Playing & Trading Strategy

Some of you may remember the chart of the week column that I used to write on HSXsanity, which compared the box office performance of current releases against historical comparisons and attempted to make a delist prediction. That column was discontinued because it was too time-consuming to write, but I've maintained the tool I used to create the charts.

With the changes at HSX, the removal of the adjust and the questions that many traders have about how to value moviestocks, I've decided to make that tool public to view.  You can check it out at this link.

Overview

Very quickly, new box office data is entered on the 'In Release' sheet. This data is used to calculate more data, such as internal multipliers, delist multipliers etc. When a moviestock delists, that data is transferred to the 'Data' sheet. It has data for all movies in wide release since 1998.

The OW sheet is designed to help predict a movie's opening weekend based on its previews and Friday box office, by using historical comparisons to generate a set of predictions. It also generates a set of delist predictions based on those comparisons.  It's set up for 3-day weekends, four day weekends and midweek openers/Thanksgiving weekends.

The Comps sheet features up to eight movies that are in release, and historical comparisons for each one. It provides some headline data for each comparison including release date, OW, delist multiplier, etc.  These comparisons are used throughout the rest of the sheets.

The DC sheet provides a lot of data between the in release movie and its comparisons - for example, it shows their weekend % drops.

The Charts - sheets C1, C2, ... C8 - show graphs of in release movies against their historical comparisons. Very quickly, the charts show an in release movie's box office over time.  It then takes a comparison movie's cumulative box office over time, expressed as a percentage of its 3-day opening weekend take, and applies that percentage to the in release movie's OW to show what its box office would be if it acted like the comparison movie.  Obviously the closer the , the better the comparison.

The other sheets - the spreadsheet is a work in progress, and has several old experiments, which I won't go into here.

The Mechanics

Here's the formula used for an three-day OW prediction based on a historical comparison:

1. Take the actual preview and Friday box office of the movie being released.

2. Calculate Saturday prediction by this formula: (Friday BO - previews) * (Comparison Sat/Comparison Fri)

3. Calculate Sunday prediction by this formula: (Saturday prediction) * (Comparison Sun/Comparison Sat)

4. Sum the actual preview and Friday box office with the Saturday and Sunday predictions to get the opening weekend prediction.

The formula for the delist prediction on the OW sheet:

1. Generate a three-day delist multiplier for a historical comparison by this formula: (Box office at delist - Box office before opening weekend) / Three day opening weekend box office.

2. Multiply this delist multiplier by the weekend prediction generated using that comparison.

The formula used for the charts:

(Comparison Box Office - Comparison Box office earned before OW) / Comparison OW * (Release Box Office - Release Box office earned before OW) + Release Box office earned before OW

Because adjusts and adjust multipliers will no longer feature on HSX, I'm using a basic three day OW for all movies including those with irregular openings - four day weekends, etc.  This allows direct comparison between movies that have different release patterns.

Caveats

The sheet is updated manually. I try to do this every day but sometimes I get behind. One of the main reasons why I haven't gone public before is because I don't want to feel obligated to keep it going.

There's still some work to do to update it to make it suitable for public view (eg ensuring 'estimate' is not used to describe a 'prediction', tools I developed that didn't work out, out of date info about the sheets etc). There'll be a few changes in the coming weeks. 

Nothing systematic has been done to compare predictions against performance.

Going forward

Those are the basics. I don't want to go into details/minutiae in this post, but if you have specific questions, ask them.

I'd appreciate a discussion about the formulas I've used and whether they are the right ones. Obviously what was fine for something used just by me is different to a tool for traders.

I'm happy to have a discussion each week about which movies I think look like making the best comparisons for each new release and fielding suggesting - though ultimately, it will be my call.