Box office predictions database
RS and MTC are no longer taggable new (see the Movies board for discussion) so I've decided to move my tracking of RS, MTC etc into the cloud and have it publicly available for anyone who wants the info.
Here's a link to a draft SS - anyone with this link can check it out
https://docs.google.com/spreadsheet/ccc?key=0AozvDcPvd7aLdGx1NUxyVXRyaE1tMkg0VkVzeEl0aHc#gid=4
I've opened the permissions for comments, so please comment in the SS and not here. Only comments of substance, please.
I'll put details into an HSXsanity column later this month, but it has 5 sheets
This Week, which shows the predictions for this week, and estimated adjusts and PSAs based on those predictions
Data, which has the historical information for predictions, going back to when I started tracking in Jan 2010
Accuracy, which shows the historical accuracy of each predictor
Accuracy Summary, which shows the average accuracy for predicting the last 50 movies
EAP God, which shows what each predictor would score if they were playing this season of EAP God against each other
The fine print:
I've chosen to include ten predictors - RS, MTC, LAT, Variety, THR, EW, boxoffice.com, CS, BOG and BOM. I'm open to including one or two more if they are truly of value to the HSX community, but I don't want to have every predictor with a website as it just ends up being a lot of noise (and a lot of work).
We all know how vague box office predictions can be. As a general rule, I've interpreted low 20s as 22, mid 20s as 25, high 20s as 28. But there are lots of permutations - just over 20, mid-to-high 20s, very high 20s, almost 30, just under 30, could reach 30. Anyway, by no means do I think these interpretations are perfect, but they should be good enough.
Five day releases - we have two this week - are a hassle. What I do is guess the Wed-Thur BO myself, and subtract that from the prediction. When actuals become available, I subtract those from the prediction. Again, not perfect, but good enough.
I also include the moviestock price at reset on opening day to gauge our accuracy as a prediction market. I use this figure because it is before we have any good info on midnights, etc and the price is easy to verify. However, the predictions before Thanksgiving 2010 have the price at halt.