A while back, we conducted a predicting contest called "How Well Do You Know Your BBWAA?" 16 BJOL members (including me) submitted entries where we predicted the % of the vote that each of the candidates on the 2022 Hall of Fame ballot would receive. The results of the balloting were announced this past week, and, as you probably heard, only David Ortiz cleared the 75% threshold necessary for election.
The scoring of our contest took the absolute difference of the actual percentage vs. the predicted percentage for each candidate on a member’s entry to come up with prediction errors. The errors were then totaled and averaged for all candidates on each member’s entry. The person with the lowest average error was the winner.
Below are the results, pending any challenges or corrections that may emerge if anyone spots any. By the way, I included an "average" entry….basically an entry that consisted of the averages of everyone else’s predictions, and then calculated the average error that that fictional entry would have generated.
Our winner for the 5th straight year (!) is Chris Bodig, with an average error of 2.8 points per prediction. Chris now sits alongside the 1949-1953 Yankees in the pantheon of great title runs. Once again, I did pretty well among the "non-Bodig" entries.
The full list of results:
Member
|
Avg. Error
|
chrisbodig
|
2.8
|
Dan
|
3.1
|
Patrick
|
3.1
|
SaKish
|
4.0
|
tigerlily
|
4.0
|
DaveNJnews
|
4.4
|
bdhopkin
|
4.6
|
Entry of average vote predictions
|
4.6
|
jaycuda
|
4.8
|
stevemaxon
|
5.2
|
patrickbconnors (Regan)
|
5.4
|
evanecurb
|
5.5
|
rtallia
|
5.6
|
the_third_angel
|
5.8
|
terrific
|
6.7
|
sfnewhall
|
11.2
|
r44fletch
|
22.9
|
Below a table listing the candidates in order of actual % of the vote received. It also compares the actual results for each candidate vs. the average prediction (that is, the average of all of the entries) and the corresponding error (in percentage points).
I also included other figures including the high prediction, the low prediction, the range (high minus low prediction), the standard deviation of the predictions, as well as data comparing to last year’s actual vote percentages, including the change vs. last year’s results that our group predicted:
Name
|
Actual Vote %
|
Avg. Prediction
|
Error
|
High
|
Low
|
Range
|
Std Dev
|
LY Actual
|
Predicted Change
|
Actual Change
|
David Ortiz
|
77.9
|
71.6
|
6.3
|
81.0
|
50.0
|
31.0
|
8.4
|
n/a
|
n/a
|
n/a
|
Barry Bonds
|
66.0
|
66.3
|
0.3
|
76.0
|
52.0
|
24.0
|
5.3
|
61.8
|
4.5
|
4.2
|
Roger Clemens
|
65.2
|
66.6
|
1.4
|
78.0
|
58.0
|
20.0
|
5.1
|
61.6
|
5.0
|
3.6
|
Scott Rolen
|
63.2
|
59.1
|
4.1
|
71.0
|
23.0
|
48.0
|
11.5
|
52.9
|
6.2
|
10.3
|
Curt Schilling
|
58.6
|
69.1
|
10.5
|
78.0
|
45.0
|
33.0
|
7.8
|
71.1
|
(2.0)
|
(12.5)
|
Todd Helton
|
52.0
|
54.4
|
2.4
|
75.0
|
40.0
|
35.0
|
7.5
|
44.9
|
9.5
|
7.1
|
Billy Wagner
|
51.0
|
47.4
|
3.6
|
60.0
|
7.0
|
53.0
|
12.5
|
46.4
|
1.0
|
4.6
|
Andruw Jones
|
41.1
|
41.3
|
0.2
|
60.0
|
29.0
|
31.0
|
9.3
|
33.9
|
7.4
|
7.2
|
Gary Sheffield
|
40.6
|
45.3
|
4.7
|
75.0
|
21.0
|
54.0
|
10.7
|
40.6
|
4.7
|
0.0
|
Alex Rodriguez
|
34.3
|
52.1
|
17.8
|
76.0
|
30.0
|
46.0
|
12.6
|
n/a
|
n/a
|
n/a
|
Jeff Kent
|
32.7
|
35.9
|
3.2
|
60.0
|
25.0
|
35.0
|
8.0
|
32.4
|
3.5
|
0.3
|
Manny Ramirez
|
28.9
|
31.6
|
2.7
|
67.0
|
15.0
|
52.0
|
10.3
|
28.2
|
3.4
|
0.7
|
Omar Vizquel
|
23.9
|
43.7
|
19.8
|
55.0
|
21.0
|
34.0
|
7.6
|
49.1
|
(5.4)
|
(25.2)
|
Sammy Sosa
|
18.5
|
21.0
|
2.5
|
54.0
|
12.0
|
42.0
|
9.6
|
17.0
|
4.0
|
1.5
|
Andy Pettitte
|
10.7
|
20.3
|
9.6
|
65.0
|
10.0
|
55.0
|
15.3
|
13.7
|
6.6
|
(3.0)
|
Jimmy Rollins
|
9.4
|
14.4
|
5.0
|
52.0
|
1.0
|
51.0
|
13.9
|
n/a
|
n/a
|
n/a
|
Bobby Abreu
|
8.6
|
12.3
|
3.7
|
28.0
|
5.0
|
23.0
|
5.5
|
8.7
|
3.6
|
(0.1)
|
Mark Buehrle
|
5.8
|
10.9
|
5.1
|
18.0
|
2.0
|
16.0
|
4.2
|
11.0
|
(0.1)
|
(5.2)
|
Torii Hunter
|
5.3
|
9.1
|
3.8
|
15.0
|
4.0
|
11.0
|
2.6
|
9.5
|
(0.4)
|
(4.2)
|
Joe Nathan
|
4.3
|
3.1
|
1.2
|
5.0
|
1.0
|
4.0
|
1.4
|
n/a
|
n/a
|
n/a
|
Tim Hudson
|
3.0
|
7.0
|
4.0
|
40.0
|
2.0
|
38.0
|
8.9
|
5.2
|
1.8
|
(2.2)
|
Tim Lincecum
|
2.3
|
3.8
|
1.5
|
22.0
|
0.0
|
22.0
|
5.9
|
n/a
|
n/a
|
n/a
|
Ryan Howard
|
2.0
|
7.6
|
5.6
|
55.0
|
0.0
|
55.0
|
13.8
|
n/a
|
n/a
|
n/a
|
Mark Teixeira
|
1.5
|
6.4
|
4.9
|
31.0
|
1.0
|
30.0
|
8.2
|
n/a
|
n/a
|
n/a
|
Jonathan Papelbon
|
1.3
|
5.9
|
4.6
|
60.0
|
0.0
|
60.0
|
14.8
|
n/a
|
n/a
|
n/a
|
Justin Morneau
|
1.3
|
2.3
|
1.0
|
12.0
|
0.0
|
12.0
|
3.9
|
n/a
|
n/a
|
n/a
|
Prince Fielder
|
0.5
|
1.8
|
1.3
|
12.0
|
0.0
|
12.0
|
3.0
|
n/a
|
n/a
|
n/a
|
A.J. Pierzynski
|
0.5
|
1.8
|
1.3
|
21.0
|
0.0
|
21.0
|
5.2
|
n/a
|
n/a
|
n/a
|
Jake Peavy
|
0.0
|
4.7
|
4.7
|
50.0
|
0.0
|
50.0
|
12.3
|
n/a
|
n/a
|
n/a
|
Carl Crawford
|
0.0
|
0.9
|
0.9
|
5.0
|
0.0
|
5.0
|
1.3
|
n/a
|
n/a
|
n/a
|
Observations:
· As a group, we missed pretty badly on Omar Vizquel (almost a 20 percentage point miss), Alex Rodriguez, and Curt Schilling. With both Vizquel and Schilling, we did expect both of them to lose support based on their respective situations (which I will assume you are likely familiar with), but we significantly underestimated just how much they would decline.
Rodriguez debuted on this year’s ballot, and we overestimated his initial level of support pretty significantly. We expected he’d get a little over half of the vote, but he was closer to one-third.
· On the flip side, we underestimated Ortiz. As a group, we predicted around 72 percent support (which would have had him just missing getting elected), where as Ortiz actually received close to 78%, and was inducted in his first year on the ballot.
· Our best group pick (the one with the lowest error) was Andruw Jones. Our average prediction was 41.3%, and he actually received 41.1%. We were also very close on Barry Bonds as the average group prediction was 66.3%, and Bonds actually received 66.0%.
· Scott Rolen was the big winner this year, as his vote total improved from 52.9% last year to 63.2% (an improvement of 10.3 percentage points), setting himself up nicely for a probable induction within the next couple of years. Andruw Jones and Todd Helton both improved by more than 7 percentage points over last year, although they aren’t as close to induction as Rolen is. Still some nice gains for those two.
Overall, our group predicted 8.2 names per ballot. The actual voting ended up at 7.1 per ballot, which is frankly a pretty poor performance – in most years, we’re normally only off by about 0.1 or so.
So congratulations to Chris Bodig. Again.
Thanks to all for participating.
Dan