If you know exactly what to expect this week in Darlington, you're either a psychic or a liar. You choose.

In theory, we should have a good idea of what'll happen. The NASCAR Cup Series ran three races there last year, a delight we never get. It would have jacked up the value of track history at arguably the most unique track on the circuit.

But Sunday's race is a whole new ballgame.

All three 2020 races in Darlington (plus the lone race there in 2019) were run using the 550-horsepower package. This year, it's shifting to the 750-horsepower package. Whereas cars could run almost full-throttle last year, they're now going to be laying into the brakes pretty heavily.

That makes the track history tough to trust. But the current form is dicey, too.

If we group Darlington with the other 750-horsepower tracks, we're lumping it in with spots like Phoenix and Martinsville. Those spots are flat and slow, and Darlington is neither.

From a length perspective, Darlington is closer to the 1.5-mile tracks. But those involve much less off-throttle time due to the different packages, meaning we're almost comparing apples to oranges.

In other words, there's no one race we can point to as being the top comp for Darlington. It's a unicorn once again. Instead, we might have to take more of a blended approach.

To me, the races we want to emphasize from 2021 are those at ovals using the 750-horsepower package and the 550 tracks with massive tire wear. This way, we're seeing which drivers have done well when braking has been more pertinent while also looking at similar-sized tracks where things have gotten slick late in runs.

This leaves us with six most relevant races: Richmond, Martinsville, Bristol, and Phoenix for the 750 tracks and Atlanta and Homestead among the 550 tracks. That's a pretty good sample. If drivers have been good in both buckets, they're likely to be good again on Sunday.

Five of those six races are in the data sheet below. The one omitted was Phoenix (in favor of last week's race in Kansas) because Phoenix is flat, doesn't have much tier wear, and occurred two months ago now. Plenty has changed since then, and I wanted a better read on who may have made gains in that time, thus leading to Kansas' inclusion over it.

As always, the numbers listed are each driver's average running position rather than their finish. If we look at finishes, we'd see that Martin Truex Jr. ($13,500 on FanDuel) and Chase Elliott ($12,000) finished 22nd and 20th, respectively, in last year's playoff race at Darlington. They finished that low, though, only because they wrecked each other while battling for the lead late.

Clearly, they had better speed than their finishes indicate. Their top-five average running positions are better reflections of how competitive they were in that race.

The other data listed is each driver's starting position, FanDuel salary, and win odds at Online Sportsbook. The win odds are in fractional form, so Truex being listed at this 7 means he's +700 to win.

For the most part, we're not going to find many high-upside drivers starting further back. The two big exceptions are Kyle Larson ($13,000) and Alex Bowman ($11,000).

These two guys have been staples at the front when there has been heavy tire wear both this year and previously. Larson had a first-place average running position in Atlanta (though he finished second) and has three podium finishes in Darlington across six career races. Bowman won this year in Richmond and last year in Fontana (both tracks with big tire falloff), and he finished third this year in Atlanta. He was also runner-up in the first Darlington race last year. As mentioned in this week's track preview, you'll want to pair Larson or Bowman with a driver at the front who can lead laps, but these guys are likely to buzz through the field quickly on Sunday.

The big issue with track-history data -- even when it's from last year, as we have above -- is that drivers change teams. That changes their outlook dramatically. Two guys impacted by that here are Christopher Bell ($9,000) and Erik Jones ($7,300).

Jones is actually a former Darlington winner, and he has never had an average running position outside the top 10. But all of those runs came with either Joe Gibbs Racing or Furniture Row Racing, much better equipment than he has now. It fully taints all of those old races.

That doesn't exclude Jones from consideration this week. The extra off-throttle time puts more control in the driver's hands, so Jones' poor equipment is less of a detriment here than it is elsewhere. He's also starting 26th and has a reasonable salary at $7,300. We can still go at Jones this week, and he's honestly a fine play among the values. We just shouldn't go too hard due to his misleading track history.

Bell now occupies Jones' old seat at JGR, so we know he has the giddy-up to compete. Darlington hasn't traditionally been a great track for him (whether in the Xfinity Series or the Cup Series), but he's easily in the best equipment of anybody starting outside the top 20. There's risk with Bell, but he does carry some upside.

One value driver who's tough to fully evaluate is Aric Almirola ($7,500). Almirola ranks just 25th in the current form section of my model, meaning he wouldn't have much place-differential upside, even starting in 27th. However, the Stewart-Haas Racing cars have been less putrid in the 750-horsepower package as Almirola's two best finishes and three of his top four finishes have come in this package. The overall lack of speed prevents Almirola from being a priority, but he's still on the radar thanks to the modified package for this week.