week I'm on holliday so I can't respond earlier than next weekend.

[stan]

[quote=Red Devil]
Like PadawanLotto, I too cannot understand the History DifferencesTable. "History to Show 0" yes that's OK but I cannot see where the figures come from on any of the following (-1 to -10). The numbers produced do not follow what I would expect, am I correct in assuming that the (-1) table for instance should show the skips from the previous draw as if the current draw hasn't been drawn yet.
[/quote]
switch to numbers history tab and select differences only in the show: combo box. values in column -1 then appear in row headers in the history differences tab for column -1, values in the differences table are the numbers that correspond to the difference value in the number history tab. so basically it's an inverted number history table.
the same aplies to other history columns.

Also can you explain the difference between "Draw Simulation" & the Simulate Button. Is it that the simuation operates on all 11 columns and the Simulate Button only operates on the current "History to Show".

when draw simulation button is on then the application takes the numbers from edit fields next to that button and recalculates all wn history values as if those numbers were drawn. it's useful to test how an individual ticket projects into wn history sum values.

simulate button recalculates the sum values for 10000 random tickets and shows the min/max sum value and also how many random tickets lead to sum decrease and how many lead to sum increase (the red/green bar).

[stan]

[quote=Simon69]
Here a tool I developed last week showing some Differences statistics.

I hope it wil be useful for making differences estimates.

Here the link : http://home.tiscali.nl/lsystems/winsum.html

<snip>
[/quote]

looks like an interesting tool, thanks for sharing! btw, do you happen to know java by any chance?
if there's enough interest and if you don't mind i can incorporate it into future expert lotto versions.