Here’s Why You’re Not Going to Win Storrow Pool

Share This

Storrow

It’s that time of the year – hell, trucks already have gotten stuck on Storrow Drive. We love the game and story of Storrow Pool, when we all crystal ball and joke about the inevitability of a truck getting stuck on Storrow Drive. It brings me a lot of joy.

It’s not the only thing: you see, the immense amount of joy I get from FOIA is kind of sad, too. Sometimes these things collide (no pun intended).  For example, did you know that massDOT does a phenomenal job of collecting crash data on car accidents around the state and makes it available for review?

Of course, with this new database at my fingertips, I want to let all of you know that you have no chance against me in this year’s #StorrowPool. I downloaded the data with the sole purpose of Nate Silvering the crap out of it and proving trends to pin point when bridge incidents with trucks on the historic trailway happen on September 1st.

There’s just one problem: looking at the data…Storrow Pool is a lie.

In order to get the data to a working system and agreement as to what constitutes an actual Storrow Pool incident, I developed the following methodology:

  1. Query the official report for all accidents in Boston that include Storrow Drive.
  2. Evaluate the incidents based on (a) what time of year it occurred, (b) involved vehicles, specifically light trucks or vans and (c) if the “first harmful incident” was a collision involving a bridge or overhead bridge structure.
  3. Make charts.

Out of 2,746 total incidents that occurred on Storrow Drive between 1/1/01 and 12/31/11, 47 incidents involve overhead bridge collisions, as close to the criteria above as we want to narrow down for our first examination. I didn’t limit to September, so here’s the break down of which month (year not limited) those 47 incidents occurred:

IncidentByMonth

The data points to the fact that September seems just as regular as any other fall month, and it isn’t the highest overall. Here comes the fun caveat: only one of those sixSeptember incidents, according to these reports, happened on the 1st – back in 2007. Now, that August outlier is worth considering, too, because of timing and the fact that incidents, like yesterday’s, could happen the week around September 1. No truth to that, at least in this data: August 8th, 2003 is the data point there. Interestingly, the only trend about those September incidents (and trend should be used lightly since this is a small data set) is that four of the six happened around the middle of the month – two on September 18th (2005 and 2011), one on 14th (2006) and one on the 12th.

Fact is, the 1st of the month is a lot less likely than the middle of the month, which doesn’t make logical sense. To check this out, I split the data by date, independent of the month or the year. Check this:

storrowbyday

Nothing comes through, again, in terms of anything – except perhaps that the trend doesn’t hold for the beginning of months, but the middle. In fact, it looks like we shouldn’t be playing StorrowPool around student move in, but move out. It is the third week of May where there is a pile up of similar incidents – four of the five May incidents occurred between the 17th and 26th of the month.

As a note, I was worried about the data set being too limited or incomplete – perhaps a specific roadway didn’t get included in the report when filed. So, I removed the Storrow Drive limitation and repulled the data to see if anything else came up trend wise. We jump from 47 to 86 incidents in that time frame, but many of these involve places like parking garages and other roadways/onramps. It did bring in a few valid data points and an important clarification: there are parts of “Storrow Drive” that are technically Soldier’s Field Road, thus not in the initial search (and #SoldiersFieldPool is a pretty shaunty hashtag). When we add those to the data set, we get an additional 10 incidents, and I’d put three of them (9/1/2005, 8/31/2002 and 9/7/2010) squarely in range. Still, we have not enough data to suggest that this week is any more likely than any other time of the year.

Ok, so all we’ve proven here is that Storrow Pool is a bit of a myth in terms of its relation to the first of the month or the bulk of September 1st Moving Day. That doesn’t mean we shouldn’t be vigilant. Good luck, friends.