answered question

0
Votes

2
Answers

M$1.00

adamnelson | April 30, 2009 04:55 PM | in Probability

How can the probabilities over time of non-exclusive outcomes of an event happening be represented?

I'd like to represent, based on votes over time, the probability that different outcomes will happen. For instance:

Event: The New York Marathon in September
Outcome 1: David Wins
Outcome 2: John Wins
Outcome 3: David beats his personal best

People vote on whether they think each outcome will happen and the computer does not know which outcomes are exclusive (1 and 2) and which ones are non-exclusive (1 or 2 and 3) so the assumption is that they are all non-exclusive. These votes are timestamped and I want to represent the probabilities to an end user. My current thinking is a line chart with a line for each outcome and just showing the sum of the votes at every given time in the past for that outcome.

Any other thoughts from an Edward Tufte type direction?

Report

| RSS

| Share

answers (2)

0
Votes

BEST ANSWER decided by votes | howulikethemapples | May 01, 2009 03:42 AM

Whoa...I would suggest a line graph, on the x-axis put time, on the y-axis put the percentage likely (on a scale of 0-100%), of the number of people saying 'yes' to the assumption that the event would occur. But that doesn't account for exclusivity of factors...how important is that?

Tip howulikethemapples for this answer | Permalink | Report

Voted as best: bbrookin, lesliec

Comment

0
Votes

morriss003 | April 30, 2009 05:14 PM

Maybe a 3D representation of a cube graph and then different colored lines to represent the different runners, where x equals time, y equals number of votes for outcome a and z represents the number of votes for outcome b

Tip morriss003 for this answer | Permalink | Report

Voted as best: pensivefox

Comment