Can Anyone Really Win at Talladega?

Every year, we hear that Talladega is a wild card because “Anyone can win”.  Which, of course, made me wonder — can anyone win?

Who Wins Races in General?

I analyzed the last three years and everything we have so far for this year. I looked for correlations between who wins races and how “good” a driver they are, as determined by how high they finish in the standings at the end of the year.

The number in each box is the percent of all wins run by drivers in the top 5, top 10, top 15, top 20 and the Chase.  I discarded some situations, like Brian Vickers, who won a race in 2013, but sat out much of the season due to illness and finished 78th in points. Same thing for Hamlin and Stewart, neither of whom ran all the races that year, but won a race.

Note that the new rule – that anyone who wins is automatically in the top 16 — will invalidate this type of analysis because someone who would’ve finished lower in points gets boosted by the win.

Year	T5	T10	T15	T20	Chase
2011	44.4	63.9	83.3	91.7	77.8
2012	48.6	88.6	94.3	100	88.6
2013	61.8	70.6	91.2	94.1	85.3
2014	35.5	71.0	96.8	100	100

Here’s a gratuitously colorful graph of the same data:

The take-away message:  It is unusual for a driver who ends the season outside the top 15 to win a race. In fact, for the last three years, more than 70% of the races are won by the top ten drivers. (And I don’t know about the goofy perspective Excel uses in those graphs.  It makes it look like the numbers for 2013 and 2014 are less than 70% – but they’re not. I promise.)

Winning at Talladega?

If Talladega really is an ‘equal opportunity racetrack’ in terms of winning, then the stats ought to look very different over the years. I analyzed Talladega races all the way back to 1990, which is almost 50 races. You know what? It’s not that different from the average.

Year	T5	T10	T15	T20	Chase
2011	44.4	63.9	83.3	91.7	77.8
2012	48.6	88.6	94.3	100	88.6
2013	61.8	70.6	91.2	94.1	85.3
2014	35.5	71.0	96.8	100.0	100.0
Talladega	44.7	72.3	91.5	95.8

The stats are almost identical relative to every other race track out there. Out of the 47 races I included, only two were won by drivers outside the top twenty.

• Jamie McMurray – 2009 Fall (22)
• David Ragan – 2013 – Spring (28)

I omitted the Spring race in 2009 because the driver (some guy named Brad Keselowski (?)) finished in 38th place – but only ran 15 out of the 36 races. So if you’re currently running below 20th place, you’ve got less than a 5% chance of winning.

Even the year Michael Waltrip (the patron saint of teams hoping for an upset at a superspeedway) won, he finished the season in 15th.

Wait a Minute… That Can’t Be Right

We all remember David Ragan winning Talladega and Daytona and Trevor Bayne winning the Daytona 500.  Is it true that if you’re not in the top 15 and you’re going to win, it’s likely going to happen at a plate track?  Let’s look at the exceptions.

Year	Driver	Finishing Rank	Track
2013	Martin Truex, Jr.	16	Sonoma
2013	David Ragan	28	Talladega
2012	Joey Logano	17	Pocono
2012	Marco Ambrose	18	Watkins Glen
2011	Trevor Bayne	53	Daytona
2010	Regan Smith	26	Darlington
2010	Paul Menard	17	Indy
2010	Marcos Ambrose	19	Watkins Glen

This year Aric Almirola won Daytona, and I’ve left that out because we don’t know where anyone is finishing yet. He could be 15th or better still.

But even if you counted him, not even half of the “upsets” take place at restrictor plate tracks.

But I swear I remember…

I sweated this one out. I have looked at Dega Data for two straight days because I knew there had to be something interesting in there.

And I finally found it – but it runs counter to all my intuition. This is one of those things scientists have to be very, very careful about – not letting our expectations get in the way of reality. If you expect to see something, you’re more likely to see it.

Does Qualifying Matter at Talladega?

So why does everyone think anyone can win Talladega?

It’s called the von Restorff or isolation effect. It’s named after a woman named Hedwig von Restorff (1906-1962), a psychiatrist and children’s doctor who conducted a set of memory experiments and found that an isolated dissimilar item surrounded by otherwise similar items would be better remembered.  In other words, it basically says that when something stands out as being very unusual, we tend to remember it.  For example, consider two lists:

A	B
9	TCE
21	GTS
3	JWN
16	PDY
QXK	QXK
5	RWV
13	MTX
8	LEB
54	DVQ
3	PYN

The same three-letter sequence is in both lists. If I showed you the lists, then took them away and asked you what you remembered, you’d remember the letters better if I’d given you the A list than if I’d given you the B-list.  We tend to remember the unusual. And there’s a reverse effect, in that you may actually remember less about the things that don’t stand out.

So while there’s a chance for an upset at a superspeedway, it’s small. It is much more likely that our winner will be a driver who’s experienced success already.