We welcome Alex's data expertise to the fold with Bama-MSU 2015: Experimental charting

Ed. Note: We'd like to welcome Alex to the fold. Saxon is probably not going to be able to churn out his excellent data analyses this fall owing to silly things like real life, but, like a trooper, he found the next man up. I was covering SECMD16 last week, yet I still should have been posting Balloons' first article. For that, I apologize to both Alex and the readers -- you deserved better. The link to the first article is embedded herein. Please give him a warm welcome in the comments, and enjoy his hard work. I know that I did. RTR -Erik

Last week I published a project on charting football games, using 2015's Cotton Bowl as the sample game. The game ended up being a hilarious rout, so I thought I'd put up a few of the charts here for your viewing pleasure. As a reminder, the box score and stats are available on RollTide.com (where I pulled all this play-by-play data from).

Box score, ALA-MSU Cotton Bowl

Let's start with an easy one: Success Rate by quarter. Everyone loves a bar chart with a good story, right?

Success Rate by quarter, bar chart, ALA-MSU Cotton Bowl

I'm using the same success metrics that Saxon does; See the Football Study Hall definition for details. The results are straightforward: once punts are removed from the data, the Tide saw higher success rates across every quarter.

The consistent performance gap surprised me: the game was tied 0-0 throughout the first quarter and was only a two-score game (10-0) by the half, but creeping beneath that box score was one offense significantly outperforming the other. That third quarter was an absolute monster performance from Alabama (which is obviously reflected in the box score, too), but notably, the Tide backups still edged out MSU during the garbage-time 4th quarter. Roll Tide, folks.

If we roll the numbers up cumulatively, we see a similar (obvious) story: of all of the "success" to be had in this game, one team was taking up more than its fair share.

Plays-by-success, area chart, ALA-MSU Cotton Bowl

The cumulative view makes the story extra punchy: that chunk of successful plays on the Alabama side is downright meaty (it helps that it's red). Interestingly, MSU did have more explosive plays, and they started happening earlier in the game, but the come-from-behind playbook was a likely contributor...big pass attempts tend to lead to lower success rates but higher explosiveness.

To get down into the details, we need to talk about plays. I broke out a few charts that represent plays with individual dots or squares.

Here's one:

Plays dot plot, ALA-MSU Cotton Bowl

To show more of the game's "flow," this is broken down into quarters, then into "quarter quintiles" (just 3-minute game clock increments, five per quarter, to show more detail). "Good plays" for each team are close to the horizontal axis, while worse plays are farther away.

The basic story is the same: the Tide dominated the 2nd and 3rd quarters across all play types: successful, explosive, scoring, turnovers, and special teams. But there are some interesting nuances here: see the spurts of success that the MSU offense had in the 2nd and 4th quarters. Conner Cook and team were able to rally a bit in moments to create some momentum, only to be snuffed out by interceptions (most notably Cyrus Jones's red zone INT to basically close out the first half). Also, the Spartans had just enough explosive plays, sprinkled out fairly evenly, to keep their offense on the field instead of 3-and-out'ing every time: the teams ended up with similar total play counts by the end of the game, despite the lopsided nature of successful plays and score.

If you like what you see, check out the full article. Admittedly I haven't yet gone into comparisons to season averages, or personnel stats, or (a particular fan base favorite) run/pass splits, but there's more to come. In fact, I'm hoping to put together some charts for RBR's football game reviews ("Charting the Tide") this fall.

As always, comments and questions are appreciated.