Martinsville Practice Intelligence

Practice session intelligence and signals for the Cook Out 400 at Martinsville Speedway.

Practice Signals

Driver	Start Pos	Classification	Summary
Denny Hamlin	1	CONTENDER	Fired off 4th, fell to 24th at 10 laps, then locked in at 12th for every long-run segment — honest mid-pack speed, not a dominant car. But after the session he said: "Nothing's a given — we're going to have to make adjustments, and I think we know exactly what to do." That's a six-time Martinsville winner telling you the 10-lap dip is already solved. He's starting from the pole with 735 NextGen laps led here. The car doesn't need to be the fastest — it needs to be the most manageable, driven by the most experienced driver at this track. Model impact: when a driver who has won here six times tells you he knows exactly what the overnight fix is, the model adjusts.
William Byron	2	CONTENDER	Byron fired off 26th — dead last of the competitive cars — then climbed steadily to 11th by 30 laps, which is the right shape for a 500-lap race but not the dominant profile his track history suggests. He said after practice: "Our balance wasn't perfect in practice, but feel like it's still in a good place for tomorrow. Hopefully we don't over-adjust tonight." Model impact: a driver with two Martinsville wins who knows this track cold is flagging both a balance issue and the overnight adjustment risk in the same breath — both things matter.
Josh Berry	3	CONTENDER	Berry fired off 19th but from 10 laps onward he locked in at 5th and never moved — one of the most consistent long-run profiles on the entire board. The broadcaster said it best: "His car was just steady Eddie all the way through the corner on both ends. When a car's fast here, it looks just like Josh Berry's." Model impact: that's not a lap time summary — that's a description of technique. When a broadcaster says your car looks like the blueprint for fast at Martinsville, the model treats it as a speed confirmation.
Ty Gibbs	4	FADE RISK	Gibbs fired off 2nd — legitimate top-tier short-run speed — then fell off a cliff to 34th at 10 laps and never fully recovered, stabilizing in the 15-18 range late. Kevin Harvick said in the broadcast: "I think Ty Gibbs is going to be in victory lane and it's going to happen this year." Model impact: Harvick's confidence may be right about the season, but the practice data says this weekend's setup has a serious long-run problem that needs to be resolved overnight.
Shane Van Gisbergen	5	LIVE LONG-RUN THREAT	SVG fired off 34th — last of anyone who completed meaningful laps — but climbed steadily to 13th by the 20-lap average and held there, which is a legitimate late-run improvement curve from a driver the field tends to underestimate. After qualifying 5th on a single lap he said: "It was certainly on edge, but really awesome. Felt pretty good in practice, and then yeah, got a lap down, so I'm stoked." Model impact: a driver who qualifies 5th with a single-lap effort while his long-run data sits 13th is either hiding race pace or genuinely fast over one lap — the practice improvement curve from 34th to 13th suggests there's more to his weekend than a fortunate qualifying session.
Austin Cindric	6	STRONG FACTOR	Cindric fired off 30th then climbed steadily to 11th by the 20-lap average and held there — a clean improvement curve that mirrors what Berry and Keselowski also showed. After practice he said: "Our Rich Menards Ford Mustang had some pace short run, long run. I wouldn't say we were the best or the worst by any means, but it was good to feel it out." Model impact: the understatement matches the data — 11th consistently in the long run is better than the market will price him after a quiet practice narrative.
Carson Hocevar	7	NON-FACTOR	Hocevar went 29th → 33rd → 35th → 32nd and never ran a meaningful 25 or 30-lap average. Before qualifying he said: "We had a few issues like with the brakes — I don't know if they're broke or something going on — but we're just going to continue working." Model impact: a driver who doesn't know if his brakes are broken going into qualifying at one of the heaviest-braking tracks on the circuit is describing a non-factor race day.
Tyler Reddick	8	FADE RISK	The most alarming fade on the board. Reddick fired off 7th — looked like the short-track breakthrough was coming — then dropped to 32nd at 10 laps and never recovered, settling in the 18-24th range for the rest of the session. The broadcaster said it plainly: "Tyler Reddick. Guy's on fire. He's above everybody, a little bit slower all the way through with some problematic situations. Got to get some handling issues done on this 45 car." Model impact: the broadcast confirmed handling issues in real time, and the data confirmed the car never solved them across the run.
Joey Logano	9	LIVE LONG-RUN THREAT	Logano fired off 28th and looked like a non-factor — then climbed all the way to 4th by 25 and 30 laps, one of the two or three most dramatic late-run improvement curves on the entire board. No quotes from Logano in the session, but MRN noted his 13-race top-10 streak at Martinsville coming in — and the practice data looks exactly like the car that built those results. Model impact: a driver who consistently finishes top-10 here showing a 28th-to-4th improvement curve in practice is not a coincidence — it's a profile.
Chase Elliott	10	CONTENDER	Elliott was 3rd at 5 laps, 1st at 10 laps, and settled into a consistent 4th-5th range for every long-run segment — the cleanest combined short-and-long profile of any Hendrick driver on Saturday. He said about this track: "It's pretty similar. This place is extremely consistent — it doesn't change a lot year to year." The broadcaster went further: "I think Hendrick needs a win and I think Chase Elliott's going to deliver it this weekend." Model impact: when the car looks like a consistent top-5 across every practice interval and a credible analyst names the driver as the win vehicle, the model responds.
Christopher Bell	11	STRONG FACTOR	Bell fired off 23rd but climbed steadily to 7th by 15 laps and held there across every remaining interval — a legitimate mid-run builder that locked in and never faded. MRN flagged him as one of the drivers "not showing up" in practice alongside Reddick and Briscoe, which makes the consistent 7th-range long-run data one of the more interesting mismatches of the session. Model impact: when the broadcast narrative says a driver is missing and the lap data says he's 7th in the long run, the data wins.
Ryan Blaney	12	CONTENDER	Blaney fired off 6th and just kept getting better — 3rd at 10 laps, 3rd at 15, 1st at 20, 2nd at 25, 2nd at 30. Best 20-lap consecutive average in the entire field. MRN's summary said it plainly: "The long run belonged to both Ryan Blaney and Kyle Larson." The broadcaster described his technique: "Blaney has a little bit different approach than everybody else. His big thing — he wraps the curb a lot better than everybody else." Model impact: technique-specific broadcaster analysis confirming elite long-run data is the strongest signal combination on the board.
Kyle Larson	13	CONTENDER	Larson fired off 5th and kept climbing — 4th at 10, 2nd at 15, 2nd at 20, 1st at 25, 1st at 30. Best 25-lap and 30-lap averages in the entire field. MRN said: "Best 30 consecutive: Kyle Larson." There are no Larson quotes from the session, but he once said about this track: "Our stats are really good there, but it is actually not a place that I'm very comfortable at — it takes me half the race to get into a rhythm." Model impact: a driver who admits needing 80 laps to find rhythm but still shows the best 30-lap average in the field is telling you the ceiling, not the floor.
Zane Smith	14	MID-PACK GRIND	Smith qualified 7th on lap one but the practice data tells a completely different story — 21st at 5 laps fading steadily to 26th by 30 laps. After practice he said: "You certainly see a difference and feel a difference late in the run — just with the amount of fall off, especially the transition from concrete back to asphalt with wheel spin and trying to control that. So definitely going to be a big thing we have to manage tomorrow." Model impact: a driver describing the exact problem the data shows — rear tire management and the concrete-to-asphalt transition — is confirming the car has a structural challenge that a good qualifying lap didn't fix.
Bubba Wallace	15	CONTENDER	The fastest car in Saturday's practice session — 1st at 5 laps, 2nd at 10, 1st at 15, 3rd the rest of the way. MRN called it: "The single lap speed belonged to Bubba Wallace." The broadcaster on Prime broke down his technique in real time: "What did we say? Fast cars go fast? That one's fast" — then showed his line on camera: low entries, diamonds the corner, exits straight and low. "Called it the blueprint for what a fast car looks like at Martinsville." Model impact: the fastest car in practice with technique confirmation from the broadcast and the broadcaster calling it the blueprint — this is as strong a combined signal as Saturday produced.
Chris Buescher	16	NON-FACTOR	Buescher went 27th → 28th → 31st → 33rd — a profile that got worse as laps accumulated, which is the worst shape you can have at Martinsville. No quotes that change the read. Model impact: a car that gets progressively slower over a practice run at a 500-lap race track is not a race-day contender.
Ryan Preece	17	STRONG FACTOR	Preece went 14th → 6th → 8th → 10th and held the 10th range across every long-run segment — a consistent, legitimate profile that the market consistently underprices. After practice he said: "I think it drove really good. Honestly, for all the things that we wanted to be when we came here, I felt like we checked a lot of those boxes. So the Bodyguard Fastenal Ford Mustang Dark Horse is solid today." Model impact: when a driver says "we checked all the boxes" and the lap data shows a locked-in 10th-range long-run profile, that's a convergent signal.
Ross Chastain	18	STRONG FACTOR — BRAKE FLAG	Chastain showed one of the most consistent practice profiles on the board — hovering between 8th and 10th across every single lap interval without fading. But the broadcast delivered a red flag that the lap times didn't show. The broadcaster said on two separate occasions: "Saw that car bounce getting in the corner right there. I didn't like that, the way that that left front tried to lock up and that left front would bounce like that." Then: "Did it again. And then you see wheel lock right in the center of the corner. That's by himself with no pressure. You start in race conditions, he's already at his maximum limit right there." Model impact: consistent 9th-range lap data paired with a structural brake/left-front issue observed by the broadcast twice is a split signal — the speed is real, the durability concern is real.
Erik Jones	19	NON-FACTOR	Jones fired off 32nd, improved slightly to 19th by 15 laps, then dropped to dead last (37th) at 20 laps before ending his run. Before qualifying he said: "Martinsville has been a challenging one for us in general. Just fought tight kind of in practice and figuring out the power and what that's like in the long run." Model impact: "figuring out the power" at a new horsepower track that every other driver adapted to quickly is a team-specific problem, not a universal one.
Michael McDowell	20	MID-PACK GRIND	McDowell went 22nd → 15th → 28th → 25th — an inconsistent shape with a mid-session peak that didn't hold. After qualifying he said: "We got things to work on overnight. There's no doubt about it. But everybody does. Track's going to change a lot tomorrow with the rubber. So you get 70, 80 laps into the race, you'll know what you have and start working on it from there." Model impact: honest self-assessment of a car that isn't ready on Saturday — that language from a veteran is usually accurate.
Justin Allgaier	21	DEEP SURVIVAL PROFILE	Dead last in every practice category — 37th, 36th, 37th, 35th. Allgaier is a substitute for Alex Bowman and was honest about the challenge. He said: "We've got a work cut out for us today. I think that the differences between the two cars is really significant here. So just trying to maximize this car as best you can." Model impact: a Cup substitute openly acknowledging the car gap is not a betting case — it's a survival mission.
Daniel Suarez	22	MID-PACK GRIND	Suarez went 20th → 12th → 18th → 16th — an inconsistent shape with a mid-session improvement that didn't hold across the longer run. He said: "I've been nothing but impressed with Spire Motorsports. The energy is great. We're running well. We're executing well. I truly believe that we're inching our way there." Model impact: the organizational confidence is real, but "inching our way there" describes a team still building — not one ready to compete with the front of the field at Martinsville.
Brad Keselowski	23	LIVE LONG-RUN THREAT	Keselowski fired off 33rd — near the back of the field — then climbed steadily all session to 15th by 30 laps, one of the more significant improvement curves on the board. He said after practice: "I always feel like Martinsville is a track that you can kind of cheat practice a little bit, overdrive your car, and that's not what you're going to have to race. I was really happy with where we ended up after 30, 40 laps." Model impact: a driver who explicitly says he prioritized race trim over practice speed and then shows a 33rd-to-15th improvement curve is describing the exact Martinsville setup philosophy that works here.
Cole Custer	24	FADE RISK	Custer fired off 13th — legitimate early-session speed — then collapsed to 35th at 10 laps and never recovered. The 13th-to-35th drop in a single interval is one of the sharpest falls on the board. No quotes from Custer in the session. Model impact: a car with strong fire-off speed that immediately collapses under tire wear is the wrong profile for a 500-lap race.
Connor Zilisch	25	SHORT-RUN ONLY	Zilisch fired off 12th with genuine early-session speed, then faded to 18th-25th range as laps accumulated. He said after the session: "We were very good on the short run, but everything has its cost. We definitely need to work on getting it better for the long run because the race is pretty long." Model impact: a rookie who can self-diagnose his car's limitation clearly in post-session comments is ahead of where most rookies are at this stage — but the limitation he's describing is the one that gets punished hardest at Martinsville.
Riley Herbst	26	MID-PACK GRIND	Herbst sat consistently between 21st and 32nd across all intervals — no meaningful improvement curve, no alarming collapse, just mid-to-back-of-pack steadiness. No direct quotes in the session. Model impact: a stable but uncompetitive profile from a driver with two Cup Martinsville starts that both ended badly.
Chase Briscoe	27	NON-FACTOR	Briscoe was rough all session — 31st, 30th, 24th, 27th — and the garage confirmed the problem. He said: "Yeah, it definitely wasn't great. I don't know. This place has been a struggle for whatever reason at JGR. SHR, I always thought this was one of my better tracks, but just haven't been able to quite get the feel under braking that I like at least." Before qualifying he added: "In practice that was my... I was just all over the place under braking. I would lock the left front up, then I would lock the rears up." Model impact: a driver who can't trust his braking at a track defined by heavy braking, starting 27th, is a non-factor.
AJ Allmendinger	28	MID-PACK GRIND	Allmendinger ran 24th → 31st → 29th → 22nd with a partial long-run profile. No meaningful signal from the session. No direct quotes. Model impact: a veteran Cup presence with limited recent competitive Martinsville history, mid-pack data, and no broadcast signal to upgrade the read.
Todd Gilliland	29	FADE RISK	Gilliland fired off 9th — genuine early-session pace from a driver who ran top-15 laps at a 99.8% rate in fall 2025 — then faded steadily to 25th by the end of the run. No quotes from the session. Model impact: a driver with a historically strong Martinsville profile showing a 9th-to-25th fade across a single practice session is a concern the track history can't fully offset.
Austin Dillon	30	HIDDEN VALUE	Nobody is talking about Austin Dillon — and the practice data suggests that's a mistake. He ran 8th → 10th → 6th → 8th → 9th → 8th across every practice segment, one of the most consistently competitive profiles on the entire board. No quotes from Dillon in the session, but a 6th-best 15-lap average and an 8th-best 30-lap average from a driver starting 30th and drawing zero broadcast attention is the definition of a hidden signal. Model impact: consistent 6th-to-10th long-run range from a driver with no market or narrative attention is one of the most interesting mismatches of the weekend.
Noah Gragson	31	FADE RISK	Gragson went 16th → 17th → 34th → 36th — steady early pace that fell off a cliff at the 15-lap mark and never recovered. No meaningful quotes from the session. Model impact: a car that collapses at 15 laps is not built for a 500-lap race.
John Hunter Nemechek	32	FADE RISK	Nemechek showed genuine mid-session speed — 15th at 5 laps, 9th at 10 laps — then fell apart completely to 30th at 15 laps and 29th at 20 laps before the run ended. No quotes in the session. Model impact: promising early data destroyed by a sharp mid-run collapse before any meaningful long-run picture emerged.
Ricky Stenhouse Jr.	33	FADE RISK	Stenhouse showed legitimate early-session speed — 11th at 5 laps, 7th at 10 laps, 10th at 15 laps — then collapsed to 34th at 20 laps and the run ended. No quotes from the session. Model impact: a driver who runs 7th at 10 laps and then falls to 34th at 20 laps has a very specific long-run setup problem that typically shows up in the race exactly when it matters most.
Kyle Busch	34	MID-PACK GRIND	Busch went 17th → 11th → 14th then faded to 20th without completing a meaningful 25-lap run. After practice he said: "Typical Martinsville, you know, pretty similar to what we've been fighting here the last couple times. Just not bad. Car turns really good. You just got to be able to continue to connect that corner and have that drive off, especially the long run speed." Model impact: "typical Martinsville" and "pretty similar to what we've been fighting" from a driver who has finished 17th and 13th in his last two starts here is confirmation the setup hasn't changed.
Cody Ware	35	DEEP SURVIVAL PROFILE	36th → 37th → 36th — near last the entire session. No quotes. No signal. Model impact: none.
Ty Dillon	36	MID-PACK GRIND	Dillon ran 18th → 21st → 22nd → 24th — consistent but sliding backward as laps accumulated. No quotes from the session. Model impact: gradual fade profile starting 36th is a recipe for an unremarkable day.
Austin Hill	37	LIVE LONG-RUN THREAT — DEEP FIELD	Hill fired off 35th — last in the field — then climbed steadily to 14th by 30 laps, one of the most dramatic improvement curves of the session. No quotes from Hill in the transcript, and the broadcast gave him almost no attention. Model impact: a driver who climbs from 35th to 14th over the course of a practice run — at a track where almost no one moves that significantly — has something real in his car. The absence of broadcast narrative is itself the signal. Nobody is pricing this.

PIT BY NUMBERS