A computer script designed by Dragons flight was used to parse 100 days of AFD logs from June 12005 - September 82005 searching for bolded keywords (e.g. delete, keep, merge, redirect, kill, cleanup, etc) in signed comments. This has allowed a large statistical sample to be generated from which important patterns in voting and article deletion behavior might be identified.
These were broken into groups by searching for the === Nomination Headers ===. The first link to User space after the nomination header was assumed to belong to the nominator and was recorded in a separate category.
Obviously, there are many ways that this script can be fooled if people format things in unusual ways (or even some fairly mundane variations), but the hope is that by capturing a large enough sample it will be possible to derive meaningful patterns even if people who forget to bold their vote or sign in an unusual place are ignored. (Note: if more than one link to User space was present, the last one on the line was assumed to be the signature.)
The "vote text" was interpreted by removing a long list of modifiers (e.g. strong, weak, super, borderline, massive, etc.) and creating lists of common synonyms (e.g. delete = kill, nuke, destroy; keep = cleanup, revise, expand, don't delete). In this way it was possible to categorize 96% of all vote text as either: keep, delete, merge, speedy, speedy keep, redirect, bjaodn, rename, transwiki, or comment. The remaining 4% consist of a variety of ambiguous statements that could not be interpreted and rarely used phrases (e.g. "grind into a pulp") that were not considered frequent enough to be worth teaching to the parser even though their meaning may have been clear. If someone used multiple keywords, e.g. "delete or merge", the vote was usually recorded based on the first occurring recognized word. Some inversion terms (e.g. "don't" in "don't delete") were also processed to handle exceptions where the keyword wasn't assigned its normal meaning.
Obviously, this can never be as accurate or as complete as someone processing the AFD votes by hand, and it is likely a variety of mistakes and misinterpretations were made, but I believe this methodology is more than sufficient to get a broad understanding of AFD patterns.
Overview patterns
Number of days:
100
Range of days:
2005 June 1 - 2005 September 8
Number of nominations:
11211
Number of voters:
7202
Number of votes:
86773
Number of closers:
170
Percent deleted:
75.2 %
Empirical threshold:
63.5 %
Consistent with threshold:
93.8 %
Percent deleted includes content removal outcomes (delete, speedy, bjaodn and redirect) as compared to content retaining outcomes (keep, no consensus, merge, speedy keep, move, and transwiki).
Full voting patterns
Category
# of Voters
votes
delete
keep
merge
speedy
sp. keep
redirect
bjaodn
move
transwiki
nominate
comment
other
All Users
7202
86773
47.5%
21.9%
3.9%
2.4%
0.3%
2.9%
0.2%
0.7%
0.4%
12.9%
3.7%
3.1%
At least 250 votes
51
27957
59.4%
16.6%
4.2%
2.8%
0.3%
3.4%
0.2%
0.6%
0.4%
6.5%
3.2%
2.6%
Between 100 and 250 votes
129
21055
47.8%
19.2%
3.7%
3.0%
0.4%
3.3%
0.3%
0.5%
0.4%
14.4%
3.9%
3.0%
Between 20 and 100 votes
451
20500
44.9%
20.2%
3.8%
2.5%
0.4%
2.7%
0.3%
0.8%
0.4%
16.6%
4.3%
3.1%
Less than 20 votes
6570
17261
31.3%
35.7%
3.6%
1.1%
0.3%
1.9%
0.2%
0.8%
0.3%
17.1%
3.7%
3.9%
Condensed voting patterns
Category
# of Voters
votes
delete
keep
All Users
6912
79077
70.6%
29.4%
At least 250 votes
46
23343
76.0%
24.0%
Between 100 and 250 votes
121
19615
74.0%
26.0%
Between 20 and 100 votes
430
19578
72.7%
27.3%
Less than 20 votes
6314
16541
56.2%
43.8%
Content removal options (delete, speedy, nominate, bjaodn and redirect) consolidated under "delete".
Content preserving options (keep, merge, speedy keep, move, and transwiki) consolidated under "keep".
Comments and unparsed options are ignored and removed from counts.
Expanded low voter count
Category
Voters
Votes
Delete
Keep
Voted 20 to 49 times
276
8837
68.3%
31.7%
Voted 10 to 19 times
369
5020
66.0%
34.0%
Voted 5 to 9 times
591
3855
61.1%
38.9%
Voted 2 to 4 times
1476
3787
54.8%
45.2%
Voted once
3878
3878
40.2%
59.8%
Same as above but for the very infrequent voters. Done on request of User:Fubar Obfusco.
Deletionist vs. Inclusionist tendencies
Expressed in terms of how often they vote delete, this table summarizes the tendencies of AFD regulars.
Deletion Percentage
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
Users with more than 250 votes
1
1
0
1
1
6
0
13
12
11
Users with more than 100 votes
4
3
2
3
7
13
9
39
53
34
Users with more than 20 votes
17
9
13
17
28
48
71
126
170
98
AFD outcomes
Outcomes
Number
Percentage
delete
6721
60.0 %
keep
2201
19.6 %
merge
296
2.6 %
speedy
805
7.2 %
speedy keep
27
0.2 %
redirect
360
3.2 %
bjaodn
8
0.1 %
move
36
0.3 %
transwiki
42
0.4 %
uncertain
715
6.4 %
"uncertain" represents all of the AFDs whose outcome the program was unable to parse.
no consensus results are included under keep.
Note: Combining the content removal options (delete, speedy, redirect, and bjaodn) and discounting the 6.4% of "uncertain" conclusions, indicates that 75.2% of AFDs result in content being "deleted", versus 24.8% with content preserving conclusions (keep, speedy keep, merge, move, transwiki).
User patterns
The voting patterns for the top 30 participants on AFD.
See also: An expanded list for all participants averaging more than one vote per day.
Estimated threshold is the percentage of delete votes this person most usually requires before closing an AFD as a deletion. This threshold can be significantly distorted for closers that avoid controversial votes (see deviation below).
Adherence is the fraction of closes performed that appear consistent with this admin's threshold.
Estimated deviation is the estimated number of AFD results that would have to be changed if this admin adopted the 63.5% threshold which is the average. Admins who rarely close controversial votes (e.g. nothing in the 50-80% range) may have a deviation of 0 even if their estimated threshold is substantially displaced.
Article voting patterns
Category
AFDs
Percentage
Unanimous delete
6523
58.2%
Strong delete (>80%), not unanimous
855
7.6%
Others (50-80% delete)
1932
17.2%
Majority keep (>50%), not strong
803
7.2%
Strong keep (>70%), not unanimous
463
4.1%
Unanimous keep
635
5.7%
Deletion as a function of vote percentage
Delete %
occurrences
Deleted
Percentage
0 - 5 %
18
0
0.0 %
5 - 10 %
138
0
0.0 %
10 - 15 %
188
2
1.1 %
15 - 20 %
318
9
2.8 %
20 - 25 %
248
6
2.4 %
25 - 30 %
134
3
2.2 %
30 - 35 %
279
14
5.0 %
35 - 40 %
237
17
7.2 %
40 - 45 %
120
6
5.0 %
45 - 50 %
344
46
13.4 %
50 - 55 %
64
9
14.1 %
55 - 60 %
289
82
28.4 %
60 - 65 %
102
48
47.1 %
65 - 70 %
320
208
65.0 %
70 - 75 %
417
328
78.7 %
75 - 80 %
328
307
93.6 %
80 - 85 %
249
245
98.4 %
85 - 90 %
431
420
97.4 %
90 - 95 %
135
130
96.3 %
95 - 100 %
6137
6014
98.0 %
No compensation for sockpuppets / anon votes
Some of the more perverse results also reflect parser error. For example, some of the 2% of AFDs that were kept despite apparent unanimous delete resulted from strangely formatted or labeled keep votes that the parser was unable to count.
Votes per article
Votes received
AFDs
Frequency
1
391
3.5%
2
635
5.7%
3
1204
10.7%
4
1582
14.1%
5
1483
13.2%
6
1212
10.8%
7
994
8.9%
8
714
6.4%
9
528
4.7%
10
395
3.5%
11
352
3.1%
12
266
2.4%
13
206
1.8%
14
168
1.5%
15
126
1.1%
16
103
0.9%
17
111
1.0%
18
89
0.8%
19
74
0.7%
20
70
0.6%
21
62
0.6%
22
55
0.5%
23
44
0.4%
24
37
0.3%
25
29
0.3%
26
31
0.3%
27
24
0.2%
28
25
0.2%
29
18
0.2%
30
22
0.2%
31
11
0.1%
32
16
0.1%
33
13
0.1%
34
8
0.1%
35
9
0.1%
36
7
0.1%
37
5
0.0%
38
4
0.0%
39
8
0.1%
40
8
0.1%
41
6
0.1%
42
2
0.0%
43
7
0.1%
44
6
0.1%
45
4
0.0%
46
1
0.0%
47
3
0.0%
48
1
0.0%
49
5
0.0%
50
4
0.0%
51
1
0.0%
52
1
0.0%
54
2
0.0%
56
1
0.0%
57
1
0.0%
59
2
0.0%
61
1
0.0%
63
3
0.0%
64
2
0.0%
65
1
0.0%
67
1
0.0%
69
1
0.0%
70
2
0.0%
71
1
0.0%
73
1
0.0%
78
1
0.0%
79
1
0.0%
81
1
0.0%
82
1
0.0%
90
1
0.0%
96
1
0.0%
98
1
0.0%
114
1
0.0%
115
1
0.0%
167
1
0.0%
185
1
0.0%
192
1
0.0%
The nomination counts as one vote. Those AFDs with only 1 or 2 votes were typically speedy deleted shortly after listing.
Contested AFDs
Number of AFDs with at least the specified number of both keep and delete votes, and the percentage of such AFDs as a fraction of all AFDs
Votes
occurrences
Percentage
1
4688
41.8%
2
2643
23.6%
3
1614
14.4%
4
1062
9.5%
5
737
6.6%
6
515
4.6%
7
372
3.3%
8
273
2.4%
9
211
1.9%
10
167
1.5%
11
130
1.2%
12
110
1.0%
13
80
0.7%
14
65
0.6%
15
54
0.5%
16
46
0.4%
17
38
0.3%
18
29
0.3%
19
26
0.2%
20
25
0.2%
21
23
0.2%
22
21
0.2%
23
16
0.1%
24
14
0.1%
25
14
0.1%
26
12
0.1%
27
11
0.1%
28
10
0.1%
29
9
0.1%
30
9
0.1%
31
7
0.1%
32
6
0.1%
33
6
0.1%
34
5
0.0%
35
5
0.0%
36
5
0.0%
37
4
0.0%
38
2
0.0%
39
2
0.0%
40
2
0.0%
41
2
0.0%
42
2
0.0%
43
2
0.0%
44
2
0.0%
45
2
0.0%
46
2
0.0%
47
2
0.0%
48
2
0.0%
49
2
0.0%
50
2
0.0%
51
2
0.0%
52
2
0.0%
53
2
0.0%
54
2
0.0%
55
2
0.0%
56
2
0.0%
57
2
0.0%
58
2
0.0%
59
2
0.0%
60
1
0.0%
61
1
0.0%
62
1
0.0%
63
1
0.0%
64
1
0.0%
65
1
0.0%
66
1
0.0%
67
0
0.0%
AFDs with at least 2/3 delete
Shows the number of AFDs with at least the specified number of keep votes and those with at least twice as many delete votes as keep votes.