To make it easy for HSX traders to find current and historical information on box office predictions, HSXsanity has made its box office prediction spreadsheet available at this link.
The spreadsheet includes predictions from ten different sources - RS, MTC, the Los Angeles Times, Variety, the Hollywood Reporter, Entertainment Weekly, boxoffice.com, Coming Soon, Box Office Guru and Box Office Mojo - with data going back to January 2010. The first five sources were selected because they provide predictions based on tracking, the second group of five were selected because of the high quality of their analysis.
In addition, the spreadsheet includes the moviestock price for each opener, taken at reset on its opening day (except for movies opening before Thanksgiving weekend in 2010, which have the price at halt.) This is to provide a gauge for our own accuracy as a prediction market - and also a check on our complaints that other predictors are often wrong.
There are six sheets on the spreadsheet:
- This Week, which shows the predictions for this week, and estimated adjusts and PTAs based on those predictions
- Data, which has the historical information for predictions
- Accuracy, which shows the historical accuracy of each prediction, based on weekend estimates
- RS/MTC rule, which shows whether the RS/MTC rule has worked for each prediction
- Accuracy Summary, which shows the average accuracy for each predictor for the last 50 wide openers, as well as the accuracy of the RS/MTC rule
- EAP God, which shows what each predictor would score if they were playing this season of EAP God against each other. (EAP God is a box office prediction game you can check out here where you try to predict the HSX adjust price for each opener. And if you already play, you can compare your score against the predictors)
As you would expect, there is some fine print regarding this spreadsheet and how it has interpreted data:
- Tracking information is often vaguely reported, and many predictions have had to be estimated. For example, high teens is estimated as 18, mid teens as 15, low teens as 13, low double digits as 11 and so on.
- Tracking information does not always take holiday weekends into account. For example, a tracking provider may make a three-day prediction for a four-day weekend. When this occurs, a four-day prediction is approximated by applying an appropriate multiplier to the three-day prediction.
- Similarly, for mid-week releases, there may be a five day-prediction instead of a three-day. When this occurs, weekday box office is guesstimated, and subtracted that from the prediction to approximate a three-day prediction. When actual numbers become available for the weekday box office, those figures replace the original guesstimate and the three-day prediction is re-calculated.
- For calculations related to the RS/MTC rule, the rule is "triggered" if there is a 20 per cent difference between RS and MTC's predictions, and MTC is considered to be "close" if its prediction is within 20 per cent of the weekend estimate. If you'd like more information about the RS/MTC rule, you can check out this column.
None of these interpretations provide exact predictions, but they are good enough to be workable.
Finally, if you are going to pay attention to the Accuracy Summary sheet, you should also read this column about weekend predictions and how the value is not really in the numbers, but the analysis that comes along with it. Similarly, if you're going to pay attention to tracking, you should also read this column about why predictions based on tracking are so often inaccurate.
This comment has been removed by the author.
ReplyDeleteThis comment has been removed by the author.
DeleteWhat's the "Br/even Friday Reset" column?
ReplyDeleteIt stands for "Breakeven Friday reset". It the prediction of the HSX market for the opening weekend, based on the price at reset on Friday morning.
ReplyDeleteRoger - are you available to walk us through this data or provide some guidance on where to get the historical HSX stock price by release?
DeleteHi there - I've been looking for the raw HSX historical data - doesn't anyone know how to acquire that?
ReplyDeleteDid you find Any?
ReplyDelete