Fantasy Movie League Fall ’15 Week 10 Picks

UPDATE 11/5 ~Noon Pacific BELOW

After a week in which only one film earned more than $10M, the heavy hitters are back with two movies predicted to bring in four to eight times that, if not more.  The long-awaited return of the super villainous organization “Spectre” gives us what is possibly the last time we see Daniel Craig as James Bond.  “The Peanuts Movie” brings beloved characters to a 3D computer animated format and potentially to a new generation of audiences.

How does that shape up the week for Fantasy Movie League picks?

Week 10 Perfect Combo Probabilities

I’ve been playing around with my predictive model here quite a bit and running regression tests on prior weeks.  What I’m finding is that averaging ShowBuzzDaily and ProBoxOffice.com weekly forecasts does better than relying on either of them exclusively.  ShowBuzzDaily had a good run of new film forecasts that led me to using them exclusively for debuting films but I’m now seeing over the longer haul it averages out well.

With that as background, if you average ShowBuzzDaily and ProBoxOffice.com forecasts this week, figure out the average value for each of those films, and then apply that average value to the price of unforecasted films you get what is currently loaded into the FML Comparator.  If I run my variance simulator on that, which takes the top 6 Best Performer candidates and varies each by +/- 10% for returning films in that top 6 and +/-30 for new films in the top 6, there are 729 possibilities (three states, six films, three to the sixth power is 729) and results in the following table:

Week 10 Probabilities

When I ran regression tests on this tweaked model for the Fall season, 8 of the 9 weeks already played, this technique had the Perfect Combination somewhere in its first 10 entries.  There are 355,083 legal combinations to play this week and now let’s apply some common sense to these 10 because, more than likely, the Perfect Combination is in here somewhere.

Interpreting the Perfect Combo Probabilities

“Spectre” or “Peanuts”? Pricing this week has us trying to choose between one of the Bond film or two of the children’s classic characters to lead our lineups with. There’s a great thread in The Chatter discussing the merits of each and the ProBoxOffice.com article has a good list of pros/cons for each as well.

To summarize those sources, “Spectre” would logically compare to 2012’s “Skyfall” and its $88M opening but this latest edition to the Bond series doesn’t have 50th anniversary buzz or a popular Adele single to help its marketing like its immediate predecessor, but instead a Sam Smith single that isn’t getting nearly as much attention. Is “Peanuts” a reboot of a classic for a new generation of kids or does it instead compare to 2011’s “The Muppets”, which had a $27M opening and can be argued appealed mostly to nostalgic Gen Xers instead of their children?

Further down the line, the top two “Spectre” combinations get paired by the variance simulator with “Goosebumps” and either “The Last Witch Hunter” or “Burnt”. Similarly, most “Peanuts” x2 combinations use “The Martian” and one of “Our Brand is Crisis”, “The Last Witch Hunter”, or “Burnt”.

For what it’s worth, and taking a page from a comment by Phil’s Phun Phlicks (#11 in this thread), if I do a Fandango search on my zip code, five local theaters come up on the first page of results.  Here is the number of Friday showtimes for second level films across those movie houses:

Burnt: 23
Our Brand is Crisis: 14
Last Witch Hunter: 13

Skew that how you’d like given that I live in Orange County, CA but at least my local market seems to be giving Bradley Cooper a second chance, although that hardly guarantees that people will show up to those times. All three films, when averaging the professional estimates, look to drop between 47.5% and 51.6% and whichever one you think will do better than that should guide the completion of your screens this week.

UPDATE 11/5 ~Noon Pacific

One of the things that is extremely rewarding about doing this forecast modeling in public view is how kind people are in pointing out errors, which in turn makes the model better.  A commenter below saw something peculiar in my output this week, which I initially thought could be accounted for in the normal variances, but when I took a closer look over lunch today I discovered a bug where I was varying every movie, not just new ones, by +/- 30%.  When I fixed that to vary the top 6 BP candidates by +/-10 if returning and +/- 30% if new, it produces the following:

Week 10 Probabilities2

Most of the same logic applies as discussed earlier related to “Spectre” vs “Peanuts” and the merits of secondary films, but the combinations differ slightly.  I’ll run regressions over the weekend here as well to see how past forecasts were impacted by this bug.

3 thoughts on “Fantasy Movie League Fall ’15 Week 10 Picks

  1. I think there is an error with your top pick. You have 4 “Our Brand Is Crisis”, but with that selection, it leaves you with $8 left over. However, if you only selected 3 of them, you are left over with $24, which allows you to select “The Last Witch Hunter” which is set to out perform “Crisis” by almost $1 million. Maybe not a big difference, but with the margins between the top 5 choices so small, it may make a difference. This combo also makes you have $0 left over.

    Like

    1. Hi there and thanks for the comment. Sometimes the math works out fluky and I didn’t catch it at first but thanks for drawing attention to it. The combination you refer to is listed as #9 in the table and the top combination wins when Our Brand is Crisis does well and The Last Witch Hunter does poorly. This is why I always write a disclaimer sentence that you have to use some common sense when looking at what the math outputs. While I didn’t mention it specifically in the article, you are correct that it probably makes more sense to play #9 than #1.

      Like

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s