Vancouver Blog Awards Updates
This page describes how I updated the Vancouver Election Blog Awards, based on votes received. Starting 2008-05-20 we awarded a total of $300 per week to the top-voted blogs (skipping those indicated *, who are waiving their awards). The voting system design has changed several times, as described below.
2008-11-17:
Second week of approval voting. Ranking determined by number of approval votes received, with 5-week vote decay as described under 2008-09-08 below.
2008-11-10:
Changed to approval voting. (Had been voting dollar amounts.) See spreadsheet for details on votes and how awards are determined.
2008-11-03:
See spreadsheet for details on votes and how awards are determined.
2008-10-27:
See spreadsheet for details on votes and how awards are determined.
2008-10-16:
See spreadsheet for details on votes and how awards are determined.
2008-10-06:
See spreadsheet for details on votes and how awards are determined. This week the award cap was lowered from $150 to $100.
2008-09-29:
See spreadsheet for details on votes and how awards are determined.
2008-09-22:
See spreadsheet for details on votes and how awards are determined.
2008-09-15:
See spreadsheet for details on votes and how awards are determined.
2008-09-08:
See spreadsheet for details on votes and how awards are determined. I'm using a 5-week linear vote decay, which means: I count the last vote received from each IP address, weighted according to how long ago it was cast. Weight for the current week is 1.0, for the previous week 0.8, the week before that 0.6, then 0.4, then 0.2. Votes cast more than 5 weeks ago are not counted, i.e. weight = 0.0. This trades off two principles: (a) blog quality changes through time, so older votes are less relevant; and (b) we can't expect busy people to vote every week, so 1-week decay is too fast.
2008-09-01:
See spreadsheet for details on votes and how awards are determined.
2008-08-25:
See spreadsheet for details on votes and how awards are determined.
2008-08-18:
Second week of absolute dollar voting. Rankings are still shifting, but less dramatically than last week, so we seem to be reaching more of an equilibrium. See spreadsheet for details on votes and how awards are determined.
2008-08-11:
This was the first week of "absolute dollar voting". Rankings changed dramatically! See this spreadsheet for details on the votes and how awards are determined. I plan to blog about the results soon.
For 3 weekly revisions 2008-07-21 through 2008-08-04, we tried out "dollar voting" for the contestants receiving awards, as described in this blog post. This could also be called "relative dollar voting" to distinguish it from "absolute dollar voting" which we started after 2008-08-04.
2008-08-04:
Rankings changed quite a bit this week -- see #1 through #6 on dollar vote spreadsheet; #7 through #13 below:
| Media/Blog Name |
OldRank |
Up |
Stay |
Down |
NewRank |
| Price Tags* |
7 |
58% |
20% |
22% |
7 |
| In Support of Public Education* |
8 |
0% |
36% |
64% |
10 |
| Paul Hillsdon |
9 |
58% |
20% |
22% |
8 |
| David Talks/The Berner Monologues |
10 |
58% |
20% |
22% |
9 |
| Miss 604 |
11 |
0% |
36% |
64% |
13 |
| ursa minor: bear604 |
12 |
0% |
48% |
52% |
12 |
| Walter Schultz |
13 |
0% |
67% |
33% |
11 |
We're trying another version of dollar voting starting 2008-08-04 -- see blog.
2008-07-28:
We saw a couple more adjustments this week than last: Ranks #5 and #6 changed places (see dollar vote spreadsheet), as did #9 and #10 (see table below).
| Media/Blog Name |
OldRank |
Up |
Stay |
Down |
NewRank |
| Price Tags* |
7 |
43% |
43% |
14% |
7 |
| In Support of Public Education* |
8 |
14% |
57% |
29% |
8 |
| David Talks/The Berner Monologues |
9 |
33% |
50% |
17% |
10 |
| Paul Hillsdon |
10 |
57% |
29% |
14% |
9 |
| Miss 604 |
11 |
43% |
57% |
0% |
11 |
| ursa minor: bear604 |
12 |
0% |
71% |
29% |
12 |
| Walter Schultz |
13 |
25% |
50% |
25% |
13 |
I'm trying out a new dollar voting design this coming week.
2008-07-21:
We seem to be near an equilibrium for now: no rank changed this week. But this first week of dollar voting brought an incremental change: $10 less for #1 The Tyee, and $10 more for #3 David Eby. See calculations on dollar vote spreadsheet.
| Media/Blog Name |
OldRank |
Up |
Stay |
Down |
NewRank |
| Price Tags* |
7 |
3% |
67% |
30% |
7 |
| In Support of Public Education* |
8 |
31% |
43% |
26% |
8 |
| David Talks/The Berner Monologues |
9 |
33% |
67% |
0% |
9 |
| Paul Hillsdon |
10 |
3% |
67% |
30% |
10 |
| Miss 604 |
11 |
44% |
43% |
13% |
11 |
| ursa minor: bear604 |
12 |
1% |
84% |
15% |
12 |
| Walter Schultz |
13 |
4% |
57% |
39% |
13 |
Hopefully Frances Bula will rejoin us soon, and we'll see some more dramatic movement!
From 2008-05-20 through 2008-07-14, weekly awards to the top five were $100, $80, $60, $40, $20 (skipping those indicated *, who are waiving their awards). The ballot design and tallying method have evolved over time, as described below. From 2008-06-23 through 2008-07-14, because of the general bias for more Down votes than Up, I applied a simple across-the-board adjustment to equalize Down and Up totals. Then I focused on the median voter in each row, so only rows with "Up" or "Down" greater than 50% call for a move. I highlight them in bold, and rerank the blogs in response.
2008-07-14:
| Media/Blog Name |
OldRank |
Up |
Stay |
Down |
NewRank |
| The Vancouver Observer |
1 |
54% |
19% |
27% |
2 |
| The Tyee |
2 |
71% |
7% |
22% |
1 |
| David Eby |
3 |
49% |
9% |
42% |
3 |
| The Vancouver Manifesto |
4 |
38% |
5% |
57% |
5 |
| Bill Tieleman |
5 |
44% |
8% |
48% |
4 |
| Stephen Rees |
6 |
28% |
35% |
37% |
6 |
| Price Tags* |
7 |
28% |
24% |
48% |
7 |
| In Support of Public Education* |
8 |
28% |
23% |
49% |
8 |
| David Talks/The Berner Monologues |
9 |
16% |
52% |
32% |
9 |
| Paul Hillsdon |
10 |
33% |
30% |
37% |
10 |
| Miss 604 |
11 |
40% |
12% |
48% |
11 |
| ursa minor: bear604 |
12 |
24% |
38% |
38% |
12 |
| Walter Schultz |
13 |
29% |
26% |
45% |
13 |
Most of the ranks are at equilibrium this week, without a majority vote for a move. The Tyee's up move was strong enough to overtake the Observer's weaker up move.
Starting today there's a new ballot format, with "dollar voting".
2008-07-07:
| Media/Blog Name |
OldRank |
Up |
Stay |
Down |
NewRank |
| The Tyee |
1 |
20% |
60% |
20% |
2 |
| David Eby |
2 |
22% |
36% |
42% |
3 |
| The Vancouver Observer |
3 |
88% |
5% |
7% |
1 |
| City States by Frances Bula* |
4 |
27% |
29% |
44% |
|
| The Vancouver Manifesto |
5 |
44% |
5% |
51% |
4 |
| Stephen Rees |
6 |
42% |
6% |
52% |
6 |
| Bill Tieleman |
7 |
49% |
4% |
47% |
5 |
| Price Tags* |
8 |
37% |
6% |
57% |
7 |
| In Support of Public Education* |
9 |
44% |
5% |
51% |
8 |
| Paul Hillsdon |
10 |
33% |
13% |
54% |
10 |
| David Talks/The Berner Monologues |
11 |
7% |
50% |
43% |
9 |
Frances Bula has left the Vancouver Sun, so her City States blog will soon be replaced by her next blog. She'll let me know when it's launched, and I'll include it ASAP.
In three cases above (OldRank 5, 8 & 9), a small majority vote to "move down" wasn't big enough to overtake the next contestant below, which was also moving down. Then they moved up one rank because #4 dropped out.
Not enough votes on Page 2 to be confident of voters' wishes, so we'll accumulate more through next Monday.
2008-07-01:
(I was a day late with this week's update -- travelling all day Monday.)
| Media/Blog Name |
OldRank |
Up |
Stay |
Down |
NewRank |
| The Tyee |
1 |
22% |
57% |
21% |
1 |
| David Eby |
2 |
31% |
36% |
33% |
2 |
| The Vancouver Manifesto |
3 |
25% |
14% |
61% |
5 |
| City States* |
4 |
34% |
21% |
45% |
4 |
| Stephen Rees |
5 |
23% |
32% |
45% |
6 |
| Bill Tieleman |
6 |
35% |
15% |
50% |
7 |
| Price Tags* |
7 |
31% |
20% |
49% |
8 |
| In Support of Public Education* |
8 |
37% |
17% |
46% |
9 |
| Paul Hillsdon |
9 |
28% |
22% |
50% |
10 |
| Miss 604 |
10 |
35% |
13% |
52% |
12 |
| The Vancouver Observer [NEW] |
11 |
95% |
5% |
0% |
3 |
| ursa minor: bear604 |
12 |
30% |
70% |
0% |
13 |
| Walter Schultz |
13 |
22% |
58% |
20% |
14 |
Congrats to The Vancouver Observer for moving from #11 to #3 in its first week!
We welcome new contestant David Talks/The Berner Monologues. (Inserted at rank #11 at the end of ballot Page 1.)
2008-06-23:
| Media/Blog Name |
OldRank |
Up |
Stay |
Down |
NewRank |
| The Tyee |
1 |
20% |
54% |
26% |
1 |
| David Eby |
2 |
28% |
56% |
16% |
2 |
| The Vancouver Manifesto |
3 |
31% |
22% |
47% |
3 |
| City States* |
4 |
47% |
43% |
10% |
4 |
| Stephen Rees |
5 |
17% |
42% |
41% |
5 |
| In Support of Public Education* |
6 |
29% |
20% |
51% |
8 |
| Price Tags* |
7 |
39% |
13% |
48% |
7 |
| Bill Tieleman |
8 |
55% |
10% |
35% |
6 |
| Paul Hillsdon |
9 |
37% |
19% |
44% |
9 |
| Miss 604 |
10 |
42% |
11% |
47% |
10 |
| The Vancouver Observer [NEW] |
|
|
|
|
11 |
| ursa minor: bear604 |
12 |
33% |
67% |
0% |
12 |
| Walter Schultz |
13 |
33% |
67% |
0% |
13 |
This week the only reranking was to swap #6 with #8. Also, the Report Card blog dropped out voluntarily (taking a holiday; focus not ideal for this contest anyway).
We welcome new contestant The Vancouver Observer! (Linked from ballot.)
2008-06-16:
| Media/Blog Name |
OldRank |
More |
Less |
NewRank |
| The Tyee |
1 |
92% |
8% |
1 |
| David Eby |
2 |
50% |
50% |
2 |
| The Vancouver Manifesto |
3 |
60% |
40% |
3 |
| Stephen Rees |
4 |
40% |
60% |
5 |
| City States* |
5 |
50% |
50% |
4 |
| Bill Tieleman |
6 |
17% |
83% |
8 |
| Price Tags* |
7 |
0% |
100% |
7 |
| In Support of Public Education* |
8 |
50% |
50% |
6 |
| Paul Hillsdon |
9 |
50% |
50% |
9 |
| Report Card* |
10 |
33% |
67% |
11 |
| Miss 604 |
11 |
|
|
10 |
No one voted on Page 2 (ranks 12 & 13) this week.
Voters don't seem to like this pairwise ballot design, so I'm reverting to the Up/Stay/Down design we were using until 2008-06-02. We'll look into other ways of handling the downward bias I referred to below.
2008-06-10:
(I was a day late with this week's update -- travelling all day Monday.)
| Media/Blog Name |
OldRank |
More |
Less |
NewRank |
| David Eby |
1 |
37% |
63% |
2 |
| The Tyee |
2 |
78% |
22% |
1 |
| City States* |
3 |
46% |
54% |
5 |
| Stephen Rees |
4 |
47% |
53% |
4 |
| The Vancouver Manifesto |
5 |
54% |
46% |
3 |
| Price Tags* |
6 |
30% |
70% |
7 |
| Bill Tieleman |
7 |
65% |
35% |
6 |
| Miss 604 |
8 |
42% |
58% |
11 |
| Report Card* |
9 |
44% |
56% |
10 |
| Paul Hillsdon |
10 |
39% |
61% |
9 |
| In Support of Public Education*[NEW] |
11 |
100% |
0% |
8 |
| ursa minor: bear604 |
12 |
50% |
50% |
12 |
| Walter Schultz |
13 |
|
|
13 |
This was the first trial week for a pairwise ballot design. Instead of voting each contestant to move Up or Stay put or move Down, you vote that each contestant deserves More or Less than the next contestant down. So rerankings only happen when "Less" is over 50% (bolded).
There seem to be too many cases where a majority voted "Less", but I reranked obediently. Let's try this ballot design for at least one more week before trying to decide whether it's an improvement or not. Comments welcomed via my blog or email.
2008-06-02:
| Media/Blog Name |
OldRank |
Up |
Stay |
Down |
NewRank |
| David Eby |
1 |
40% |
39% |
21% |
1 |
| The Tyee |
2 |
30% |
33% |
37% |
2 |
| City States* |
3 |
25% |
20% |
55% |
3 |
| Stephen Rees |
4 |
22% |
17% |
61% |
4 |
| Public Eye Online |
5 |
26% |
30% |
44% |
|
| The Vancouver Manifesto |
6 |
27% |
20% |
53% |
5 |
| Price Tags* |
7 |
34% |
19% |
47% |
6 |
| Bill Tieleman |
8 |
25% |
25% |
50% |
7 |
| Miss 604 |
9 |
11% |
36% |
53% |
8 |
| Paul Hillsdon |
10 |
25% |
34% |
41% |
10 |
| Report Card* |
11 |
63% |
24% |
13% |
9 |
| Walter Schultz |
12 |
0% |
0% |
100% |
13 |
| ursa minor: bear604 |
13 |
58% |
9% |
33% |
12 |
Public Eye Online I dropped out at this point because I've had no response from the author. I'd be happy to add it again if the author wishes. Its focus is mostly provincial, so I hope it will be part of the B.C. contest I plan to start in November when municipal elections are done.
There seems to be a general downward bias in the voting, which I think is partly caused by poor ballot design. Many people voted for their 2 or 3 favourites to move up, and for all the others to move down. So I've changed the ballot design. The new pairwise design should also make the reranking more automatic, less subjective.
For this last reranking under the old system, I basically ignored the "Down" moves, except for the one strong 100% signal (which is reinforced by an "Up" on the next contestant anyway).
I added newcomer In Support of Public Education* as #11, which appears at the bottom of Page 1 and at the top of Page 2.
From 2008-05-15 through 2008-05-26, I focused on the median voter in each row, so only rows with "Up" or "Down" greater than 50% called for a move. I highlighted them in bold, and reranked the blogs in response to how strong the vote was.
I planned to develop a more objective standardized method for deciding how to rearrange the ranking based on these vote percentages. But meanwhile I used subjective judgment. The response of voters after the reranking would correct my judgment errors and help me develop a good method for reranking in the future.
2008-05-26:
| Media/Blog Name |
OldRank |
Up |
Stay |
Down |
NewRank |
| David Eby |
1 |
38% |
30% |
32% |
1 |
| The Tyee |
2 |
25% |
35% |
40% |
2 |
| City States* |
3 |
35% |
16% |
49% |
3 |
| Public Eye Online |
4 |
15% |
23% |
62% |
5 |
| The Vancouver Manifesto |
5 |
23% |
15% |
62% |
6 |
| Stephen Rees |
6 |
38% |
19% |
43% |
4 |
| Bill Tieleman |
7 |
21% |
24% |
55% |
8 |
| Miss 604 |
8 |
9% |
18% |
73% |
9 |
| Price Tags* |
9 |
81% |
6% |
13% |
7 |
| Paul Hillsdon |
10 |
45% |
30% |
25% |
10 |
Newcomer Report Card has also been added as #11 on Page 1. We haven't received enough votes on Page 2 to be confident of voters' wishes, so we'll accumulate more through next Monday.
2008-05-20:
| Media/Blog Name |
OldRank |
Up |
Stay |
Down |
NewRank |
| David Eby |
1 |
32% |
25% |
43% |
1 |
| City States* |
2 |
27% |
32% |
41% |
3 |
| The Tyee |
3 |
52% |
22% |
26% |
2 |
| The Vancouver Manifesto |
4 |
11% |
26% |
63% |
5 |
| Public Eye Online |
5 |
37% |
30% |
33% |
4 |
| Bill Tieleman |
6 |
40% |
25% |
35% |
7 |
| Stephen Rees |
7 |
52% |
16% |
32% |
6 |
| Miss 604 |
8 |
35% |
21% |
44% |
8 |
| Walter Schultz |
9 |
5% |
11% |
84% |
11 |
| Paul Hillsdon |
10 |
44% |
28% |
28% |
10 |
| Price Tags* |
11 |
65% |
10% |
25% |
9 |
| ursa minor: bear604 |
12 |
24% |
35% |
41% |
12 |
As planned, I'll now split the ballot into two pages -- #1 - #10 on Page 1, #11 etc on Page 2.
2008-05-15:
Here's how the votes tallied (Up, Stay, Down columns) since 2008-05-13 --
| Blog |
OldRank |
NewRank |
Up |
Stay |
Down |
| City States |
1 |
2 |
23% |
34% |
43% |
| The Tyee |
2 |
3 |
15% |
52% |
33% |
| The Vancouver Manifesto |
3 |
4 |
31% |
34% |
35% |
| The Public Eye |
4 |
5 |
16% |
30% |
54% |
| David Eby |
5 |
1 |
91% |
0% |
9% |
| Miss 604 |
6 |
8 |
19% |
19% |
62% |
| Walter Schultz |
7 |
9 |
6% |
33% |
61% |
| Stephen Rees |
8 |
7 |
32% |
30% |
38% |
| Bill Tieleman |
9 |
6 |
53% |
19% |
28% |
| Paul Hillsdon |
10 |
10 |
42% |
40% |
18% |
| Price Tags |
11 |
11 |
29% |
50% |
21% |
| ursa minor: bear604 |
12 |
12 |
29% |
54% |
17% |
Today's reranking was less drastic than the previous one (2008-05-13 below), as we're getting closer to an equilibrium reflecting voter preferences. I expect the next reranking will be even less drastic. Of course there are short-term fluctuations caused by blog postings (e.g. at City States and David Eby), but those too will calm down somewhat as we get more voters and votes.
2008-05-13:
I started the ranking in late April / early May 2008, in a fairly random order. The ballot is designed for incremental updates (e.g. switch two neighbouring ranks, like a tennis ladder). But for the first update only, on May 13, I did a comprehensive reordering based on votes to that date, since the initial ranking was so uninformed.
I used ordinary least-squares regression, so if you're familiar with that you can follow the calculations on this spreadsheet. Feel free to contact me with questions. One of the voters suggested adding Vancouver Blog Miss 604, so I inserted it in the middle position, #6. (Insertion points for future additions are likely to be lower.)
Future ranking updates will probably be incremental. I'll design and tweak the algorithm as we go along, and may base it on Median Voter Theory. |