Bitcoin Forum>Bitcoin>Mining>Pools
Author

Topic: Strange, bad things happening with deepbit stats. (Read 1175 times)

Lem
newbie
Activity: 78
Merit: 0
Strange, bad things happening with deepbit stats.
June 06, 2012, 11:42:44 AM
#1
First of all, some rounds show an incredibly high hashrate, even for deepbit.
Nowadays deepbit usually collects $difficulty*0.43 in about fifteen minutes.

But look at it:
Code:
06.06 14:03:56 - 286 1340852  0.01034492
06.06 13:48:06 -
1340852 shares in 15 minutes? This means more than 6000Gh/s. It's not a bit too much? :-o

Only two rounds after, we're much below average:
Code:
06.06 14:43:09 - 242 919151  0.01276939
06.06 14:23:18

Only 1600Gh/s. :-o

Strange huge variations: deepbit is heavily hopped, I know... but in my little experience this is still unusual.
Well, who knows...

However I have to report here an IMO much, much more serious issue.

My stats on deepbit are often wrong, with shares that go from the round they've been actually computed to other rounds. I realized it some days ago (I saw two empty rounds, while I could prove I had sent shares iduring those rounds) and I sent Tycho some emails. Then no problems for a while, but now...

I paste here my latest email to Tycho, sent today.

Quote
Dear Tycho,

I'm sorry, but there's something terribly wrong in your stats. We
discussed the matter some days ago. Shares of one round are shifted to
other rounds. Problem is: this does affect proportional rewards.

My account is [email protected], the same for my worker.

Here it is a bit of my worker stats you show on your site.
I'm adding the rightmost column, in which I write how many shares I
expected to see instead (according to my logs).

          Block                 Shares  
Time/Link       Found inYour    Total   Reward          Right shares n°
06.06 14:23:18   -      228     1428291   0.00774212    228
06.06 14:03:56   -      286     1340852   0.01034492    270
06.06 13:48:06   -      248     2319673   0.00518521    230
06.06 12:58:16   -      497     2689836   0.00896133    253 !!!
06.06 12:12:58   -      212     1839576   0.00558933    211
06.06 11:50:05   -      288     1816367   0.00769008    264
06.06 11:26:03   -      523     2514285   0.01008855    263 !!!
06.06 10:44:29   -      239     1372505   0.00844551    235
06.06 10:28:57   -      10      301871    0.00160665    8
06.06 10:23:21   -      484     1352402   0.01735727    252 !!!
06.06 10:10:07   -      38      117943    0.01562619    23  
06.06 10:08:23   -      300     5313607   0.00273825    296
06.06 08:13:06   -      2       199652    0.00048585    1
06.06 08:08:49   -      1       369686    0.00013119    281 !!!
06.06 07:21:59   -      0       136706    None          285 !!!
06.06 06:33:10   -      1       653542    0.00007421    234 !!!
06.06 05:47:26   -      240     1901919   0.00612013    266
06.06 05:02:52   -      5       3219479   0.00007532    5
06.06 03:50:02   -      3       525427    0.00027692    2
06.06 03:38:12   -      1       632138    0.00007672    2
06.06 03:23:52   -      263     1107742   0.01151487    263
06.06 02:59:04   -      4       2647862   0.00007327    4  
06.06 01:59:12   -      242     1060949   0.01106274    243
06.06 01:35:09   -      0       1700663   None          0
06.06 00:57:04   -      233     1085236   0.01041294    232
06.06 00:32:57  

TOTAL                   4348                         4351              

Look, for example, at this particular round (blocks 183248, 183249, and
183250):

06.06 10:23:21   -      484     1352402   0.01735727
06.06 10:10:07  

484 shares in  13 minutes: 37 shares/min, while my system is capable of
about 20 at best! This isn't luck, this is nonsense!

In fact I have logged 45 shares for block 183248, 39 for 183249 and 168
for 183250. A total of 252, so 19 shares/min, which has sense.
And my prop rewards are bounded to these wrong shares numbers, so
sometime I gain, sometime I lose. It's easy to calculate the net income
or loss.

IMO this bug is extremely serious, and it must be fixed as soon as
possible! I'm really concerned.
What do you think about it?

I think I have to post to bitcointalk, too.

By the way: this below is my log for the rounds above: shares belonging
to blocks. My count may be wrong for very few shares that can shift from
one block to the next, but no more than that. And my total count is
right: every time your server accepts one of my shares, my count climbs
by one. And I have logged  also the partial hash of every share, as
usual.

mer 6 giu 2012, 00.46.11, CEST 183186 222
mer 6 giu 2012, 00.54.45, CEST 183187 10
mer 6 giu 2012, 00.57.07, CEST 183188
mer 6 giu 2012, 00.59.30, CEST 183189
mer 6 giu 2012, 01.13.20, CEST 183190
mer 6 giu 2012, 01.24.51, CEST 183191
mer 6 giu 2012, 01.35.14, CEST 183192
mer 6 giu 2012, 01.40.04, CEST 183193 77
mer 6 giu 2012, 01.56.50, CEST 183194 165
mer 6 giu 2012, 01.59.16, CEST 183195 1
mer 6 giu 2012, 02.00.03, CEST 183196 1
mer 6 giu 2012, 02.24.03, CEST 183197
mer 6 giu 2012, 02.29.26, CEST 183198 1
mer 6 giu 2012, 02.36.05, CEST 183199 1
mer 6 giu 2012, 02.40.12, CEST 183200
mer 6 giu 2012, 02.59.10, CEST 183201 1
mer 6 giu 2012, 03.15.40, CEST 183202 259
mer 6 giu 2012, 03.22.24, CEST 183203 3
mer 6 giu 2012, 03.23.56, CEST 183204
mer 6 giu 2012, 03.24.14, CEST 183205 1
mer 6 giu 2012, 03.24.28, CEST 183206
mer 6 giu 2012, 03.32.55, CEST 183207
mer 6 giu 2012, 03.38.15, CEST 183208 1
mer 6 giu 2012, 03.40.02, CEST 183209 2
mer 6 giu 2012, 03.50.05, CEST 183210
mer 6 giu 2012, 03.53.22, CEST 183211 1
mer 6 giu 2012, 04.08.50, CEST 183212
mer 6 giu 2012, 04.09.17, CEST 183213
mer 6 giu 2012, 04.17.04, CEST 183214
mer 6 giu 2012, 04.33.05, CEST 183215 2
mer 6 giu 2012, 04.45.21, CEST 183216
mer 6 giu 2012, 04.53.35, CEST 183217
mer 6 giu 2012, 05.02.56, CEST 183218 2
mer 6 giu 2012, 05.05.56, CEST 183219 61
mer 6 giu 2012, 05.12.15, CEST 183220 120
mer 6 giu 2012, 05.29.34, CEST 183221 83
mer 6 giu 2012, 05.47.29, CEST 183222 2
mer 6 giu 2012, 06.01.51, CEST 183223 232
mer 6 giu 2012, 06.18.14, CEST 183224
mer 6 giu 2012, 06.26.22, CEST 183225
mer 6 giu 2012, 06.33.12, CEST 183226 1
mer 6 giu 2012, 06.50.42, CEST 183227 283
mer 6 giu 2012, 07.03.33, CEST 183228
mer 6 giu 2012, 07.22.01, CEST 183229 2
mer 6 giu 2012, 07.46.40, CEST 183230 278
mer 6 giu 2012, 07.49.40, CEST 183231 1
mer 6 giu 2012, 08.01.14, CEST 183232 1
mer 6 giu 2012, 08.08.51, CEST 183233 1
mer 6 giu 2012, 08.13.09, CEST 183234 1
mer 6 giu 2012, 08.39.55, CEST 183235 282
mer 6 giu 2012, 08.53.52, CEST 183236
mer 6 giu 2012, 08.59.29, CEST 183237
mer 6 giu 2012, 08.59.52, CEST 183238
mer 6 giu 2012, 09.00.11, CEST 183239
mer 6 giu 2012, 09.06.03, CEST 183240 2
mer 6 giu 2012, 09.26.17, CEST 183241 2
mer 6 giu 2012, 09.39.51, CEST 183242 3
mer 6 giu 2012, 09.47.00, CEST 183243 1
mer 6 giu 2012, 09.57.54, CEST 183244 4
mer 6 giu 2012, 09.59.59, CEST 183245 1
mer 6 giu 2012, 10.08.26, CEST 183246 1
mer 6 giu 2012, 10.10.09, CEST 183247 23
mer 6 giu 2012, 10.13.06, CEST 183248 45
mer 6 giu 2012, 10.15.27, CEST 183249 39
mer 6 giu 2012, 10.23.24, CEST 183250 168
mer 6 giu 2012, 10.28.59, CEST 183251 8
mer 6 giu 2012, 10.36.42, CEST 183252 127
mer 6 giu 2012, 10.37.34, CEST 183253 9
mer 6 giu 2012, 10.44.30, CEST 183254 99
mer 6 giu 2012, 11.26.05, CEST 183255 263
mer 6 giu 2012, 11.50.24, CEST 183256 264
mer 6 giu 2012, 11.51.24, CEST 183257 3
mer 6 giu 2012, 11.52.56, CEST 183258 18
mer 6 giu 2012, 11.55.56, CEST 183259 50
mer 6 giu 2012, 12.01.29, CEST 183260 88
mer 6 giu 2012, 12.12.59, CEST 183261 52
mer 6 giu 2012, 12.17.24, CEST 183262 79
mer 6 giu 2012, 12.21.53, CEST 183263 83
mer 6 giu 2012, 12.27.07, CEST 183264 90
mer 6 giu 2012, 12.48.15, CEST 183265
mer 6 giu 2012, 12.50.39, CEST 183266 1
mer 6 giu 2012, 12.58.17, CEST 183267
mer 6 giu 2012, 13.03.40, CEST 183268 84
mer 6 giu 2012, 13.07.27, CEST 183269 60
mer 6 giu 2012, 13.31.08, CEST 183270 78
mer 6 giu 2012, 13.31.35, CEST 183271 3
mer 6 giu 2012, 13.42.22, CEST 183272 3
mer 6 giu 2012, 13.43.48, CEST 183273
mer 6 giu 2012, 13.48.07, CEST 183274 1
mer 6 giu 2012, 14.03.59, CEST 183275 270
mer 6 giu 2012, 14.13.01, CEST 183276 150
mer 6 giu 2012, 14.17.56, CEST 183277 77
mer 6 giu 2012, 14.23.21, CEST 183278 1

Thanks in advance, and best regards.

        Bye, Lem

Tycho keeps answering that no shares are lost. Well, but for prop rewards this is not enough.

The problem is that if his system makes my shares fly from a short round to a longer one, I lose
bitcoins. If the contrary happens, I gain bitcoins. This is because I have chosen *proportional* reward. Obviously people that go with PPS are not affected.

In other words: I do pool-hopping, so I want my shares to stay exactly
where I put them: at the beginning of rounds. I don't want them to fly
from block to block, from round to round or whoever knows where they go.

If I can't be sure that my shares are accounted for in the round
they belong to, then I don't think deepbit is reliable for prop reward. Period.
Bitcoin Forum>Bitcoin>Mining>Pools
Jump to: