Hit interestingness scores (phase 2)

Concerns? Let us know by posting here.

Moderators: avij, Phaseolus, Fons, dserrano5

User avatar
avij
Forum Moderator
Forum Moderator
Posts: 6120
Joined: Mon May 27, 2002 10:45 pm
Location: Helsinki Finland
Contact:

Re: Hit interestingness scores (phase 2)

Post by avij »

lmviterbo wrote:The remark made by an-148 might make sense in fact. Here's a new experiment:

Code: Select all

 taking into account rarity of hits per denomination, as suggested by an-148
+--------------+-----------|--------+------------------|----------------------+
| denomination |   notes   |  hits  |   notes / hits   | =100*D2/MAX(D$2:D$8) |
+--------------+-----------|--------+------------------|----------------------+
|            5 |  46025746 | 559156 |  82.312889426207 |     16.7415696556489 |
|           10 |  27569367 | 159650 | 172.686295020357 |     35.1225629037242 |
|           20 |  24223518 | 141445 | 171.257506451271 |     34.8319624458943 |
|           50 |  14774917 |  38525 | 383.515042180402 |     78.0028964772086 |
|          100 |   2074346 |   4219 | 491.667693766295 |    100.0000000000000 |
|          200 |    275461 |    846 | 325.604018912530 |     66.2244078756371 |
|          500 |    430090 |   1040 | 413.548076923077 |     84.1112975626275 |
+--------------+-----------|--------+------------------|----------------------+
Hm, I have some second thoughts about this.. In particular, this would mean that hits with 50e notes would be more interesting than hits with 200e notes. Sure, yes, I understand that if you get a bundle of used 50e and 200e notes from a bank, it's more likely that you'll get a hit with a 200e note than with a 50e note. I think 50e is the most common denomination in circulation, whereas 200e is the least common denomination. This will definitely affect the hit ratios as well.

Take this as a test -- have a look at the hit listing and see which ones raise your eyebrows more, the 50e hits or 200e hits. That should be the metric that we should be using, not how easy it WOULD be to get a hit with some denomination. If you really want to use the "how difficult it is to get a hit with some denomination" way of thinking, you should take into account how difficult it is to get the note in the first place. By basing the calculations on the number of hits that have really occurred we'd get real data for how unusual such hits are.

There are 431 users who have hits with both 50e notes and 200e notes. Of those, there is only one user who has more hits with 200e notes (2) than with 50e notes (1). I think the users who have both 50e and 200e hits also think their 200e hits are more interesting than the 50e hits. I can't really come up with a good justification for ranking the 50e notes higher than the 200e notes.
Money makes the world go round. We track how the money goes round the world.
User avatar
an-148
Euro-Master
Euro-Master
Posts: 4596
Joined: Wed Oct 11, 2006 5:04 pm
Location: summer: Ivanica, BiH ; winter: La Calamine (Liège) BE
Contact:

Re: Hit interestingness scores (phase 2)

Post by an-148 »

yes, but:

1. it's not uneasy to get :note-200: : simply ask them instead of other notes at the bank (I know, you will not get very much of them for the same amount of money :D :D :D but that's also true for other high denominations)
2. next remark is personnal (but maybe other users share the opinion): the :note-50: is the most boring note to register: you have more of them to register than higher values ......... with unendless "no-hit" registrations; the printer code is displaced , and ... simply boring note !! a hit with one of them is highly valuable :D :D
(in addition, :note-200: - hits and :note-500: - hits are often the consequence of registrations by note collectors : this diminishes a bit the interrestingness)

I believe the pure mathematic is the adequate way to approach the interrestingness by value (only my humble opinion :wink: )
http://meine.flugstatistik.de/image/an148.gif" onclick="window.open(this.href);return false;
mizar
Euro-Regular in Training
Euro-Regular in Training
Posts: 136
Joined: Tue Jun 05, 2012 5:18 pm
Location: Slovenia

Re: Hit interestingness scores (phase 2)

Post by mizar »

I'm with avij on this one. I'd say we don't start with maths, the starting point is what EBT users consider interesting, and then we try to translate as many elements of this as possible into formulas.

(Also I like the printer code location on :note-50:, it's the easiest to see :P And after that :note-5:.)
luckbox
Euro-Regular
Euro-Regular
Posts: 260
Joined: Fri Aug 05, 2011 10:20 am

Re: Hit interestingness scores (phase 2)

Post by luckbox »

I think you all are making this score too complicated and annoying for many big EBT-ers. There are very large EBT-ers who enter a lot of :note-5: and I don't think you should not discourage these EBT-ers. Therefor the denomination should not be part of the interesting score. We should be happy with these EBT-ers, because they keep the site active.

In my opinion we could create a rating for very interesting hits (for example less common international hits), but we should not go so far that every :note-5: hit is rated as less interesting than any hit with another denomination.

P.s. I do not post this because of personal interest (a little less than 35% of my notes is :note-5:) but I'm worried that some large EBT-ers would be discouraged and become less active or even stop with EBT.
User avatar
avij
Forum Moderator
Forum Moderator
Posts: 6120
Joined: Mon May 27, 2002 10:45 pm
Location: Helsinki Finland
Contact:

Re: Hit interestingness scores (phase 2)

Post by avij »

If you think the denomination should not be taken into account, you can propose to assign it a minimal weight. This will be taken care of in the fourth phase, ie. two steps from now. Let's not concern ourselves with the weights yet.

I personally think the denomination should be taken into account when calculating the scores.

But don't get stuck on the denominations -- there are a few dozen other metrics that'd need to be converted into a number, feel free to take a stab at the other metrics as well.
Money makes the world go round. We track how the money goes round the world.
User avatar
klapotec
Euro-Master
Euro-Master
Posts: 2604
Joined: Tue Dec 27, 2005 12:18 pm
Location: Steiermark / Štajerska / Styria

Re: Hit interestingness scores (phase 2)

Post by klapotec »

Speaking as a "large" EBTer - albeit one who tries to track all denominations somewhat equally, as can be seen by my value of entered notes by denomination - I find that hits with :note-5: are not as interesting as, say, with :note-100: ; I don't currently have the time to spare to put numbers to all this, but the denomination should definitely count, adjusted for region/country and how many bills of that denomination have been tracked by the hitpartners, i.e. if someone enters mostly :note-20: and has a :note-5: hit with someone who predominantly enters :note-10: then it's presumably more interesting for either than a :note-10: or :note-20: hit.
User avatar
bob9
Forum Moderator
Forum Moderator
Posts: 203
Joined: Sat Aug 29, 2009 7:58 pm
Location: Vaasa

Re: Hit interestingness scores (phase 2)

Post by bob9 »

Give comments a fixed value. If there is a comment, score +N, and if there's no comment score -N. It's hard to judge if a comment is interesting or not, but I think we can all agree that no comment at all is always boring.
Cash is king. | My EBT profile
User avatar
avij
Forum Moderator
Forum Moderator
Posts: 6120
Joined: Mon May 27, 2002 10:45 pm
Location: Helsinki Finland
Contact:

Re: Hit interestingness scores (phase 2)

Post by avij »

bob9 wrote:Give comments a fixed value. If there is a comment, score +N, and if there's no comment score -N. It's hard to judge if a comment is interesting or not, but I think we can all agree that no comment at all is always boring.
Right. However, I prefer to keep things positive, so the scoring for this metric would be simply "if the note entry has a comment, score=100; otherwise score=0".
Money makes the world go round. We track how the money goes round the world.
User avatar
avij
Forum Moderator
Forum Moderator
Posts: 6120
Joined: Mon May 27, 2002 10:45 pm
Location: Helsinki Finland
Contact:

Re: Hit interestingness scores (phase 2)

Post by avij »

Some food for thought:
http://miuku.net/tmp/hit_days.txt" onclick="window.open(this.href);return false;
http://miuku.net/tmp/hit_kms.txt" onclick="window.open(this.href);return false;

These figures are for the total days and kilometres for a hit, so a triple with 100km + 200km travel distances is tabulated as 300km in the above list, likewise for days. Moderated (and obviously fake) hits are also included in the numbers.
Money makes the world go round. We track how the money goes round the world.
User avatar
avij
Forum Moderator
Forum Moderator
Posts: 6120
Joined: Mon May 27, 2002 10:45 pm
Location: Helsinki Finland
Contact:

Re: Hit interestingness scores (phase 2)

Post by avij »

This information was requested from me privately, but I don't see a reason why it couldn't be published here on the forum for others as well.

Here is a breakdown of the hits per denomination per year. The years are tallied starting from the 2nd entry of the hit, so a triple from 2009-2010-2011 counts as zero for 2009, one for 2010 and one for 2011. Initially I thought about counting only the 2nd entry of each hit, but perhaps this method is more descriptive of the hit frequencies.

Including moderated hits:

Code: Select all

+------+--------------+-------+
| year | denomination | count |
+------+--------------+-------+
| 2002 | 5            |   987 |
| 2002 | 10           |   713 |
| 2002 | 20           |   402 |
| 2002 | 50           |   254 |
| 2002 | 100          |    34 |
| 2002 | 200          |     3 |
| 2002 | 500          |     1 |
| 2003 | 5            |   622 |
| 2003 | 10           |   392 |
| 2003 | 20           |   343 |
| 2003 | 50           |   168 |
| 2003 | 100          |    19 |
| 2003 | 200          |     5 |
| 2003 | 500          |     3 |
| 2004 | 5            |  3807 |
| 2004 | 10           |  2233 |
| 2004 | 20           |  1912 |
| 2004 | 50           |   742 |
| 2004 | 100          |    70 |
| 2004 | 200          |    17 |
| 2004 | 500          |    12 |
| 2005 | 5            | 24842 |
| 2005 | 10           |  9128 |
| 2005 | 20           |  6997 |
| 2005 | 50           |  2154 |
| 2005 | 100          |   243 |
| 2005 | 200          |    40 |
| 2005 | 500          |    43 |
| 2006 | 5            | 41551 |
| 2006 | 10           | 13421 |
| 2006 | 20           | 12351 |
| 2006 | 50           |  3734 |
| 2006 | 100          |   353 |
| 2006 | 200          |    53 |
| 2006 | 500          |   132 |
| 2007 | 5            | 68093 |
| 2007 | 10           | 18458 |
| 2007 | 20           | 17193 |
| 2007 | 50           |  4624 |
| 2007 | 100          |   761 |
| 2007 | 200          |   118 |
| 2007 | 500          |   254 |
| 2008 | 5            | 71307 |
| 2008 | 10           | 22479 |
| 2008 | 20           | 19372 |
| 2008 | 50           |  5526 |
| 2008 | 100          |   525 |
| 2008 | 200          |   133 |
| 2008 | 500          |   217 |
| 2009 | 5            | 98103 |
| 2009 | 10           | 24953 |
| 2009 | 20           | 23101 |
| 2009 | 50           |  5724 |
| 2009 | 100          |   609 |
| 2009 | 200          |   170 |
| 2009 | 500          |   114 |
| 2010 | 5            | 83970 |
| 2010 | 10           | 23486 |
| 2010 | 20           | 22607 |
| 2010 | 50           |  5413 |
| 2010 | 100          |   554 |
| 2010 | 200          |    87 |
| 2010 | 500          |    84 |
| 2011 | 5            | 82251 |
| 2011 | 10           | 22041 |
| 2011 | 20           | 19012 |
| 2011 | 50           |  4969 |
| 2011 | 100          |   421 |
| 2011 | 200          |   105 |
| 2011 | 500          |    73 |
| 2012 | 5            | 80904 |
| 2012 | 10           | 20404 |
| 2012 | 20           | 16982 |
| 2012 | 50           |  4705 |
| 2012 | 100          |   606 |
| 2012 | 200          |   114 |
| 2012 | 500          |   125 |
| 2013 | 5            | 13813 |
| 2013 | 10           |  4381 |
| 2013 | 20           |  3388 |
| 2013 | 50           |   946 |
| 2013 | 100          |    81 |
| 2013 | 200          |    21 |
| 2013 | 500          |    20 |
+------+--------------+-------+
Only unmoderated hits:

Code: Select all

+------+--------------+-------+
| year | denomination | count |
+------+--------------+-------+
| 2002 | 5            |    55 |
| 2002 | 10           |    51 |
| 2002 | 20           |    36 |
| 2002 | 50           |    15 |
| 2002 | 500          |     1 |
| 2003 | 5            |   259 |
| 2003 | 10           |   151 |
| 2003 | 20           |   133 |
| 2003 | 50           |    50 |
| 2003 | 100          |     3 |
| 2003 | 500          |     1 |
| 2004 | 5            |  2538 |
| 2004 | 10           |  1574 |
| 2004 | 20           |  1317 |
| 2004 | 50           |   437 |
| 2004 | 100          |    22 |
| 2004 | 200          |     4 |
| 2004 | 500          |     5 |
| 2005 | 5            | 16896 |
| 2005 | 10           |  6858 |
| 2005 | 20           |  5624 |
| 2005 | 50           |  1579 |
| 2005 | 100          |   101 |
| 2005 | 200          |    18 |
| 2005 | 500          |    15 |
| 2006 | 5            | 27200 |
| 2006 | 10           | 10231 |
| 2006 | 20           | 10086 |
| 2006 | 50           |  2765 |
| 2006 | 100          |   209 |
| 2006 | 200          |    37 |
| 2006 | 500          |    66 |
| 2007 | 5            | 38991 |
| 2007 | 10           | 14293 |
| 2007 | 20           | 13756 |
| 2007 | 50           |  3557 |
| 2007 | 100          |   384 |
| 2007 | 200          |    68 |
| 2007 | 500          |   108 |
| 2008 | 5            | 53145 |
| 2008 | 10           | 18929 |
| 2008 | 20           | 17121 |
| 2008 | 50           |  4344 |
| 2008 | 100          |   365 |
| 2008 | 200          |    82 |
| 2008 | 500          |    93 |
| 2009 | 5            | 69235 |
| 2009 | 10           | 21078 |
| 2009 | 20           | 20953 |
| 2009 | 50           |  4830 |
| 2009 | 100          |   421 |
| 2009 | 200          |    97 |
| 2009 | 500          |    65 |
| 2010 | 5            | 63766 |
| 2010 | 10           | 19353 |
| 2010 | 20           | 19994 |
| 2010 | 50           |  4470 |
| 2010 | 100          |   378 |
| 2010 | 200          |    56 |
| 2010 | 500          |    39 |
| 2011 | 5            | 63501 |
| 2011 | 10           | 18361 |
| 2011 | 20           | 16550 |
| 2011 | 50           |  3986 |
| 2011 | 100          |   304 |
| 2011 | 200          |    72 |
| 2011 | 500          |    50 |
| 2012 | 5            | 61123 |
| 2012 | 10           | 16231 |
| 2012 | 20           | 13994 |
| 2012 | 50           |  3572 |
| 2012 | 100          |   321 |
| 2012 | 200          |    45 |
| 2012 | 500          |    42 |
| 2013 | 5            | 10758 |
| 2013 | 10           |  3554 |
| 2013 | 20           |  2797 |
| 2013 | 50           |   713 |
| 2013 | 100          |    58 |
| 2013 | 200          |     8 |
| 2013 | 500          |     9 |
+------+--------------+-------+
Money makes the world go round. We track how the money goes round the world.
User avatar
lmviterbo
Euro-Master
Euro-Master
Posts: 6518
Joined: Thu Aug 21, 2003 5:23 pm
Location: Lisboa, Portugal
Contact:

Re: Hit interestingness scores (phase 2)

Post by lmviterbo »

Thanks, avij!

Is it possible to have also notes per denomination per year?
User avatar
avij
Forum Moderator
Forum Moderator
Posts: 6120
Joined: Mon May 27, 2002 10:45 pm
Location: Helsinki Finland
Contact:

Re: Hit interestingness scores (phase 2)

Post by avij »

Sure, here you go:

Code: Select all

+------+--------------+---------+
| year | denomination | count   |
+------+--------------+---------+
| 2002 | 5            |  153505 |
| 2002 | 10           |  131557 |
| 2002 | 20           |  137378 |
| 2002 | 50           |   89657 |
| 2002 | 100          |   11150 |
| 2002 | 200          |    3974 |
| 2002 | 500          |    3865 |
| 2003 | 5            |  280148 |
| 2003 | 10           |  252815 |
| 2003 | 20           |  264616 |
| 2003 | 50           |  185750 |
| 2003 | 100          |   21410 |
| 2003 | 200          |    5922 |
| 2003 | 500          |    7139 |
| 2004 | 5            | 1046720 |
| 2004 | 10           |  830950 |
| 2004 | 20           |  825786 |
| 2004 | 50           |  537566 |
| 2004 | 100          |   59042 |
| 2004 | 200          |   11117 |
| 2004 | 500          |   13713 |
| 2005 | 5            | 2523166 |
| 2005 | 10           | 1825060 |
| 2005 | 20           | 1746465 |
| 2005 | 50           | 1109666 |
| 2005 | 100          |  126729 |
| 2005 | 200          |   21892 |
| 2005 | 500          |   31095 |
| 2006 | 5            | 3948420 |
| 2006 | 10           | 2648206 |
| 2006 | 20           | 2481731 |
| 2006 | 50           | 1536572 |
| 2006 | 100          |  179187 |
| 2006 | 200          |   26948 |
| 2006 | 500          |   45546 |
| 2007 | 5            | 5183345 |
| 2007 | 10           | 3349721 |
| 2007 | 20           | 3052646 |
| 2007 | 50           | 1813229 |
| 2007 | 100          |  256769 |
| 2007 | 200          |   36457 |
| 2007 | 500          |   63805 |
| 2008 | 5            | 5789160 |
| 2008 | 10           | 3653454 |
| 2008 | 20           | 3178807 |
| 2008 | 50           | 1975420 |
| 2008 | 100          |  283212 |
| 2008 | 200          |   40110 |
| 2008 | 500          |   61220 |
| 2009 | 5            | 6790741 |
| 2009 | 10           | 3814997 |
| 2009 | 20           | 3315598 |
| 2009 | 50           | 1987617 |
| 2009 | 100          |  313728 |
| 2009 | 200          |   37275 |
| 2009 | 500          |   56755 |
| 2010 | 5            | 6327523 |
| 2010 | 10           | 3609974 |
| 2010 | 20           | 3147402 |
| 2010 | 50           | 1833156 |
| 2010 | 100          |  276084 |
| 2010 | 200          |   31834 |
| 2010 | 500          |   50039 |
| 2011 | 5            | 6446632 |
| 2011 | 10           | 3482225 |
| 2011 | 20           | 2882419 |
| 2011 | 50           | 1745973 |
| 2011 | 100          |  262063 |
| 2011 | 200          |   30202 |
| 2011 | 500          |   47137 |
| 2012 | 5            | 6377711 |
| 2012 | 10           | 3310396 |
| 2012 | 20           | 2682303 |
| 2012 | 50           | 1651911 |
| 2012 | 100          |  238910 |
| 2012 | 200          |   25086 |
| 2012 | 500          |   41907 |
| 2013 | 5            | 1291512 |
| 2013 | 10           |  728090 |
| 2013 | 20           |  560406 |
| 2013 | 50           |  338967 |
| 2013 | 100          |   50350 |
| 2013 | 200          |    5079 |
| 2013 | 500          |    9108 |
+------+--------------+---------+
(this list is technically of note entries, not individual notes)
Money makes the world go round. We track how the money goes round the world.
User avatar
lmviterbo
Euro-Master
Euro-Master
Posts: 6518
Joined: Thu Aug 21, 2003 5:23 pm
Location: Lisboa, Portugal
Contact:

Re: Hit interestingness scores (phase 2)

Post by lmviterbo »

avij wrote:(this list is technically of note entries, not individual notes)
Thanks!

Am I right to assume that number of individual notes = note entries - hits?
User avatar
avij
Forum Moderator
Forum Moderator
Posts: 6120
Joined: Mon May 27, 2002 10:45 pm
Location: Helsinki Finland
Contact:

Re: Hit interestingness scores (phase 2)

Post by avij »

lmviterbo wrote:
avij wrote:(this list is technically of note entries, not individual notes)
Thanks!

Is the number of individual notes = note entries - hits?
Triples (and higher) might make such blanket statements inaccurate.. But if you subtract the "hit" figures from the "note entries" figures that I have posted above, you should end up with the count of individual notes regardless of the hits' cardinality, due to the way the "hit" numbers were calculated.

Although for many of these calculations we should disregard the current moderation status of the hits, it's possible that for this particular case the "only unmoderated hits" figures might be more useful. This is because of the planned hit grouping feature -- I believe the current "only unmoderated hits" is a closer approximation of the number of the eventual hit groups than the "including moderated hits" figures.
Money makes the world go round. We track how the money goes round the world.
User avatar
lmviterbo
Euro-Master
Euro-Master
Posts: 6518
Joined: Thu Aug 21, 2003 5:23 pm
Location: Lisboa, Portugal
Contact:

Re: Hit interestingness scores (phase 2)

Post by lmviterbo »

As I suspected, EBTers do not register notes according to the same proportion of denominations in circulation.

Here's the rate at which EBTers registered notes in 2012, per denomination (on the two rightmost columns):

Code: Select all

+--------------+---------+-------+------------+-------------+-------------+-------------+----------+----------+
|              |    E    |   H   |  N = E - H |      U      |      U      |      C      | R1 = C/E | R2 = C/N |
+--------------+---------+-------+------------+-------------+-------------+-------------+----------+----------+
|              |   note  |  hits | individual |     hits    | unmoderated | circulation | entering | entering |
| denomination | entries |  all  |    notes   | unmoderated |  percentage |    August   |  rate 1  |  rate 2  |
+--------------+---------+-------+------------+-------------+-------------+-------------+----------+----------+
|   5          | 6377711 | 80904 |    6316588 |       61123 |      75.55% |   1536076.0 |   240.85 |   243.18 |
|  10          | 3310396 | 20404 |    3294165 |       16231 |      79.55% |   2042139.0 |   616.89 |   619.93 |
|  20          | 2682303 | 16982 |    2668309 |       13994 |      82.40% |   2786663.1 |  1038.91 |  1044.36 |
|  50          | 1651911 |  4705 |    1648339 |        3572 |      75.92% |   6031228.6 |  3651.06 |  3658.97 |
| 100          |  238910 |   606 |     238589 |         321 |      52.97% |   1652287.1 |  6915.94 |  6925.24 |
| 200          |   25086 |   114 |      25041 |          45 |      39.47% |    184715.5 |  7363.29 |  7376.52 |
| 500          |   41907 |   125 |      41865 |          42 |      33.60% |    590872.6 | 14099.62 | 14113.76 |
+--------------+---------+-------+------------+-------------+-------------+-------------+----------+----------+
So approximately

one in every   240 :note-5:
one in every   620 :note-10:
one in every  1040 :note-20:
one in every  3650 :note-50: ┣ in circulation was registered on EBT
one in every  6920 :note-100:
one in every  7370 :note-200:
one in every 14100 :note-500:

Some possible explanations for this:
  • All forms of forced tracking (paying with large denomination notes to get small ones, breaking large notes and "unbreaking" small ones) lead to smaller denominations being much more commonly registered.
  • Notes of high currency may be less prone to be entered on EBT because :note-500: and other high currencies are more pervasive in big transactions or illegal ones that trackers might want to avoid to register for safety or privacy reasons.
  • Notes of high currency are to be found outside Europe in larger quantities than in the Euro Area.
Now arguably because of this disparity in how banknotes are registered we value large denominations more than small ones.

Here's a probability (or combinatorics?) problem I unfortunately can't solve: given the amount of notes in circulation (C) and the amount of times someone entered a note (E), what would be the expected number of hits (double entries)?


On another note, I am appalled at the moderation percentage! As high as 2/3 of :note-500: are moderated. Even the less moderated denominations have 1/4 of their notes moderated. That was even worse in the first years. I had absolutely no idea of this. By the way, this too might have an effect on how EBTers consider large denomination hits more valuable.


Note: I used circulation figures from August since it's the month that best correlates with the average for this year (and possibly always), as can be derived from data on ECB circulation stats.
Post Reply

Return to “Feedback and Development”