Friday, March 7, 2014

What the Numbers Say... 2013 Races

Hello Internet! Look, look, it's a blog post the week after a blog post. That must mean that it's March, and it's almost time for the [as yet to be named title sponsor] IndyCar Series to get underway for 2014! I am excited beyond excited for this year, and I was even before the recent "good news every day" tidal wave. So, you can imagine that I am ready for 2014 to happen. Like immediately. But, before we can do that, I've been slowly digesting all of last year's numbers over the winter. (Let's be fair, it's still winter here in Wisconsin.) And, over the next two posts, I'll be presenting them to you in preparation for next season. Think of it as a "How to Read a Race Score" guide. (Even though I'm still figuring this all out as well...)

The Numbers of Note

Ok, enough goofing off. Here's the cold, hard, numerical truth about 2013 in IndyCar. This is each race, listed with the high Race Score, the low Race Score and the Average.

St. Pete80.821.2736.78
Long Beach*86.81-17.1630.80
Sao Paulo*82.67-14.9628.90
Indy 50079.740.9935.77
Belle Isle 190.531.3337.55
Belle Isle 277.371.3336.94
Toronto 175.731.1436.68
Toronto 298.19-4.5736.68
Houston 184.741.4036.71
Houston 295.62-8.3436.99
*Scored with the Old Formula

You get all that? Good. Confused by it? Me too. Let's chat shall we? I'll be referring back to the chart, so don't worry about scrolling up and down. It's not going anywhere.

A Primer on Race Scores

The "Race Score" (or sometimes "Aggregate Race Score") is a numerical tool that I use to help me try to quantify who races well on a particular day. The score takes into account: 1) Finishing position, 2) Laps led, 3) Positions gained on track (relative to both start and qualifying), 4) Average running position on track, and 5) Laps completed. If a driver qualifies on pole, starts on pole, leads every lap, and wins; that driver would have a Race Score of 100. That number is arbitrary, but it seemed like a good place to start.

For further reading on my current Race Score Formula, and how I arrived at it. I would suggest you check out a couple of posts from last year: Behind the Curtain on the New Formula and Pay No Attention to the Formula behind the Curtain. Just click the links and you'll go right there. Or, I'm bad at links. That could be the case.

Well, now that we're all caught up, onward to the data analysis. That's why we're all here! 

What the High Scores Mean

The High Score goes to the driver with the best race. This is often the winner. It is not always the winner. What the magnitude of the high score can tell us is how dominant/impressive that driver was over the course of the race.

High Highs

A very high top score indicates a very impressive drive. If you watched last year's race at Iowa Speedway, you probably aren't surprised that the highest score of the year goes to James Hinchcliffe at that race. Hinch led 226 of 250 laps; started second; finished first; and had an average running position of 1.23. That's dominant. Similar instances of high scores indicating dominance are found in: 

- Will Power's performance in the second Houston Race
- Scott Dixon at Toronto 2
- Mike Conway, DESTROYER OF WORLDS, at the first Belle Isle Race
- And, Helio Castroneves at Texas

But, a high top score can also indicate a very impressive run that is put together by a non-winner. This is mostly done by Sebastien Bourdais. Bourdais had the highest score in Baltimore last season despite finishing the race in P3. This was accomplished in a couple of ways. First, he carved up the field, coming from P22 on the grid to nab the lead less than halfway into the race. Second, he led the second most laps behind only Power, who failed to finish the race.

In short, a High High indicates something special. Of course, something special is only fun to watch if it's something like Bourdais at Baltimore (worth noting that he did a similar thing at Toronto 1, but Dixon still outscored him). If something special is Hinchcliffe at Iowa, we'd better all hope they're battling farther back in the field. Regardless, I can appreciate special. My hat's off to these 90+ performances from 2013.

Low Highs

A "low" top score (we're talking in the 70s) usually means competition for the lead. Now, this can take on a variety of forms. Let's look at last year's "Under 80" crowd for a sampling:

- Indy 500: The top score in the 500 last year was 79.74, just barely low enough to count as low. Let's look at the contributing factors. 1) Three drivers led over 30 laps, and no driver led over 40. This lowers everyone's points for laps led, and brings scores down as a result. 2) No driver had an average running position less than P4. This makes sense with all the lead-shuffling that went on at IMS last year, and also indicates that less points will be had for average running position, which drops scores again. I would point out that this is a good example of how a low top score can actually indicate a dramatic race.

- Belle Isle 2: Top score here goes to Simon Pagenaud with a 77.37. This is an instance of a dominant driver (here that is Mike Conway, DESTROYER OF WORLDS) getting shuffled back late in the race due to strategy, and Pagenaud, who benefited from being able to stay out longer before his last stop, was able to strategize his way to the top, without leading as many laps as Conway and with a worse average running position. I find race strategy fascinating. I was glued to my TV for this. I understand that's not everyone's thing.

- Toronto 1 and Sonoma: These are very similar, so we'll lump them together. In both cases the Top 3 finishers in the race all led similar amounts of laps, with very similar average running positions. This leads to a flatter distribution of points. This leads to lower top scores. And, for my money, it leads to fascinating strategy and good road racing.

- Fontana: Finally we get to the finale at Fontana. Top score here goes to race winner and (dare I say) Oval Maestro, Will Power. The mathematical reasons for this being a lower top score are similar to the reasons at Indy. How we get to the math, however, is different. Indy had 19 cars finish on the lead lap. Fontana felt more like an Indy 500 from the early 1990s with only 5 entrants completing the posted race distance. Note, this is NOT a bad thing. Fontana was a compelling race. There was attrition. There was dust. There was more dust. There were worries about championship contenders being able to finish. There was even more dust. The point is, that a low top score usually indicates drama. Sometimes that drama is the pass-fest known as the 2013 Indy 500. Sometimes that drama is the attrition-fest known as the 2013 MAVTV 500.

What the Low Scores Mean

Let's just take a moment to glance at the "Low Score" column. This isn't nearly as fascinating as the high scores but there are a couple things worth noting. First, you'll see that in the three early-season races that didn't get re-scored with the New Formula, the bottom score is ABYSMAL. The "best" worst score from those races is -14.36. Basically, if a car that qualified well got punted out of a race early or something broke on it, that driver was faced with a score of worse than negative ten. That stinks. That's a lot of feel bad. Note: not that any drivers are actually reading me and therefore feeling bad, but I have a lot of feelings, so I'd feel bad for them.

Anyway, with the New Formula, you'll see that, most of the time, the worst score is between two and zero. I can live with that. Second, now those big negatives mean something. Let's look at the two "worst" offenders.

- Helio Castroneves at Houston 2: Not a lot went right for Helio in Houston, and this is an example. Starting from pole, Helio led the first 10 laps and then things went wrong. The car decided that it had had enough for the day and quit on him. He finished P23. That's a really bad day. Race Score: -8.34

- James Hinchcliffe at Toronto 2: The Toronto weekend did not end on a high note for Hinch, in much the same vein that Houston 2 stunk for Helio. Hinch's car didn't even make it out for the formation laps. He was down laps before even turning one. Let's just say, an average running position of 23.15 doesn't bode well for your scoring day. Race Score: -4.57

Basically, in the Brave New Formula World, if you're more than two points worse than zero, you've had something go terribly bad when you weren't supposed to.

What the Averages Mean

Finally, the "Average" column. The primary thing this column tells us is that the New Formula is higher scoring than the Old Formula. Now, this is something we know. Look at the averages for the three Old Formula races. We have: 31.35, 30.80, and 28.90. This led me to postulate (using 2012 scores, which exist only in the dark recesses of my computer, and these scores from 2013) than a "Good Score" was 35.00 or better. Usually, a 35+ score would put you in the Top 10, and score you  points for the Race Score World Championship, which is my totally made up championship where I award points to drivers based on race scores.

Now, we can see from the "Average" column above (I know, I know, it's a long way to scroll now...) that under the New Formula, an average score is between 35 and 38. Now, a 35 is no longer a points-paying score. With this mathematical construction in place, we're looking at what once was (if only in my mind) the indication of "ya done alright" now signifying "better pick it up." So, the question that we now wind up with is wondering what makes a "good score?" Hopefully, we can answer this below.

What Gets You RSWC Points

Last thing today, here's a chance for us to determine, going into 2014, a benchmark for "good score." In order to do this, let's take a look at the lowest points-paying scores from every New Formula race in 2013:

St. Pete47.01
Long Beach~~~
Sao Paulo~~~
Indy 50049.81
Belle Isle 145.26
Belle Isle 249.79
Toronto 143.37
Toronto 243.26
Houston 145.66
Houston 247.33

As you can see, there is quite a range in scores, from 38.78 grabbing a single point at Mid-Ohio to needing 49.81 to score the single at the Indy 500. The average of these scores, however, is 45.00. And, the median score, within the group, is 44.97.

I think that (armed with this) we can draw the "good score" line at 45.00. This score will, most of the time, score you points, and even in a 23-ish car field, Top 10 is nothing to sneeze at. Let's take a quick look at the biggest outlier to make sure our position is solid.

- Mid-Ohio: The tenth-best score at Mid-Ohio was a very low 38.78, posted by Sebastien Bourdais, who qualified 14, stared 13, finished 12, and had an average running position of 11.58. No one would say that Bourdais set the world on fire at Mid-Ohio, so even though he sneaked into the points, I'm fine saying that his run is less than optimal. Additionally, the P9 score at Mid-Ohio was Hinchcliffe's 45.59: right there at "good score."

So, going forward, we'll call 45.00 and greater a good score, meaning drivers will have to do well above average to earn it.

Stay Tuned

I hope this look back at last year's numbers has brought some fond memories back. I also hope that with the information we've put together here today, we can all go into the 2014 IndyCar Series with a good idea of just how all these numbers work. And, with that in our pocket, we can get to another year of quantifying the race!

Remember to follow @ScoringIndy on Twitter for blog updates and race predictions. Seriously, you'll want to pay attention to those race predictions. They're sure to tell you exactly what won't happen.

Next week, we'll take a look at some driver's seasons. And, try to see who's primed for a good run in 2014, by looking back at entrant (not driver, entrant) performance in 2013.

See you then!

-- Guido

No comments:

Post a Comment