Lead Shares and Clutch Passing

by Sean Wrona

I do not have much of a knack for timing, do I? The last column was released during the middle of the 24 Hours of Daytona while this one is being released during the Super Bowl, both times when I can't imagine many people would be reading my columns. I did watch quite a bit of the Rolex and collected average lap times from it, and the most surprising observation I made was that Fernando Alonso's co-driver Lando Norris, making his first ever start in any major league series, actually had a faster average lap time than Alonso did. This further contributes to my impression of him as the most impressive racing prospect in any series in years. However, I am not writing about that this time but am continuing with observations from my previously-released NASCAR lead change project. One of my principal goals in collecting the lead changes for all NASCAR Cup races from 1985 to present (I've also done a few races prior to that as well but no complete seasons pre-1985, which I have not yet included in the study) was to come up with a single quantifiable measure of passing ability on the same unit scale as wins or number of races but solely based on drivers' ability to pass for the lead. In this column, I introduce the concept of the lead share.

The lead share, adapted from Bill James's concept of win shares in sabermetrics, is defined as the proportion of on-track passing for the lead in a race for which each driver was responsible, weighted according to the order in which lead changes took place in the race. This statistic is a composite that essentially reflects all at once how many races each driver led, how often they took the lead in those races, and also whether they tended to take the lead early or late in a race. If there are n lead changes in a race, I award 1/x lead shares to the driver who was the first driver to take the lead on-track, 2/x to the second, and so on until awarding n/x lead shares to the final driver to take the lead on track (previously defined as the terminal natural leader or TNL), where x is the summation of all the numbers between 1 and n. This allows for a measure of passing ability that rewards all drivers who pass for the lead, but most rewards drivers who make a lot of passes for the lead in a given race and those who make passes later in the event. It also heavily rewards drivers who pass for the lead in races where passing is more difficult because in a race with extremely few natural lead changes, each pass becomes much more important and I think that makes sense. We celebrate the passes that are most rare or unexpected, while many scoff at passing being too easy at restrictor plate tracks (or similarly the recent Indy 500s or the CART Handford device races). There's a reason Alex Zanardi's pass of Bryan Herta at Laguna Seca was so celebrated, and it was because hardly anyone ever passed at Laguna Seca in that era. I think it's right that passes should count more if they are rarer and this statistic also reflects that perspective.

Lead shares improve upon the previous statistic I invented called cumulative races led (which simply sums each driver's average percent led for each races on a race unit scale) because lead shares are entirely based on drivers' ability to put themselves into the lead while CRL is based on how much drivers lead regardless of how they obtained the lead, which will inevitably create a bias in favor of drivers who have stronger teams, pit crews, strategists, and the like. I noted previously that despite leading over half the race in the 2005 spring Pocono race, Brian Vickers never actually took the lead on track as he beat polesitter Michael Waltrip out of the pits, got passed by Carl Edwards, beat Edwards out of the pits, got passed by Edwards again, beat Edwards out of the pits again, and finally got passed for the win a third time by Edwards. Even though Vickers dominated the race on the surface, he did essentially nothing on his own to assist with that domination as his pit crew was almost singlehandedly responsible for it. However, because Vickers led 121 laps out of 201 to Edwards's 46, Vickers would receive a substantially higher CRL for that race of 121/201 (.602) to Edwards's 46/201 (0.229) even though Edwards did essentially all the work. His three passes of Vickers were indeed the only natural lead changes in the race, so he gets an entire lead share for that race because he was the only driver to take the lead on track and Vickers does not receive any portion of a lead share even though he dominated. This gives you a markedly different view on what domination means, and this definitely affects a lot of races. If everyone acknowledges drivers can luck into individual race wins (by beating the previous leader out of the pits and the like), it naturally follows that you can luck into merely leading as well for the same reasons. This statistic corrects for that flaw.

I will provide two further examples to show how this works in practice. Although I am not known for my concision in these articles, I am very concise when it comes to summarizing the lead changes in a race. I have a kind of shorthand I use to list all the lead changes in a given race in a single line in order. For instance, if Dale Earnhardt passed Jeff Gordon in a race, I would write 3p24 as an abbreviation for "the #3 car driven by Dale Earnhardt passed the #24 driven by Jeff Gordon." Following that model, here are two of the most famous races of the '90s and how I would summarize them and score them.

1992 Hooters 500 - 26p1, 4p3, 3p4, 7p11, 11p7, 7p11, 11p7, 6p11, 6p28, 7p11

In longform this would read: Brett Bodine passed Rick Mast on the opening lap, but they crashed on the next lap, handing the lead to Dale Earnhardt (which does not count as an on-track pass for the lead.) Then Ernie Irvan and Earnhardt swapped the lead before they pitted and got trapped a lap down during a green-flag pit cycle. That gave the major advantage to Alan Kulwicki and Bill Elliott who were able to pit under caution, and they exchanged the lead four times. Mark Martin passed Elliott on the track shortly before a caution came out and Davey Allison stayed out to collect five bonus points for the championship. Martin almost immediately passed Allison on the next restart and then dominated until Elliott beat him out of the pits and Martin eventually blew an engine. Kulwicki passed Elliott again on a late restart before Elliott beat him on a green-flag pit cycle to claim the win although he came up short in the championship battle.

How would these drivers be awarded lead shares for this race? Since there were ten on-track natural lead changes in this race, the denominator for the lead share fraction would be 55, since that is the sum of all numbers between 1 and 10. Bodine had the first lead change in the race and would score 1/55 of a point, while Irvan would score 2/55, and Earnhardt would score 3/55. Kulwicki had the 4th, 6th, and 10th lead changes in the race, giving him 20/55 lead shares, indicating he had the most impressive race from a passing perspective. Elliott had the 5th and 7th lead changes in the race, giving him 12/55 lead shares. Martin had the 8th and 9th lead changes in the race, giving him 17/55 lead shares. This reflects the general trend of Ford domination that year particularly on the high-speed ovals, although the three Fords certainly had a lot of help considering that Earnhardt and Irvan got trapped a lap down early and eventually crashed.

1999 Goody's Headache Powder 500 - 2p20, 20p2, 3p5, 3p5, 5p3, 5p3

Although one of the most famous (or perhaps infamous) races in modern NASCAR history, this race didn't have a whole lot of on-track lead changes. Front row starters Tony Stewart and Rusty Wallace swapped the lead twice early before Jeff Gordon beat Stewart out of the pits on yellow flag pit stops. Terry Labonte took the lead on two successive sets of yellow flag pit stops before being passed by Earnhardt, but four laps after Earnhardt claimed the lead for the second time, Labonte passed him back seemingly well on way to victory before being spun out by Darrell Waltrip. He then made a furious charge to recover from 5th to the lead in the final five lap sprint, bumping past Earnhardt on the next to last lap before Earnhardt wrecked him on the final lap. I gave credit for bump-and-run passes so Labonte's pass on Earnhardt on lap 499 counted, but I certainly did not give Earnhardt credit for the wreck. Since there were six natural lead changes in this race, and the numbers 1 through 6 add up to 21, Wallace got 1/21 of a lead share for the 1st lead change, Stewart 2/21 for the 2nd, Earnhardt 7/21 for the 3rd and 4th, and Labonte 11/21 for the 5th and 6th, indicating Labonte had the most impressive race from a passing perspective, unless you want to give Earnhardt credit for the wreck.

The below table ranks all drivers who led races between 1985-2017 that are included in the lead change study according to all the various race count statistics I have invented along with their actual win total. Lead shares reflects the career total of all such race lead change summaries by this list, while as already stated above TNL reflects the number of times each driver was the last driver to take the lead in the race, CRL reflects the sum of each driver's percentage led for all races, and W is the list of each driver's wins. This only counts the races for which I had enough information to confidently do a lead change summary. It includes all races from 1994 to the present, but there are 13 missing races in 1985, six missing races in 1986, two in 1987, and one each in 1988, 1989, 1991, 1992, and 1993. Hopefully those later ones will be uploaded to YouTube sometime in the near future. There were two new races uploaded since I last checked (the 1988 spring Atlanta race and the 1989 spring Richmond race, which are included in these tables, but not in my previous ones), and I removed one lead change from the 1993 Loudon race I discovered was erroneous today (as it happened in the pits, not on track), although I know that there are other lead changes that I will likely remove later if/when I review these races again for potential print publication of some of this material (for instance, occasions where drivers made a pass for the eventual lead, which I counted as lead changes before but wouldn't now.)

Drivers who are underrated will generally be those drivers who have more lead shares than cumulative races led, implying they are leading more than they should be through their own efforts (i.e. the Edwards/Vickers scenario.) Having LS > CRL would generally imply that a driver's teams on average were letting them down. However, drivers with CRL > LS are more likely to be overrated as it indicates that those drivers were leading more often than they should have been leading through their own efforts alone. This is essentially the dominance equivalent of the comparison between TNL and wins, where drivers with more TNLs than wins are unlucky because they place themselves in position to win more than they actually win, while drivers with more wins than TNLs are lucky because they win more than they should win based solely on their clutch ability late. Generally, the lead shares/CRL and TNL/win ratios correlate together, so drivers with more lead shares than CRL also tend to have more TNL than wins and vice versa. Clutch ability can be measured by taking the ratio of TNL or wins to lead shares, as this reflects performance on the final pass alone rather than a weighted average of all of them.

Once again, this supports the conclusion that Jeff Gordon is overrated, much like my previous lead change record tables did. He did pass others more often than he was passed on most tracks, but he wasn't quite the standout that Tony Stewart, Jimmie Johnson, or Dale Earnhardt were, and it would seem based on most of this evidence that ranking Gordon over Johnson or Earnhardt would seem to be untenable, though not primarily due to the championship count argument. Him having 3 more CRL than lead shares and 3 more wins than TNL implies that he was a bit lucky (particularly due to the quality of his Rainbow Warriors pit crews at his peak), although not as lucky as his detractors and harshest critics usually say he was. Just because he wasn't the best of his generation doesn't mean he wasn't one of the top handful.

Earnhardt however actually appears to be both underrated and not clutch (in individual races, probably because to some degree especially after 1990 he was taking it easy so he would make sure he finished the races to remain clutch in the championship battles). He only has 12 fewer lead shares than Gordon despite having 33 fewer wins included. It's not inconceivable he might in fact have more lead shares than either Gordon or Johnson if I include his earlier wins despite having fewer wins than either of them. Unfortunately, I think that might be impossible since most of them weren't likely televised nor did newspapers mention every single lead change in reports. It's fairly surprising he actually has 7.5 fewer CRL than lead shares considering he did have some extremely fast pit crews in that era, with the Flying Aces every bit as renowned in the early '90s as the Rainbow Warriors were in the late '90s, but it does say something about Earnhardt's passing ability. Earnhardt having more lead shares than CRL does accurately reflect his having a higher lead change percentage. Johnson and Stewart who also had very high passing ratios also had fairly large differences between their lead shares and CRL as well. This indicates that all three of those drivers likely had to keep repassing other drivers throughout the race to make up lost ground more than Gordon did, and were thus unluckier, except perhaps in championship battles.

Most of Earnhardt's rivals however appear to have been overrated. In addition to Gordon, Rusty Wallace, Mark Martin, and Bill Elliott all had more CRL and lead shares and all had more wins than TNL. In each case, they seemed to have about three or four more wins and TNL than they should have had based on the general level of their dominance, and that does make sense because they all had very obvious flaws (Wallace's superspeedway struggles, Martin's conservatism, and Elliott's short track struggles.) While they were all dominant on their specialties, they did probably get lucky in the sense that they led more than they should have led on the tracks that weren't their specialties due to the strengths of their teams, which slightly overrated all of them. It seems those four drivers snatched several wins even though Earnhardt outran them based on his own efforts.

The driver who had the highest ratio between lead shares and cumulative races led is Ricky Stenhouse, Jr. Not only did he win two restrictor plate races on rather few laps led in 2017, but he also was the TNL at the 2013 Kansas race in his rookie season, and astonishingly, he was the only driver to take the lead on track in that race (all other passes were off-track.) It was not an unimpressive pass either as he passed his teammate Carl Edwards and his predecessor Matt Kenseth 3-wide on a restart before being trapped a lap down after a pit stop. Stenhouse has managed 2 wins, 3 TNLs, and 1.5 lead shares despite only leading 0.5 races cumulatively in his career to date, and he has definitely punched above his weight when it has mattered more than a lot of other drivers. Stenhouse is probably one of the most underrated drivers in the series at the moment given how weak the Roush equipment is, how much people tend to sneer at restrictor plate-only winners because "even Michael Waltrip can win there" and so on.

Among the drivers who were more prolific winners, Jamie McMurray has the highest ratio of lead shares to CRL at 1.318. McMurray in the fairly recent past has had some seasons that were more dominant than most people realized. In his 2010, he took the lead on track in 11 races (which was tied for 5th with Kurt Busch and his teammate Montoya.) Although he won three races that year and factored in others, I don't think most people would have guessed he led 11 races naturally. Far fewer would guess that he actually led 10 races in 2014, a year where all the hype was about how badly Kyle Larson was dominating him (Larson only led two races naturally that season.) Yes, Larson is totally dominating McMurray now, but at the time it was premature. However, McMurray's 2014 was not really noticed by anyone because of his inconsistency and because he would often tend to take the lead early in the race or right before a pit stop (for instance at Sonoma he made a pass of A.J. Allmendinger for the lead and then pitted the next lap.) These aren't the sorts of things you notice unless you are really paying attention, but he was better than you think he was.

McMurray is followed by Tim Richmond (1.274), Alan Kulwicki (1.235), Carl Edwards (1.205), Neil Bonnett (1.177), Ken Schrader (1.174), Bobby Labonte (1.170), Sterling Marlin (1.148), Tony Stewart (1.144), and Kyle Larson (1.138) among drivers with four or more race wins in lead share/CRL ratio. These drivers did the best job of taking the lead through their own efforts more than the efforts of their team, and in general most of them were likely let down for their teams, and in a lot of cases this makes sense (Kulwicki, Schrader, Marlin, and Larson especially.) This is the first lead change-related statistic where Bobby Labonte has fared relatively well. It seems that while Labonte was extremely lucky at getting the lead at the end and winning, he was paradoxically unlucky at lucking into the lead during other portions of the race. He didn't luck into the lead much except when it mattered.

On the other side, Harry Gant had the lowest ratio between lead shares and CRL among major drivers, which isn't really a surprise when you think about it because he relied on fuel mileage and pit strategy races a lot in his late career (particularly his comeback years.) In his earlier years, I think it was a bit different, but unfortunately I think most of Gant's early years may be lost to history here, as three of the missing 1985 races were all three of his wins that year, so if he was more of a hard-charger and duelist when he was younger as I suspect he was, that is improperly reflected here. Gant is followed by Kyle Petty (.768), Ryan Newman (.845), Geoff Bodine (.865), Michael Waltrip (.870), Bill Elliott (.876), Terry Labonte (.882), Clint Bowyer (.890), Kyle Busch (.895), and Mark Martin (.904). These were drivers who benefited a lot from leading due to strategy, and I don't think any of them really should be surprising to anyone except for Kyle Busch. That one I don't really understand although JGR has had the fastest pit crews by far as of late so it may simply be that in his case. Most of these drivers would generally be overrated relative to the results they have posted, although Busch doesn't really feel like he belongs on this list and his predecessor Bobby Labonte doesn't really feel like he belongs on the other list, so this is certainly not the be all end all to evaluating drivers.

I already compared TNL to wins earlier so I won't be doing that again, but it is instructive to compare lead shares to wins because I think the most underrated drivers will likely be drivers with significantly higher lead shares (or TNL) than wins. Earlier Ernie Irvan and Sterling Marlin were the standouts in terms of having large TNL/win differences. Martin Truex, Jr. has become a standout in recent years with his horrible luck prior to 2017. Indeed it turns out that the lead share/win ratio rankings are very similar.

Ken Schrader tops the list of drivers with the largest ratio between their number of lead shares and wins. With 9.61 lead shares and 4 wins, he is the only driver with his number of wins who has twice as many lead shares as wins, although Jerry Nadeau, Johnny Benson, and Marcos Ambrose also do. Among drivers with 4 or more wins, Schrader is followed by Sterling Marlin (1.832), Alan Kulwicki (1.745), Jeremy Mayfield (1.603), Martin Truex, Jr. (1.485), Ernie Irvan (1.440), Bobby Hamilton (1.403), Kyle Larson (1.391), Geoff Bodine (1.354), and surprisingly, Michael Waltrip (1.352). These are drivers who based on their ability to put themselves in the lead should have had more wins than they had. Schrader, Marlin, Kulwicki, Truex, and Irvan each had a long history of their teams letting them down. It should be no surprise that a lot of these drivers (like Hamilton and Irvan) showed particularly nicely on the lead change percentage list. If you're a great passer and you have a lousy pit crew, you are going to have to make passes again and again and again, which will help you on a ranking like this but won't necessarily make the wins come any easier. Mayfield I sort of disregard because a lot of his bad luck was self-inflicted, and Waltrip I sort of disregard due to his relative lack of success outside plate tracks (although his 1991 Darlington run may be enough for him almost to belong here by itself.) Except for Hamilton, Larson, and Mayfield who had the same number of TNLs as wins, all these drivers had more TNLs than wins, and in some cases a lot more. Schrader had 7 TNLs to 4 wins, Marlin had 17 TNLs to 10 wins (making him I think along with Hamilton the most underrated driver of the period), Irvan had 22 TNLs to 15 wins, and Truex had 21 TNLs to 15 wins. All these drivers are better than they are regarded, Marlin especially, and based on this, I don't think Truex is the worst modern era champion.

On the flip side, the most overrated drivers would tend to be those drivers with the lowest ratio between lead shares and their actual win total. These are as follows: Darrell Waltrip (.654), Bobby Labonte (.730), Ryan Newman (.735), Joey Logano (.755), Davey Allison (.756), Terry Labonte (.764), Harry Gant (.765), Jimmie Johnson (.787), Bill Elliott (.793), and Tim Richmond (.799). I disregard Waltrip to some degree because this excludes his earlier and greatest seasons, although I already rated him last of the top eight winners a long time ago (and his cumulative races led are below all the other seven winners and barely ahead of Rusty Wallace even when counting his entire career) so I do think he's still overrated due to spending most of his career with the two worst cheating teams in NASCAR history. I disregard Jimmie Johnson and Tim Richmond because both of them had significantly higher lead shares than cumulative races led, implying that they were responsible for most of their leading themselves and are only on this list because they were clutch. I also disregard Terry Labonte because he actually had more TNL than wins. However, I think besides Johnson, Richmond, and Labonte, this is a very accurate picture of the most overrated drivers of this period. These were the guys who tended to luck into race wins most, and in ridiculous fashion. Think of how Darrell Waltrip managed to luck into both his only Daytona 500 (1989) and his only Southern 500 (1992) win on fuel mileage. Think of how Bobby Labonte inherited the lead on the last lap two different times and on a rain-ending caution in his only Southern 500 win. Think of how Newman's one supposed great season came because of fuel strategy. Think of Bill Elliott totally backing into the 1992 spring Atlanta race because a caution came out trapping everyone else a lap down when he was the only driver who hadn't pitted. Think of Logano's 2015 hot streak where he won three straight races but was the TNL in none of them (spinning Kenseth and winning the other two in the pits.) Think of Gant's fixation on fuel mileage and strategy in his late career, especially the 1991 spring Talladega race. Now try to think of a time Sterling Marlin backed into a win. It never happened. All ten of his wins were natural. Maybe I was wrong to posit Davey Allison as the most overrated driver, but I do think he is rated higher than these other drivers, except for Waltrip, Johnson, Elliott and Richmond, and Elliott is the only other one of those drivers I likely do feel is that overrated. In Johnson's case, he doesn't really belong here since in his case it's a matter of clutch more than actually lucking into things as his 81 TNL is very close to his 83 wins. Johnson is more like Labonte where he is unluckier early in the race and lucky late, but Labonte benefited from late-race luck far more than Johnson did.

Terry Labonte has one very different thing in his favor here that most other drivers don't however. With a ratio of 18 TNL to 12.993 lead shares, Labonte is the most clutch driver in NASCAR in the period. No other driver with a significant number of wins had a higher ratio between their performance on the final pass and their performance on all passes in general. He was always renowned for taking it easy for the first three quarters of a race before coming along strong at the end and it was obviously an effective strategy as he by all accounts got more wins than he should have based on the level of his dominance, and unlike his brother, did not back into them as Terry actually had more TNLs than wins. This should end any debate on which Labonte brother was the better driver, if there was still any debate. Labonte may have a lead change ratio of less than 50 percent and may have had a conservative style, but he obviously did have the ability to turn it up at the end, and most of the other drivers with that same reputation, like Mark Martin, Dale Jarrett, and Ricky Rudd are not even close. Labonte made that "be there at the end" strategy work. Nobody else really did, at least not to that degree. This may explain why he ended up with two championships even though there are several drivers who beat him in almost every other statistic who have none.

Following Terry Labonte's ratio of 1.385, the next most clutch drivers are Brad Keselowski's 1.345, Carl Edwards (who always comes up in the top two or three in this statistic, no matter how I attempt to calculate it) at 1.331, Kasey Kahne's 1.268, Jimmie Johnson's 1.239, Dale Jarrett's 1.231, Ryan Newman's 1.210, Harry Gant's 1.163, Bill Elliott's 1.146, and Kevin Harvick's 1.131 (not quite the closer, but not bad.) In a way, this is sort of the flip side of the TNL/lead share ratio. This makes sense since TNL will be highly correlated with wins and therefore this will be highly correlated with the reciprocal of the other list. Seeing how clutch Labonte and Johnson and apparently Elliott are in terms of being the last driver to take the lead on track, that does mean they really shouldn't be sharply criticized along with the others. They won more than they were supposed to because they were clutch, not because they lucked into the lead. However, the drivers who do not appear here like Darrell Waltrip, Bobby Labonte, Joey Logano, Davey Allison, and surprisingly (to me) Tim Richmond appear that they were more lucky than clutch, because if they were more clutch their TNLs would be close enough to their win totals that they would appear on both lists. I would say it's pretty clear that Logano, Allison, and Labonte are three of the most overrated drivers of the last 30 years - maybe the most, and as indicated by the previous column, all three of them tended to be better duelists on the high-speed superspeedways when they had fast cars than anywhere else, which is not a good sign. I think I must conclude that late-career Waltrip and Gant were overrated as well. Waltrip does not appear on this list and he had way more wins late in his career than he theoretically should have. He got pretty lucky. Newman and Gant had very similar styles in that they won lots of races on strategy but admittedly did often pass people for the lead who were also on strategy to claim those wins. This does make them both listed as clutch passers, but in most of their fuel mileage wins, they were usually non-factors. I have much more sympathy for Gant however because he had a much bigger equipment advantage, was a much better duelist, and was doing things at an older age when most people would not be competitive. The interesting drivers here are those who appear on the clutch list and do not appear on the lead shares to wins list, which include Keselowski, Edwards, Kahne, Jarrett, and Harvick. Those five drivers all have more TNL than wins and except for Keselowski all have more lead shares than CRL. They are all clearly strong duelists when they have the car to win, as we saw once again from Kahne when he managed to win the Brickyard 400 in almost the only race he was competitive during in the last three years.

The least clutch drivers with four or more wins were Jeremy Mayfield (0.624), Bobby Hamilton (0.713), Kyle Larson (0.719), Ken Schrader (0.728), Kyle Petty (0.825), Jeff Burton (0.864), Ricky Rudd (0.873), Joey Logano (0.883), Jamie McMurray (0.907), and Geoff Bodine (0.909). These were the drivers who made the pass for the win much less often than you'd expect based on their total number of lead shares. Drivers like Mayfield, Larson, Schrader, Petty, and Bodine should not be surprising. They were never regarded as clutch. Burton and Rudd were probably regarded as too conservative. I'm rather astonished to see Logano both in the top ten on this list and on the top ten lowest list on the lead shares/win ratio. Since lead shares/wins would be correlated with lead shares/TNL one would expect that a driver who was lucky enough to substantially more wins than lead shares would also have more TNL than lead shares, but Logano does not. He wins more often than he puts himself into the lead through his own efforts, which is more often than he puts himself into the lead at the end. While McMurray, Schrader, Mayfield, Hamilton, Larson, Petty, Bodine are at least on some of the good lists as well, Burton and Rudd are not, which likely makes them overrated as well. Burton, like Logano, has more wins than lead shares than TNL, which is not a good sign, but it is nowhere near as extreme in his case as it is for Logano's. The fact that Logano is on both negatively-themed lists and on none of the positive reciprocal lists probably means that he rivals Newman as the most overrated driver in the sport, and is probably more overrated now because Newman is no longer highly rated.

Another way to evaluate clutch passing is not to use ratios but to take a difference between drivers' lead change percentages on all passes and drivers' lead change percentages on TNL passes only. The most clutch drivers should theoretically be those who have the largest difference between their overall lead change percentage and their TNL lead change percentage. That table is below. Here I am not going to list everyone as I did above but only list those drivers who were involved in the final lead change of the race ten or more times. TNL counts will slightly differ from the above because there were five races where there were no on track lead changes. In those races, I awarded the polesitters (Earnhardt, Hamlin, Johnson, Kulwicki, and Kyle Petty) one TNL and lead share each.

Even though they had sub-50% lead change percentages, Alan Kulwicki and Terry Labonte were certainly two of the most clutch drivers in terms of passing at the end. Kulwicki was the last driver to take the lead on track twice as often as he was the last driver to be passed, and his lead change percentage on TNL passes was nearly 20 points greater than his lead change percentage in general. The same guys who usually do well in terms of clutch performance (like Carl Edwards, Jimmie Johnson, and Terry Labonte) do so again. Johnson is probably the driver who most deserves the closer title, because his TNL lead change record of 80-34 is staggering, the greatest percentage of any driver in the period who was ever passed for the win (Jerry Nadeau and Ricky Stenhouse managed to have undefeated 3-0 TNL records and Aric Almirola 1-0, but each had far too small a sample size.) He only does not lead the list in terms of being the most clutch because he has a substantially higher lead change percentage than most others, so he overachieved his overall lead change percentage by less than Kulwicki and Edwards did, but he was easily the best clutch performer at the end, and the only person to break the 70% mark among those with a significant number of TNL passes. Harvick as usual was a closer but wasn't the best closer. Prior to his Stewart-Haas years he wasn't any better either. He was actually worse (20-17), so in spite of the myths, he's been a more successful closer in his Stewart-Haas years (22-8) than in his Childress years, and actually one of the best recently. The issue with his Stewart-Haas years is how many races he lost in the pits, which affects his number of wins, but not his number of TNL. Much like Harvick, Martin Truex, Jr., Ernie Irvan, and Sterling Marlin were clutch. They lost races because they were beat out of the pits or unlucky. They did not lose races because they were weak duelists or weak duelists at the end. They were strong duelists throughout, and even stronger at the end. Tony Stewart on the other hand despite being the best passer in general was less clutch than others and wasn't as strong a passer in the end as Edwards or Johnson were. That I imagine comes down entirely to physical fitness. Johnson and Edwards were generally considered the fittest drivers of their generation and Stewart well, was not. That may not matter as much in the early portions of the race but it likely matters at the end, particularly in a hard 500-mile or 500-lap race. Stewart was still clutch at the end, and he was still only behind Johnson, Edwards, and Kulwicki in TNL percentage, but it seems that his lack of fitness compared to Edwards and Johnson prevented him from being the best passer throughout the race, and probably cost him a bunch of additional wins. Three other drivers: Kasey Kahne, Ryan Newman, and Ward Burton had higher TNL than lead change percentages despite having below average lead change percentages. They may all be overrated but at least there's a reason they're overrated. They did perform better late in the race than earlier on.

There's a reason Bobby Hamilton is underrated too. Although he was surprisingly given the level of his equipment one of the better passers of the era, he greatly underperformed in late-race passes and that's why he didn't win more races, although he had unreliable equipment so that makes sense as the stronger teams tend to be better at getting their cars to improve throughout the race. It's not necessarily a reason to knock him. Tim Richmond's stamina may have been affected by his illness much like Stewart's likely was due to his lack of fitness so I won't criticize that one either. Jeremy Mayfield, Jeff Burton, Clint Bowyer, the Waltrip brothers, Ken Schrader, Kyle Busch, Bobby Labonte, Kyle Larson, Matt Kenseth, and Joey Logano all had strong cars though, almost all had sub-50% lead change percentages, and still managed to underachieve even compared to that. Jeff Burton's 18-31 lead change record is pretty astounding for somebody who has an above 50% lead change record in general and drove for Roush and Childress and was not lacking in cars. Burton was the exact opposite of Terry Labonte so it seems. While Labonte was conservative early in the race and aggressive late, it appears Burton was the epitome of the exact opposite: aggressive early in the race and conservative late. It made sense to be conservative considering NASCAR's points reflected consistency more than winning, but I think Labonte's strategy would obviously seem to be the much more effective. One could argue Jamie McMurray and Ken Schrader may have been underrated because their lead shares were greater than their cumulative races led, however both of them indeed struggled at the end of races badly, and that's one reason they may not have won as much as they shold have. Drivers like Earnhardt, Kyle Busch, and even Matt Kenseth (much to my surprise) again appear as not being clutch. Most would expect that from Kyle but not from the other two. Generally though the lists make sense, with many of the too conservative and too aggressive drivers not being clutch in the end.

In conclusion, after undertaking all these analyses, I declare Sterling Marlin the most underrated driver of the last 30 years. No matter what advanced statistic you look at, he has never appeared below average and he has always improved. Even considering 2002, he was a championship-caliber driver (although definitely more in the tier of drivers who didn't win championships like Ricky Rudd or Geoff Bodine) who generally wasn't thought of as such. Although few people will still agree with me solely because of his win total, I think he belongs in the Hall of Fame, although no driver except Jeff Gordon or Tony Stewart does before Smokey Yunick, John Holman, and Ralph Moody, who are long overdue. I declare Joey Logano the most overrated driver of the last 30 years. Despite having top equipment his entire career, he has a below 50% lead change percentage, is not clutch, wins more often than he leads, and leads more often early than late. Jeff Burton is the one major driver who has a very similar profile and is even less clutch, but he was never hyped as being potentially the best driver of his generation like Logano was. Perhaps Davey Allison is still yet more overrated, simply because of how much higher everybody does rate him, but he was certainly a fantastic passer both early and late, and quite clutch as is seen above. The problem is it was only on superspeedways. Allison was only great on horsepower superspeedways, and struggled at passing on ovals less than 1.5 miles. Perhaps Ryan Newman is more overrated but almost everybody sees his success as being mostly due to luck, and because of how weak a passer he was. I suspect Newman will grow to be more overrated in retrospect because of his Daytona 500 win, Brickyard 400 win, all-star win, and 8-win season than he is now, when most people remember the actual circumstances of the wins more than just the original stat lines. However, Logano did not show greatness on any track type, only barely scoring positive records on superspeedways and restrictor plate tracks by a single lead change, and having negative lead changes everywhere else, and he was hyped far more than Burton or even Newman were. Mark Martin may have thought Logano could become one of the greatest drivers ever in NASCAR, but he was wrong.

Sean Wrona is the Managing Editor of racermetrics.com, the Webmaster of race-database.com, the winner of the 2010 Ultimate Typing Championship at the SXSW Interactive Conference in Austin, and the ratings compiler and statistician for the Mensa Scrabble-by-Mail SIG. He earned a master's in applied statistics from Cornell University in 2008 and previously digitized several seasons of NBA box scores on basketball-reference.com. You may contact him at sean@racermetrics.com.

---