Data Deep Dive: Summer Season Superlatives

The Fall season is well under way but I wanted to take a look back on the Summer season while it is fresh in our minds.  There are two graphs this week.  In the first, you won’t find much actual analysis but you’ll find lists of interesting numbers for the Summer 2015 season pertaining to the historical fantasy pricing of all 57 movies that were in play at one time or another.  The second graph today is a little more insightful as it’s a closer look at the Perfect Combination from the Summer season, an especially important topic given the updated scoring system for Fall.

Why should you care?

Summer 2015 was the first full season of Fantasy Movie League and as I more deeply analyze the data for any patterns, I’m hopeful that there are lessons to be learned.  For the pricing, we have a graph that looks cool but doesn’t appear to tell us much but for Perfect Combinations, there is the beginning of a pattern emerging in terms of how likely it is to find it each week.

Where’d I get my data?

For the pricing information, I asked nicely and the good people at Fantasy Movie League gave it to me instead of making me click all over the website to figure it out for myself.  My index finger and my family time give many thanks.

In order to get percentage figures for the likelihood of Perfect Combination, I needed either a running count of registered users or a time machine so I could see them myself on old pages.  The FML staff supplied me a spreadsheet of daily user registrations, although I would have preferred a DeLorean.

What were my findings?

If you simply graph the fantasy price of each movie, each week it was in play, you get this:

Summere2015HistoricalPricing

That legend gets pretty hard to make out so I added a few labels and arrows for readability.  I haven’t taken the time to do the math on it yet, but there’s a definite curvature pattern going on there below about FB$200 and I have no idea whether or not that means anything definitive but it is oddly pleasing to look at for some reason.

A few superlatives from this graph:

  • Highest High: “Jurassic World”, Week 4, FB$894
  • Lowest High: “Straight Outta Compton”, Week 15, FB$213
  • Highest Low: “Fantastic Four”, Week 15, FB$20
  • Lowest Low: “Poltergeist”, Week 4, FB$2
  • Most Weeks with Highest High: “Straight Outta Compton”, 3 weeks

The other data I got my hands on was not only the Perfect Combination for each week but the number of registered users.  That enables us to look at not only the number of players who correctly picked the Perfect Combination, but also the percentage of the player base each week.  Given the new $10M bonus for this, it is of big interest to most players with many concerned that it gives too big an advantage.

Summer_2015_Perfect_Combinations

If you look at the larger sample, only once did 5% or more of the available player population choose Perfect Combination correctly and only twice was it even above 1.8%.  It was far more common, with five occurrences, to have less than 0.14% of available players be so fortunate.  Eleven times out of the fifteen weeks, less than 1% of players hit on the Perfect Combination.  If some super small percentage of the player population achieves something like finding the Perfect Combination, giving them a bonus seems reasonable.

However, there are two spikes towards the end of our sample.  Is that because players are getting smarter or are those random?  Were more people starting to use the hyper controversial late Thursday numbers to get early indicators on the Friday actuals or were those particularly easy weeks?

There isn’t enough data to tell whether or not these last few spikes are signal or noise at this point, but it needs examination for Fall 2015 season to see if it has become a trend. I plan on tracking this going forward (0.85% of users picked the Perfect Combination correctly for Week 1 of Fall, for example) to see.  If you have other ideas on what could be done with this data that you’d like me to explore, you can find me on Twitter @nerdguru or I follow replies on all my posts for at least 24 hours after they originally appear.

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