Daily Fantasy NASCAR Track Preview: FanShield 500

Races at short, flat tracks in 2019 were completely unwatchable.

It was the first year under a new aero package for the NASCAR Cup Series, and the huge spoilers made it impossible to close up on a car and make a pass. It also drastically changed the tracks for DFS because passing plays such a key role in how we fill out lineups.

Thankfully, NASCAR heard our gripes and reversed course for 2020.

Starting with this week in Phoenix, the Cup Series will go back to a lower-downforce package at the shorter tracks and road courses, reverting things to a similar spot to where they were back in 2018. Rejoice, and be glad.

Because of this, we have to dig into what happened last year compared to other seasons and see whether we need to throw the 2019 data out. Then, we can move on and formulate an ideal roster construction now that we're back to the pre-2019 package.

A Look Back

Phoenix has never been the easiest track from a passing perspective. Most short and flat tracks aren't as the lack of banking prevents there from being numerous lanes drivers can use like we saw last week in Fontana.

This by itself means we might not have to throw all data from 2019 out the window. But we should dig in to see just how different those two races were versus what happened in the past.

There are two routes for doing this. The first is by looking at the number of green-flag passes in each race. This will -- quite literally -- show us how many passes there were, which should be a good indicator of the difficulty of passing.

The other is by looking at how tightly finishing positions correlated to each driver's starting position. This is key for daily fantasy because -- if drivers tend to finish where they started -- it's going to be tough to snag good place-differential options.

Here are both of those numbers from the past six Phoenix races, again with the two 2019 races being under a different package than what we'll see this weekend.

Interestingly, the average number of green-flag passes per race was higher last year (1,342) than the previous two years (1,314). This means we don't need to get radical about ignoring 2019. But there were other signals that said our best data will come from the earlier races.

There was a massive tie between each driver's starting position and where they finished in last year's fall race, higher than any of the five races that came before it. The spring race also had a slightly higher correlation than the average of the 2017 and 2018 races. This means that the drivers who finished best -- a key for DFS -- tended to start near the front.

Starting up front was also critical for identifying drivers who would lead a bunch of laps. Phoenix was never a spot like Daytona where you could lead laps from anywhere, but the dominators were super front-heavy last year.

Here's a look at the starting positions of the drivers who led at least 50 laps in each season. For this chart, both the spring and fall Phoenix races have been combined into one column.

Last year, none of the five drivers who led 50 or more laps started outside the top five. Both 2017 and 2018 saw three of five drivers hit that mark after starting outside the top five.

This is a relevant discussion because the Cup Series will run 312 laps this Sunday, the longest race so far in the young season. That means there are 31.2 FanDuel points available for laps led, and we'd be wise to try to scoop up as many of those as we can.

Last year, if we wanted those laps led, we had to target almost exclusively drivers starting up in the first two rows. It seems like we should have the flexibility to drift a bit lower in 2020. We shouldn't just ignore the drivers at the front if they're fast, but we don't have to feel as tied down as we may have under the 2019 rules package.

Based on the number of laps led available, we should probably aim to have two potential dominators in each lineup. In both 2019 and the previous two seasons, there was an average of 2.5 drivers per race who led 50 or more laps, which definitely makes a difference for DFS. As such, it would be wise to pinpoint a pair of drivers capable of leading laps for each lineup. They just don't have to start as close to the front as they did last year.

The switch back to a lower-downforce package should also allow us to get high-quality mid-range and value plays from the middle of the pack and deeper.

Part of this is that passing should be easier, creating a lower correlation between starting and finishing positions. That's what we're looking for with the value plays.

The other part is that a lower-downforce package lends itself to a higher crash rate, which is more conducive to finding place-differential studs. Every time a driver crashes, it pushes everyone behind them one spot higher, inflating both their place-differential and finishing points.

In last year's races, 7 of 75 drivers (9.3%) failed to finish the race, and there was an average of 7 cautions per race. In 2018, those numbers were 15.8% and 8, respectively, and they were at 16.5% and 7.5 in 2017. If you're projecting attrition rates for this year, you should skew closer to the 2017 and 2018 numbers than what we saw in 2019. That's going to make it easier to find drivers who start further back and finish well.

To see the more back-friendly nature of the racing in action, we can look at the perfect FanDuel lineup from the 2018 fall race. There, both Kyle Busch and Kevin Harvick led at least 70 laps, but Harvick's fifth-place finish and some place-differential in the middle tier allowed the lineup to have more of a skew toward the back.

That lineup fits with our emphasis on identifying drivers who can lead laps and also exploits the opportunities to snag some place-differential for our mid-range and value plays. With the way races at Phoenix transpired prior to 2019, it wouldn't be a shock to see a similar script on Sunday.

Obviously, we have to be flexible based on what happens on Friday and Saturday. If no drivers who were competitive in Friday's practice sessions qualify poorly, then we're going to drift closer to the front. The same is true for the potential dominators: if they qualify at the front, then we should plug them in. We just have more flexibility now in the event that they slip a bit in qualifying.

That practice data will go a long way toward deciding who we target. In last year's fall race, the correlation between each driver's practice portion of my model and their average running position was the second-highest for any race the entire year. We'll likely get quality 10-lap averages from each driver in both Friday sessions, and it should give us a good idea of who will be fast once the green flag drops. That data can be our crutch both for identifying dominators and drivers with place-differential upside.

Overall, Phoenix wasn't a ton different in 2019 than it was in previous years. But the small changes should alter our approach for DFS.

First, when we are hunting for laps led, we don't have to quarantine ourselves to the front of the field. If a driver is fast, they can make their way to the front, as long as they aren't digging themselves out of too deep of a hole.

Second, we don't need to pluck all of our mid-range and value plays from the top 20 and hope they grind out a good finish (which is what happened in the 2019 fall race). You can find drivers with place-differential juice -- whether it be due to raw speed or attrition for other drivers -- a bit further back. We should be inclined to exploit that if presented the opportunity.

Phoenix isn't suddenly a "stack-the-back" track now that the 2019 rules package is in the graveyard. We still need to prioritize laps led, and it's possible our best rosters will include drivers starting closer to the front. The reversion moreso means we have more flexibility, and qualifying will not entirely dictate the way we play things this weekend.