Racermetrics race-database.com

IndyCar 2015

by Sean Wrona


In my initial column I discussed a formula for ranking drivers and applied this towards to last year's 24 Hours of Daytona. I know in that column I stated that I would cover this year's very shortly and obviously I haven't but I do intend in calculating ratings for the Tudor United SportsCar Championship soon. If I am able to obtain the necessary data (which I think won't be hard but could be), I will do Formula One next. I started with IndyCar in particular in advance of the month of May and intended to have this column finished in advance of the Grand Prix of Indianapolis. Obviously I did not manage to do so (I had the calculations done, but had not yet started the writing) but decided to proceed with writing the column rather than delaying it any further by including the Indianapolis results. I changed a few things about my formula from last year. I had already calculated these overall scores for every race from the 2014 IndyCar season to give myself experience doing this before I released any actual columns and I decided some changes needed to be made to my initial formulas. Previously when calculating drivers' average speeds I included all green-flag laps excluding pit stop laps but that had the unfortunate effect of including uncharacteristic laps such as spinouts that did not result in caution periods or running out of fuel without pitting, and although these situations don't happen as often as people may think they do (since such situations usually result in cautions) they certainly skew the results and fail to measure the driver's actual pace on competitive laps, so I needed to decide what a competitive lap was. I adjusted this for my Rolex column by considering all laps within 10% of the fastest lap of the race, but that also resulted in problems when there were cars that were considerably off the pace and therefore drivers who had very few if any laps counted, particularly the Mazdas in the Prototype class. For instance, I still want to be able to compare drivers on the same team or using the same engine even if the car is too slow to be competitive with the leaders. Instead, I've decided to count all laps within 5% of that driver's fastest lap as competitive so drivers with naturally slower cars still have a chance to be accurately listed (at least as soon as I come up with a good way to compensate for differences in equipment, which I have not done yet). I may adjust this some more since some people might find 5% to be a very stringent standard (considering Formula One only requires you to be within 7% to qualify) but I think this does tend to remove biases that result from luck and comes closer to measuring pace than anything else I've come up with (although I still might decide within 7% or 10% of a driver's fastest lap could be more accurate in the future). For instance, drivers who are running off by themselves not in traffic are likely to have faster lap times. If someone is in heavy traffic or has to battle many cars to get to the front, the lap times on those laps will invariably be slower and might not meet the standard on some occasions. Usually though on road courses 5% seems to be a good dividing line between drivers who run completely clean lap and drivers who go off-track. If I made the standard something higher than 5%, I think I'd actually end up including several laps where drivers made brief off-course excursions, more than I'd want to anyway. For my measure of passing, I award X points to the driver who passes the fastest driver in the race and -X points to the driver who is passed by the slowest driver. I've decided to only count competitive passes here. If the driver who is passed had a lap time within 5% of his or her best lap time, it always counts. Unfortunately, many of the best passing laps are on restarts when many drivers back in the pack do not meet the standard as they are getting up to speed, so for all instances where the driver passed is not within 5% of his/her best lap I then review the video footage to determine if the pass was clean or dirty (I'm not for instance giving Juan Pablo Montoya credit for knocking Charlie Kimball off the track at Barber, nor do I give drivers credit if they clip a driver's wing on their way past, or sometimes immediately before a driver has a mechanical problem. The idea is to reward natural passes as much as possible, and while determining whether a pass is natural or not is certainly a judgment call, it's better to measure passing this way than to include passes where a driver benefits from another driver's off-track excursion. Sure, you could blame drivers themselves for their errors, but this is already reflected in average running position. I want to measure passing that requires skill separately from measuring average running position. While measuring which passes count and which don't is definitely more art than science, I think I'm doing a better job with it than I was previously. If unfortunately there is no footage of the pass, I then go into the lap times. If the driver passed lost several seconds in one sector (implying an off-track excursion or electrical glitch or whatever) the pass does not count. If however the driver loses time throughout a string of several sectors in a given lap or consistently loses time throughout the lap, then the pass does count. This may unfortunately exclude passes that occurred on the same lap before a driver entered the pits, which will rarely be listed separately, although I did catch one of Graham Rahal's passes immediately before he pitted while reviewing the footage and including it even though it did not appear obviously on the IndyCar lap chart. Average running position includes green-flag laps only, just as NASCAR calculates it. Finish is self-explanatory, and natural peak is the highest position obtained by a green-flag pass. Even though I'm calling it the natural peak, for this I am not (yet) requiring that the pass was natural. If a driver makes a strategic assist pass, I will still count it as natural, so for instance Sage Karam may have ended up with a better score at St. Petersburg than how he actually ran since he did pass Jack Hawksworth on track in the top five even as the drivers controlling the race were quickly disposing of both of them. In IndyCar at least, on-track passes in the middle of pit cycles on road courses aren't that common, so this may be a rare exception, and besides Karam's score still wasn't that good anyway. I've also adjusted how much two of the categories score. Previously I awarded 10% of a driver's race score for speed and 20% for passing, but I have since changed it to 15% each. My argument for undervaluing speed is that I thought it was too car-dependent, but then I realized that all these statistics by their nature are car-dependent and speed (a driver's pace) is likely one of the best predictors for how that driver will do in the future, while my passing statistic is more nebulous since I have to make judgment calls to exclude certain non-natural passes (some people might count bump-and-run passes while I wouldn't, for instance, not that they're going to happen often in IndyCar anymore with the pieces from the new aero kits flying off the cars when drivers make contact), not to mention that the passing statistic will invariably favor drivers who started in the back of the pack...this is mitigated somewhat because I reward passing the fastest cars more, passing the slowest cars less, being passed by the slowest cars more, and being passed by the fastest cars less, but still, a driver who charges through the field will generally be higher than a driver who was up front all race. Calculating passing will be nightmarish for the Indy road course race considering that start where several drivers went off course and I have to decide which passes count; probably I am going to end up deciding none of them since almost all the drivers who passed a ton of cars there including Graham Rahal and J.R. Hildebrand got extremely lucky to make it out of there alive and probably primarily did so because most of the cars in front of them were obstructed by other cars, although I think I covered one or two of the many NOLA incidents differently. When there are so many judgment calls involved in calculating a thing, it can't be worth more than a more objective measurement such as speed, so I adjusted it so each will be worth 15% and I will do that for all series for which I calculate this formula. Drivers are still ranked in all categories using percent beat, with the driver who wins a particular category in a 24-car field receiving 23/23 of the available points for that category, 2nd place receiving 22/23, and so on.

St. Petersburg - Winning vs. Domination

Who is the better driver? The driver who dominates or the driver who wins? This is not at all an obvious question as much as most fans disagree. Invariably, most will say the driver who wins because winning is the object of the game and the only thing that really matters, and that would be fine if racing was a completely individual sport. However, when drivers have unreliable equipment and mechanical failures, run out of fuel, get involved in on-track incidents caused by other drivers, or especially get beaten out of the pits, the driver didn't actually do anything to choke the win. Drivers are representative of their teams but are not the entire team themselves. The 2015 IndyCar season-opener at St. Petersburg was a typical test case. Will Power completely controlled the race from the start until his final pit stop, when he was jumped by his teammate Juan Pablo Montoya, who beat him out of the pits. Who had the stronger run? Most fans will instantly say Montoya because he won, but remember the How the Races Were Won series? Power did not actually do anything himself that caused him to lose the race. If he had stalled in the pits or got into a scrape with a fellow driver (for instance like Simon Pagenaud and Sébastien Bourdais did at IMS today leaving the pits for the first time, which allowed Graham Rahal to jump both of them and take over 2nd), it could be his fault, but in that case it wasn't. Montoya did need to have a killer in-lap to beat Power out of the pits, but ultimately, it was the team strategy to have him pit one lap later and Power having a slower pit stop that made the difference. Essentially this is just like when mid-pack cars go off-strategy to gain track position. Did Montoya have a great drive? Sure. He was the best passer in the race, was 2nd in average running position, and was 3rd in natural peak. It was a stellar drive. I do think however that Power, who led in running position, speed, and natural peak (winning three of the five categories to Montoya's two) had the better drive by the slimmest of margins, regardless of the fact that he ultimately lost the race. Had Montoya actually made the final pass of Power on-track, I would say he would have had the better drive (and I fear that had he made that pass he still might have been behind him in this statistic...that would be an instance where I would call the result erroneous). Does that mean that Montoya didn't 'deserve' the win or anything? Certainly not. You don't have to have the best drive in a race to deserve the win if you were running up front all race and just needed a tiny pit advantage to make the difference, but Power would have been even more deserving, considering he led the entire race until being beaten out of the pits. Besides the Power/Montoya thing where the drivers were very close throughout the race and it was decided by the slimmest of margins in favor of Power's dominance (which I do feel is the best predictor of how future races will go), the race in general was very fair as the drivers with the top finishes also had the top race scores with few exceptions (Luca Filippi was the only driver who had an above average finish despite having a below average run, which I would define as a race score of less than 50), so it appears he was a bit lucky while Stefano Coletti who had a weak finish despite a strong run (especially in regards to on-track passing) was unfortunate. He easily stood out over the other rookies, but for the most part, this race was clearly straightforward so I'm just going to display the results and move to the next one.

DriverRunningFinishSpeedPassingNatural PeakOverall
de Silvestro6.521739135.2173913043.2608695650.6521739135.21739130420.86956522

NOLA - Measuring Chaos

How do you quantify a debacle? The inaugural race at NOLA Motorsports Park was one of the most harshly criticized IndyCar races in the last decade as it ended with a staggering 26 of 32 laps under caution before a mandatory time limit shortened the race (even though there was considerable time in the TV window to complete more laps since the race was started shortly after the broadcast began) leading to James Hinchcliffe inheriting the win because he only pitted once when most others pitted twice, in a Honda when no Honda driver has had close to the pace of the Chevies in the actual races (with the exception of Graham Rahal, whom I would not have expected to be the top Honda driver but it is absolutely inarguable, even though Hinchcliffe has won and Rahal did not; interestingly, Rahal was easily the best Honda driver in this race as well, and Simona de Silvestro, who made a pass in the top five while Hinchcliffe didn't, and Ryan Hunter-Reay were better too). On a personal level, I really did not think the race was that bad, on the order of the Surfer's Paradise CART race in 2002 where Mario Dominguez inherited the win under a rain caution because of mandatory pit stop cycles and time limits, or the Winston Cup race at Loudon in 2000 where NASCAR added restrictor plates to attempt to solve the problem of stuck throttles, or even the IndyCar Toronto and Houston rain races last year. The crashes were all legitimate and understandable. For the most part rookies with little IndyCar experience crashed in drying conditions when IndyCar does not have optional intermediate tires. The tires available just weren't suitable for the track conditions, but even so, most of the veterans did just fine (except for the Bourdais/Pagenaud/Hunter-Reay crash) and I felt people who considered the endless crashing in this race to be an indictment of IndyCar's talent level are wrong. Formula One has certainly had chaotic races too. I think this falls closer to the 2008 Brickyard 400, where the series made the best out of a bad situation. This race, like that race, looked ugly and was below average but wasn't really a contender for the worst ever. I guess I've just watched too many NASCAR races where the cautions aren't always legitimate...these may have been too long but I'd be careful in wet conditions anyway. The only issue I see is that it was a timed race when there was still potential time left to race. Having said that, I still think calling it chaotic or a debacle is right, but it just happens sometimes. Since everything I'm mentioning except for finish is based on green-flag laps only, this basically reflects what happened in the first 15 laps of the race when there were no cautions and Montoya, Power, and Castroneves were 1-2-3 the entire time, solely because qualifying was also rained out (I think throwing out qualifying when every driver had a set a time in the first round is the biggest thing here, since it allowed drivers who did not qualify well like Montoya but finished well at St. Pete to gain a huge advantage). The drivers who did well even though they were further back in the pack by this metric were those who made lots of passes, including Simona de Silvestro, the best passer in the race (primarily due to her jumping from 14th to 8th on the lap 21 restart, although I may have excluded cars that went on track; I'll have to review how I judged this to determine what to do with the Indy restart), Josef Newgarden, and Graham Rahal. Newgarden and Rahal's stellar passing in this race, which was little noticed, may have been an excellent predictor for their future road course success this season. Despite winning the race, Hinchcliffe only had a mid-pack run, but he did have an above average race score, and that does reflect how his race went. I will say his win was more deserved than Carlos Huertas's Houston win and Mario Dominguez's Surfer's Paradise win since he made a lot more passes earlier in the race and has shown the ability to win on all types of circuits in all kinds of fashions before, although it was shocking that he won in the Schmidt car after replacing Simon Pagenaud who has not yet won for Penske. Montoya led the other three categories (average running position, speed, and natural peak), but despite having objectively the best race performance, I think he definitely lucked into that since he failed to advance into the Fast 12 qualifiers in the aborted qualifying session and then was gifted the pole. I thought Montoya's 2014 season was a little overrated, and I didn't feel he really had much staying power as the points leader (and even though he still has the points lead, I still don't), but based on his actual performance, he deserves the best race score even though he certainly wouldn't have gotten that had qualifying proceeded as usual or if the qualifying times had actually been accepted.

DriverRunningFinishSpeedPassingNatural PeakOverall
de Silvestro9.13043478317.391304356.521739131516.5217391364.56521739

Long Beach - Dull but Fair

There is very little to say about Long Beach because very little happened. One can say IndyCar needed a race like this after the carnage in NOLA and that is true. There was only one caution in the race and every car finished. There were only 22 natural passes for position in the entire race, and four drivers neither passed another car or were passed by another car on track in the entire race (Hélio Castroneves, Marco Andretti, James Hinchcliffe, and Luca Filippi). For those four drivers, who were tied for positions 10-13, I gave them credit for the average of those four positions (11.5) and treated them as if they had beaten 11.5 of the 22 cars entered. Scott Dixon beat Hélio Castroneves out of the pits and the main differences between this race and St. Pete were that Dixon did so earlier in the race allowing him to have the highest average running position and he made on-track passes for position, which Castroneves did not. If you discount the passing category, Dixon and Castroneves were tied with the exact same race score, but Dixon's passing makes the difference. Perhaps this is unfair because Castroneves started on the pole and it would have thus been very difficult for him to pass anyone unless he fell back. This is still an off-track win just like Montoya's but I am more impressed since he did lead a greater percentage of the race and the Ganassi cars do not seem to be as strong as the Penske cars this season. Perhaps surprisingly, Charlie Kimball was the fastest driver; this is probably because he went considerably off-sequence for most of the race and was not racing in traffic but it's still notable. This race was very fair because the drivers who had the highest average running positions ran well for the entire race and finished well, with most drivers' actual finish being within a position or two of their average running position rank. When you have few cautions and every car finishing the race, you'll get a purer result even if the raw excitement isn't too much. As usual this came to the margins with Dixon leading two categories (average running position and the race) while Castroneves only led one (natural peak). Sébastien Bourdais quietly was the best passer in the race.

DriverRunningFinishSpeedPassingNatural PeakOverall

Barber - Drives of the Year

Josef Newgarden and Graham Rahal scored what were arguably the most impressive drives of the year in the fourth race at Barber Motorsports Park. In a year when the Penske cars seem to considerably have more speed than any of the other teams (even though they only won one of the first four races), Newgarden and Rahal put on a clinic to earn a surprise win and second place finish, although based on their passing ability at NOLA in wet conditions perhaps this should not have been much of a shock. Despite driving perpetually barely sponsored equipment for CFH Racing, he passed three of the principal series stars, Simon Pagenaud, Scott Dixon, and Will Power, all in Penske or Ganassi cars, on the opening lap, then later passed Castroneves on track for 9th position mid-race to take effective control of the race even though Rahal, who was on a different pit sequence was leading. Rahal was not the only driver on the alternate pit sequence but he sure made the most of it showing blistering speed and passing numerous cars as if they were standing still (which to a degree they were as he did not need to conserve fuel while he was passing drivers who were). Regardless, Rahal's drive was awe-inspiring even to the many critics he had in the 2013 and 2014 seasons, including myself (I actually think his Indianapolis drive was even MORE awe-inspiring, as he finished almost 40 seconds ahead of the next highest Honda, but I haven't calculated that race yet). While in most cases the fastest driver and the best passer may not be so obvious, especially in races with lots of cautions, everybody who watched that race was well aware that Rahal was both the fastest driver and the most productive passer, and this was even more impressive because Honda had a weak performance at Barber except for Rahal, with only James Hinchcliffe in 5th and Ryan Hunter-Reay in 9th making the top ten in race score. Newgarden, who controlled the race after passing Castroneves, led the other three categories (average running position, finish, and natural peak). When I was starting this sort of research, I did not count passes for the ultimate lead even if the pass that ultimately became the pass for the lead was in mid-pack. However, gradually as I continued work on the How the Races Were Won series I decided what mattered was taking control of the race, not actually taking the lead. If you take control of the race from the previous driver who had control of the race, even if that driver was not leading, it should definitely count as natural, and I eventually started adjudicating races where the ultimate lead change was not technically for the lead this way, even though I was somewhat inconsistent in that series on this (there aren't too many races where something like this happens). Regardless, Newgarden's pivotal passes all came on track and it was the only natural win of the IndyCar season at that point to date, yet Newgarden and Rahal theoretically had much worse cars than the Penske and Ganassi drivers and thoroughly blew them out. Who had the more impressive performance? Most people who were watching thought Rahal did. I disagree. Newgarden passed Castroneves, Dixon, Pagenaud, and Power when they had equal fuel, while Rahal made most of his passes when he had a significant fuel advantage. Newgarden was clearly more deserving for the win, and had the best drive of the year to date, especially considering his equipment deficit (sure, CFH is better than the Sarah Fisher team he drove and almost won for last year, but it's still not as good as Penske or probably Ganassi, even though it looks like it's surpassed Andretti this year and maybe even its predecessor Ed Carpenter Racing did so last year). Newgarden and Rahal not only had what I consider to be their best drives ever (I definitely consider Rahal's second place in this race and the second he earned at Indianapolis as well to be more impressive drives than even his win at St. Petersburg was...Honda has an extreme disadvantage on the road courses and he's the only one of their drivers seemingly able to compensate).

DriverRunningFinishSpeedPassingNatural PeakOverall


If you don't think Penske has been dominating this season start-to-finish (Penske, not just Chevy) because they only won one race, think again. Entering the Grand Prix of Indianapolis, the four Penske drivers had a clean sweep of the top four positions in terms of overall average race score, with Castroneves leading, Pagenaud surprisingly second (this indicates that his 9th in points is VERY unrepresentative of his performance to date, points leader Montoya in third, and Power in fourth. Penske had a virtual lock on the top five in the first three races and it was mostly bad luck that kept them from a clean sweep of the races like they swept the pole positions, although the variety of winners (five winners for four different teams) has certainly been more exciting than it could have been. Penske wasn't very far ahead of Josef Newgarden, Sébastien Bourdais, and Graham Rahal, who were surprisingly all ahead of the more expected championship contenders in the historically better cars, Tony Kanaan, Scott Dixon, and Ryan Hunter-Reay (it's also surprising that Kanaan is ahead of Dixon and Hunter-Reay since he has had less success on road courses). Newgarden and Bourdais were constant consistent threats as much as any Penske car (although I'm sure Newgarden's running position average plummeted at Indianapolis, but he was still very fast holding off Will Power's efforts to put him a lap down much longer than I expected). Despite driving for the theoretically worse KV Bourdais's speed rivaled all the Penske cars. My instinct that Bourdais has been a fairly overrated points leader and hasn't been as fast as Castroneves and Power is borne out by the data, as they were the fastest two drivers this season, and Montoya was considerably behind all three of his teammates and even Bourdais in weaker equipment. However, Montoya's racecraft probably exceeds Castroneves's and Power's which allow him to make up the difference. Pagenaud is seriously unlucky to be 10th in points right now (yes, you can argue he caused the NOLA crash, but you can also argue Hunter-Reay did, and his mechanical failure at Indy was brutal even though he was as usual a constant threat in that race before it). Don't expect that to last. Average finish pretty much goes largely as the points standings do so it's the least interesting. Newgarden and Rahal again seemed to be pulling well above their weight as they had less speed than Tony Kanaan and both KVSH drivers Bourdais and Stefano Coletti but matched Bourdais in performance and blew out Coletti. Coletti still looks way more impressive than his fellow rookies Gabby Chaves and Sage Karam though based on his speed and passing...when he puts a clean race together, as he did at Indy, he should definitely be the highest-finishing rookie more often than not...his experience racing GP2 in Europe definitely seems to prepare him for IndyCar road racing, yet too many American fans deride him for being a 'ride-buyer'. Being a ride-buyer doesn't matter much if the driver is qualified, and that overlooks the fact that Karam and Chaves had money to bring as well, just not as much. Coletti is qualified and is being vastly underrated by the IndyCar fan public at the moment (holding off Tony Kanaan at Indianapolis as long as he did was very impressive even if Kanaan isn't the best road racer in the series). Karam, by contrast, has been disappointing except for the natural peak statistic, which may be generous in his case because of how I chose to evaluate his natural peak at St. Pete. Comparing teammates, Bourdais has a massive advantage over Coletti, Newgarden has a lesser advantage over Filippi, Kanaan and Dixon have a huge advantage over Kimball and Karam, Hunter-Reay has been completely invisible but is still besting his teammates, and the other teammates seem close. While I think people are wrong to criticize Coletti, criticizing Dale Coyne Racing may be yet another matter. Francesco Dracone has indeed been as bad as advertised. He was the slowest driver in all four races and usually by a significant margin to the point where it is almost an upset if he does not finish last in each category in every race, but I still want to point out that Coyne's team in general has been terrible, as the other three Coyne drivers, Rodolfo González, Conor Daly, and Carlos Huertas, are just as far behind James Jakes (the lowest non-Coyne driver) as they are above Dracone, so clearly the team has issues beyond Dracone considering how little money they have and how a team with so little resources can contend having to buy all new equipment and aero kits, not to mention that changing drivers every week makes it even harder to build setups. Having said that, I wasn't impressed with any of their drivers' performances, including Daly, whom I think has been overhyped because he was an American racing in Europe and because his father was a past driver and broadcaster. I do definitely think Daly is better than the other three though. Overall, Castroneves led in running position (by quite a lot), was tied with Montoya in finishing position (but Montoya surpassed him at Indianapolis), was again fastest in speed, was weak in passing because he generally started up front (no surprised that Graham Rahal has been the best passer of the year so far), and ahead in natural peak. Castroneves led four of the five categories but has not won a race and was not leading the points at this point. Does that sound right? Possibly not. What this is measuring is consistent race performance, while all the winners and dominant drivers had some off races, except Montoya, who finished one position better than Castroneves when they both struggled at Barber. I guess there really hasn't been a single-season standout as far as drivers are concerned to this point (although it wouldn't surprise me at all if Power took the lead after Indianapolis, since he was the biggest standout last year as well overall and I'd say now is this year even though he's still second in points because of his bad luck at Long Beach). There has definitely been a standout in terms of teams though. The parity is only existing at the end of the race really. Penske has been destroying the field as expected. The Barber race was thrilling because Newgarden and Rahal managed to fight with and pass the Penske cars, which no one expected, and I do believe those two, Bourdais, Dixon, and Kanaan are the only ones who seem likely to fight for wins on pace with Penske on the road courses. This obviously does not adjust for equipment differences, but when you do that, I actually think Newgarden and Rahal have been as impressive as any Penske driver, perhaps more, but Pagenaud and Luca Filippi in particular are doing a lot better than you think.

DriverRunningFinishSpeedPassingNatural PeakOverall
de Silvestro7.82608695711.304347834.8913043487.82608695710.8695652242.7173913
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.