Research talk:Are the bots really fighting/Work log/2017-03-09

Thursday, March 9, 2017 edit

Getting the bot2bot revert table loaded into quarry. Check this out: https://quarry.wmflabs.org/query/17229

page_namespace reverts self_reverts mean_time_to_revert
0 84804 0.0 6278944.66728
1 10856 0.0 5441352.79136
2 1389 0.0 7148045.10151
3 3372 0.0 1580921.13345
4 31309 0.0 81432.8155482
5 355 0.0 3315169.86761
6 218 0.0 7959297.8578
7 9 0.0 1923545.22222
10 1326 0.0 6390434.58145
11 215 0.0 4377105.66977
12 2 0.0 3667504.5
14 28753 0.0 8274315.06431
15 4 0.0 5814180.75
100 27 0.0 5304038.77778
101 3 0.0 4141986.0
108 4 0.0 20.25
109 37 0.0 8385960.0
118 5 0.0 3337868.2

It looks like I must have filtered out self-reverts in a previous step. The mean time to revert in the main namespace is roughly 72 days! I bet that if we look for reverts that happen quickly, we'll find some revert events that look more like actual "conflict". --EpochFail (talk) 20:42, 9 March 2017 (UTC)Reply


I checked out the time to revert distribution. https://quarry.wmflabs.org/query/17231

time_to_revert_block COUNT(*) COUNT(DISTINCT rev_user) COUNT(DISTINCT reverting_user)
< hour 1476 30 24
> year 27022 29 26
day-week 7153 35 33
hour-day 9189 32 37
month-year 32631 49 40
week-month 7333 41 34

It looks like 1476/84804=1.7% happen within less than an hour. --EpochFail (talk) 20:51, 9 March 2017 (UTC)Reply


Here's a query that gets the reverts that happened within less than an hour. A quick spot check of a few suggests that even these aren't conflict. https://quarry.wmflabs.org/query/17233 It looks like this will be a good dataset for spot-checking for "conflict"y bot2bot reverts.


It looks like RussBot and Cydebot are some of the most prolific quick-reverters. https://quarry.wmflabs.org/query/17234

rev_user_text reverting_user_text COUNT(*)
RussBot Cydebot 515
AvicBot EmausBot 241
EmausBot AvicBot 172
Xqbot AvicBot 117
AvicBot Xqbot 87
Xqbot EmausBot 69
EmausBot Xqbot 67
ImageRemovalBot 718 Bot 61
ArmbrustBot Cydebot 40
Citation bot ArticlesForCreationBot 12
AvicBot VoxelBot 10
VoxelBot AvicBot 10
AvicBot タチコマ robot 7
ArticlesForCreationBot Citation bot 6
Xqbot Thehelpfulbot 5
Cydebot RussBot 4
AnomieBOT CheMoBot 3
AvicBot Thehelpfulbot 3
EmausBot Chobot 3
MerlLinkBot Thehelpfulbot 2
RussBot OKBot 2
Citation bot 1 AnomieBOT 2
Citation bot 1 Citation bot 4 2
Cyberbot II ClueBot NG 2
AvicBot CeraBot 2
Makecat-bot Xqbot 2
CheMoBot AnomieBOT 2
ClueBot NG AvicBot 2
CorenSearchBot Mr.Z-bot 1
Mathbot Xqbot 1
EmausBot DarafshBot 1
Bibcode Bot Citation bot 1
FrescoBot Thehelpfulbot 1
JackieBot Xqbot 1
BG19bot AnomieBOT 1
Citation bot 2 CheMoBot 1
JackieBot Thehelpfulbot 1
タチコマ robot AvicBot 1
JackieBot EmausBot 1
BattyBot Addbot 1
DixonDBot Chobot 1
MadmanBot Citation bot 1
Xqbot JackieBot 1
DixonDBot EmausBot 1
Makecat-bot EmausBot 1
Yobot Mathbot 1
ClueBot NG CheMoBot 1
LyricsBot AnomieBOT 1
Yobot Xqbot 1
ClueBot NG Xqbot 1
EmausBot Thehelpfulbot 1
ClueBot NG Thehelpfulbot 1
MerlLinkBot Chobot 1
CheMoBot Citation bot 2 1

--EpochFail (talk) 21:03, 9 March 2017 (UTC)Reply

Return to "Are the bots really fighting/Work log/2017-03-09" page.