A suggestion for a slightly silly but mysterious one

[Bok]

FWIW, doesn't seem to be much different when I do all of them

MariaDB [static]> select cpid,nick,count(*) as mycount from boinc_user where truncate(metric1,0) in (1000,10000,25000,50000,100000,250000,500000,750000,1000000,2500000,5000000,10000000,25000000,50000000,100000000,250000000,500000000,1000000000) group by cpid order by mycount desc limit 0,10;
ERROR 2006 (HY000): MySQL server has gone away
No connection. Trying to reconnect...
Connection id:    170374
Current database: static

+----------------------------------+-----------+---------+
| cpid                             | nick      | mycount |
+----------------------------------+-----------+---------+
| 0dd9749ec16c8900f065d871290156bc | BAR       |       7 |
| 2286c5351f0f2d2f5d7fee45967f22b8 | Baryna.ch |       5 |
| 8f15950e07e24a5c0c73016a77c952e9 | Nflight   |       4 |
| e33013c782482e80caf0c65b476a2d87 | JayPi     |       4 |
| 89e772f65716bba7c962b4a13d14e350 | AMDave    |       4 |
| 61c4eaaca6667bc9dda3412188c5f173 | x3mEn     |       3 |
| 476d777382a3555be83f6af9ab59c9f9 | Szopler   |       3 |
| 8496797f94775aac42635504add88c62 | $         |       3 |
| d7a3d9afaf379c6984d47f1d9672f5e2 | UL1       |       3 |
| 99fecfdc6caaa3ee9f7365297dd65634 | williami  |       3 |
+----------------------------------+-----------+---------+
10 rows in set (10.022 sec)

[MarkRBright]

I think I see, and that looks like a promising start - other than the "MySQL server has gone away" bit!
Combine it with SubProject breakpoints and Nebbers is go - assuming there's some tolerance in there somewhere - though I don't see it if it is. No I'm not looking for a free SQL lesson! ;-) Good work I think!

[MarkRBright]

The nice thing about this one is that I think everyone is as likely as anyone else to get one. OK so the more projects / SubProjects you have worked on increases the odds a fair bit, but someone could have worked on 5 projects and have one or two of these and quite probably wont know why! I always got the impression that you were quite keen on mysterious ones Bok.
I do think the 0.1% tolerance would make a difference if you haven't already got it in there. I would have thought that they will be far more commonplace with that tweak. Or would it be easier to just arbitrarily say plus or minus 500 say, or some such? I'm not bothered, but I suspect you will find more of them with percentages. Why not have a look at the data and pick what you think works best.

[Bok]

It would make it much more difficult to make it to have a range of values (tolerance) as in sql terms that would be individual between (9999 and 10001) for example

Though I certainly see the appeal of doing that.

MariaDB [static]> select cpid,nick,count(*) as mycount from boinc_user where (truncate(metric1,0) between 9999 and 10001) or (truncate(metric1,0) between 99999 and 100001) group by cpid order by mycount desc limit 0,10;
+----------------------------------+-------------+---------+
| cpid                             | nick        | mycount |
+----------------------------------+-------------+---------+
| 0dd9749ec16c8900f065d871290156bc | BAR         |       8 |
| 2286c5351f0f2d2f5d7fee45967f22b8 | Baryna.ch   |       7 |
| 61c4eaaca6667bc9dda3412188c5f173 | x3mEn       |       6 |
| e33013c782482e80caf0c65b476a2d87 | JayPi       |       5 |
| a848091602d6dade8ee661252d106d2d | rilian      |       4 |
| 89e772f65716bba7c962b4a13d14e350 | AMDave      |       4 |
| d5f358d1a7913ab61d7ea87c85aafbfd | Buster Gunn |       4 |
| 8f15950e07e24a5c0c73016a77c952e9 | Nflight     |       3 |
| 3a045e428e45f087f7cc832bdb4791ff | RAD-Poland  |       3 |
| 2e8203c09bffa7dd415284af2124eb53 | skivelitis  |       2 |
+----------------------------------+-------------+---------+
10 rows in set (9.725 sec)

but, saying that - it would only run once per day.

[MarkRBright]

I feel that it wouldn't really be worth doing if there were going to be so few of them - so I wouldn't bother unless tolerances are doable. Just had a quick look at BAR and it appears that whoever he or she is, ramped up a bunch of Trigggls and stopped and may not have Boinced for some time so this could be wasted on them!!! (On that note, would it be easy to introduce a Date Last Seen under the current date First Seen in the headers - perhaps you already track this but don't show it? Obviously you do track the hosts but that would be a much more long winded approach, and you could just update a date when you update any other stuff in the daily update. When displaying you could just show "Unknown" if the date is blank - Sorry - I digress)
In your example above I would argue that it should have been between 99990 and 100010 for the second one, though after seeing it 0.5% (maybe more?) might be more interesting, as more people would get them and quite possibly have no clue as to why. Why anyone think they would get 'rewarded' for a subproject reaching 99952, 499751 or 24876 credits after all? That would be more interesting. Maybe you should delete this thread once you have decided ;-)

[MarkRBright]

Just in case it helps you Bok, I knocked this up - it doesn't take into account SubProjects but I suspect a bit of copy and paste would do so in a few seconds. I went for a 0.5% tolerance. Why not see how long it takes to run and decide if it's doable or not

MariaDB [static]> select cpid,nick,count(*) as mycount from boinc_user
where
(truncate(metric1,0) between 9995 and 10005)
or (truncate(metric1,0) between 24975 and 25025)
or (truncate(metric1,0) between 49950 and 50050)
or (truncate(metric1,0) between 99995 and 10050)
or (truncate(metric1,0) between 249750 and 250250)
or (truncate(metric1,0) between 499500 and 500500)
or (truncate(metric1,0) between 749250 and 750750)
or (truncate(metric1,0) between 999995 and 1000050)
or (truncate(metric1,0) between 2497500 and 2502500)
or (truncate(metric1,0) between 4995000 and 5005000)
or (truncate(metric1,0) between 9999995 and 10000050)
or (truncate(metric1,0) between 24975000 and 25025000)
or (truncate(metric1,0) between 49950000 and 50050000)
or (truncate(metric1,0) between 99999995 and 100000050)
or (truncate(metric1,0) between 249750000 and 250250000)
or (truncate(metric1,0) between 499500000 and 500500000)
or (truncate(metric1,0) between 999999995 and 1000000050)
or (truncate(metric1,0) between 2497500000 and 2502500000)
or (truncate(metric1,0) between 4995000000 and 5005000000)
or (truncate(metric1,0) between 9999999995 and 10000000050)
or (truncate(metric1,0) between 24975000000 and 25025000000)
or (truncate(metric1,0) between 49950000000 and 50050000000)
or (truncate(metric1,0) between 99999999995 and 100000000500)
or (truncate(metric1,0) between 249750000000 and 250250000000)
or (truncate(metric1,0) between 499500000000 and 500500000000)
or (truncate(metric1,0) between 999999999995 and 1000000000050)
or (truncate(metric1,0) between 2497500000000 and 2502500000000)
or (truncate(metric1,0) between 4995000000000 and 5005000000000)
or (truncate(metric1,0) between 9999999999995 and 10000000000050)

group by cpid order by mycount desc limit 0,10;

[Bok]

subprojects are a different table so it that complicates..

I'm not sure you have them all correct, for instance 99995 and 10050, in fact all of your ones like that have a 5 tolerance on the prior but much more on the post side

Also don't really need the truncate part when we are doing between.

This one though gives this subset

+----------------------------------+---------------+---------+
| cpid                             | nick          | mycount |
+----------------------------------+---------------+---------+
| e33013c782482e80caf0c65b476a2d87 | JayPi         |      24 |
| 476d777382a3555be83f6af9ab59c9f9 | Szopler       |      22 |
| 61c4eaaca6667bc9dda3412188c5f173 | x3mEn         |      18 |
| 0dd9749ec16c8900f065d871290156bc | BAR           |      18 |
| 5dad1593a3bd6cd31ce4a9b507ba20fc | Ananas        |      12 |
| 3bfe3268986a7fe9ebaa5acc02e69393 | somanyroads   |       9 |
| 7907e81b12e62de3e9a0c5459a973318 | Mark Gallaher |       9 |
| 5ce7cfc3ff8b21f50b0eb24a7d29ffcd | FourOh        |       9 |
| d7a3d9afaf379c6984d47f1d9672f5e2 | UL1           |       8 |
| 8e0b099acd93adc95684151a4eef9b86 | Bruce Kennedy |       7 |
+----------------------------------+---------------+---------+
10 rows in set (21.496 sec)