Page 7 of 7

Re: nigmm - Just another statistics application

Posted: Sun Jun 11, 2017 1:25 am
by lmviterbo
I may have found a small glitch in the computation of palindromes. You seem to be counting every time a n-palindrome appears in a serial.

So, for example, a serial like M060606xxxxx is now counted twice on the "palindromes with length 5" list. In my opinion, it should be counted only once.

I am, in fact, just guessing, because my own palindrome list states

1880 times serial has palindrom with length 5.

when it should say

1857 times serial has palindrome with length 5.

However, when I count every time a 5-palindrome appears in all my serials, the sum is 1898, not 1880.


(Note: besides the counting problem, "palindrom" should be spelled "palindrome".)

Re: nigmm - Just another statistics application

Posted: Mon Jun 12, 2017 1:51 pm
by tt.01
Hi lmvirtebo,
Thanks for comments.
lmviterbo wrote:On the "Print sheet location bingoup" (Printcode analysis page), I would prefer the non-J6 plates to be blacked out, and the corresponding totals to reflect this fact.

lmviterbo wrote: A nice addition would be a percentage sum in the end of the bingo.

Both can be done. (as soon as I will find small time for it)

lmviterbo wrote:a serial like M060606xxxxx is now counted twice on the "palindromes with length 5" list. In my opinion, it should be counted only once.

If I am sure, that mathematically there two palindromes here: 06060 & 60606.

lmviterbo wrote:"palindrom" should be spelled "palindrome".)

This will be definitely fixed :)

Re: nigmm - Just another statistics application

Posted: Mon Jun 12, 2017 2:39 pm
by lmviterbo
tt.01 wrote:
lmviterbo wrote:a serial like M060606xxxxx is now counted twice on the "palindromes with length 5" list. In my opinion, it should be counted only once.

[…] mathematically there are two palindromes here: 06060 & 60606.

:D Yes, I know that. :D

I would prefer the computation to be of "serials with n-palindromes", instead of the "n-palindromes in serials". I understand that might be just a personal choice though (and this is why I wrote "In my opinion" ;) ).

Anyway, unless I wrote the wrong Excel formulas, there is nevertheless some kind of error in the counting:
1880 times serial has palindrom with length 5. (nigmm)
1857 my counting for serials with palindromes with length 5 (discounting serials with longer lengths)
1898 my counting for palindromes with length 5 in serials (also discounting serials with longer lengths)
1993 my counting for serials with palindromes with length 5 (including serials with longer lengths)
2034 my counting for palindromes with length 5 in serials (including serials with longer lengths)


tt.01 wrote:Both can be done. (as soon as I will find small time for it)
Great, thanks!

Re: nigmm - Just another statistics application

Posted: Mon Jun 12, 2017 8:57 pm
by tt.01
lmviterbo wrote:1880 times serial has palindrom with length 5. (nigmm)
1857 my counting for serials with palindromes with length 5 (discounting serials with longer lengths)
1898 my counting for palindromes with length 5 in serials (also discounting serials with longer lengths)
1993 my counting for serials with palindromes with length 5 (including serials with longer lengths)
2034 my counting for palindromes with length 5 in serials (including serials with longer lengths)

Hmm, probably it would be easier to do the check manually ;) if I will create special binary with dumping calculation result to a file. At least it works fine with small set of simple palindromes. So, pls wait a (long) bit.

Re: nigmm - Just another statistics application

Posted: Wed Jun 14, 2017 10:23 pm
by lmviterbo
Just for the record: nigmm's computation of palindromes is 100% accurate. (My counting was not.)

(With tt.01's help, I realized my Excel formulae were missing some notes and wrongly adding some other.)

Re: nigmm - Just another statistics application

Posted: Sun Sep 03, 2017 2:54 pm
by mico
please, does anybody know why the signature of my 20 Europe Z001/Z is unknown?

Re: nigmm - Just another statistics application

Posted: Sun Sep 03, 2017 5:34 pm
by lmviterbo
It is indeed somehow missing from the signatures database.

Here's what the signatures-external.csv file should be like at the moment (including that missing plate and all the plates that were added to liste.eurobillets.free.fr since the last nigmm update):

--------------------------
5;2013;Y;Y003;MD;
20;2015;Z;Z001;MD;
50;2017;E;E012;MD;
50;2017;R;R010;MD;
50;2017;R;R011;MD;
50;2017;S;S014;MD;
50;2017;U;U012;MD;
--------------------------

Re: nigmm - Just another statistics application

Posted: Fri Sep 08, 2017 12:28 am
by lmviterbo
I loved the last update, especially the plate position bingos ( :oops: ) and the Serial combinatorics. I also found out a nice feature that was probably there already but had passed unnoticed to me: the hide/show option on several tables.

Now here's a new table that might be nice along with the Huge table: a summary of the top "serial keys" (I use this name for things like U**061, for which I have never seen a name). It would be something more or less like this:
top Serial Keys (nigmm suggestion).png
top Serial Keys (nigmm suggestion).png (29.3 KiB) Viewed 235 times

It would be even nicer to have also a different table for each denomination. :roll:

Re: nigmm - Just another statistics application

Posted: Wed Sep 13, 2017 12:06 pm
by gabonik
She tells me a € 5 Y003 as unknown ....?

Re: nigmm - Just another statistics application

Posted: Wed Sep 13, 2017 7:50 pm
by giulcenc
tt.01, a big thanks for your work!

I have just some suggestions for the "hugetable" page, considering the first significant digits of the serial numbers.

First serie
printer E
:note-5: : H12345 must significant digits are H123 instead of actual H**345 (but for :note-20: H**345 is ok...).
printer G
:note-20: : H12345 must significant digits are H123 instead of actual H**345.
printers H, K, and U
M-notes and T-notes can have different first digit even in the same denomination; maybe it's better show M000, M100, M200, M300 and T100, T200, T300 and so on instead of M*000 and T*000.
printer L
:note-20: : U-notes with serial from U85000 up to U88539 must significant digits are U85**0 instead of actual U**000 (if serial is up to U84999 the table is ok).
printer T
:note-200: and :note-500: : U12345 must significant digits are U**345 instead of actual U123.
:note-200: and :note-500: : V12345 must significant digits are V1**45 instead of actual V123 (like the Z-notes).

Re: nigmm - Just another statistics application

Posted: Fri Sep 15, 2017 12:32 pm
by tt.01
Hi,
Some (usually new) signatures could missing in the current application version. I do update them time to time based on information from liste.eurobillets.free.fr. And actually lmviterbo already did very good answer to those questions. You can keep all new/unknown signatures in your local signatures-external.csv file.

lmviterbo wrote:Here's what the signatures-external.csv file should be like at the moment (including that missing plate and all the plates that were added to liste.eurobillets.free.fr since the last nigmm update):
--------------------------
5;2013;Y;Y003;MD;
20;2015;Z;Z001;MD;
50;2017;E;E012;MD;
50;2017;R;R010;MD;
50;2017;R;R011;MD;
50;2017;S;S014;MD;
50;2017;U;U012;MD;
--------------------------

Re: nigmm - Just another statistics application

Posted: Fri Sep 15, 2017 12:47 pm
by tt.01
Hi,
giulcenc wrote:I have just some suggestions for the "hugetable" page, considering the first significant digits of the serial numbers.
...

The very big problem of this topic is that there is no common opinion about "first significant digits of the serial numbers" (i call it as patterns, in short). Long time ago there were hot discussions about it. I am using one variant for this pattern's list. Could you combine your list of all patterns and send it to me by PM. And let's continue discussion there.

Re: nigmm - Just another statistics application

Posted: Sat Sep 16, 2017 4:08 pm
by tralla
Found a glitch in the Signatures:
:note-100: - M004* - Vxxxx1314xxx
are shown as WD. That is wrong, they are definitely MD.

Re: nigmm - Just another statistics application

Posted: Sun Sep 17, 2017 3:56 pm
by tt.01
tralla wrote:Found a glitch in the Signatures:
:note-100: - M004* - Vxxxx1314xxx
are shown as WD. That is wrong, they are definitely MD.

Hi tralla,
I will fix it. Btw, this note line marked red (=recent update/change) on "liste des eurobillets".