Welcome to the second part of this five-part series, in which I apply the f1metrics model of driver and team performance to simulating historical hypothetical situations. Consider this a quantitative approach to tackling some popular but difficult to resolve talking points in Formula 1 history. The point of this series is not to be taken too seriously, so please try to enjoy it in that spirit!
To review the previous article, see here:
For each hypothetical below, I simulated either the extension of a driver’s career or a change in teams during their career. Two types of outcomes are presented:
(i) The Driver Performance (in ppr) for each season. This is a ranking of absolute driver performance, taking teams out of the equation. In other words, this is how the model predicts the drivers would have relatively performed in equal machinery. This is the same approach I use for my end of season driver rankings each year. It takes a driver’s age and recent experience into account.
(ii) The predicted World Drivers’ Championship standings for each season. In cases where these tables are presented, I am including the effects of a driver’s hypothetical team on their points scored. To generate these, I mapped the model’s scoring rate function to all other historical points scoring systems.
How long could Jim Clark have dominated?
Jim Clark stood head and shoulders above his peers in the mid-1960s and remained the sport’s top driver, aged 32, at the time of his death in 1968. Notably, in my all-time driver ranking list from a few years ago, he claimed the all-time #1 spot, ahead of all-time greats such as Schumacher, Fangio, and Stewart. In short, Clark was something really special.
It’s natural to wonder how much longer Clark could have dominated the sport, especially since drivers tend to stay near their peak at least into their mid-thirties. Had he survived into the 1970s, could he have beaten Rindt, Stewart, and Fittipaldi? Could he have even still been competitive into Lauda and Hunt’s era? It’s crazy to consider, but based on his age and talent, entirely plausible.
The graph below shows Clark’s predicted performances all the way up to 1990 (age 54), relative to other drivers’ performances across that time period.
These predictions indicate that Clark could have realistically been a top driver (perhaps even the best driver on the grid in some seasons) up to about 1974. Imagining Clark dicing for position with the likes of Reutemann and Fittipaldi is jarring, since they tend to be mentally compartmentalized into different eras, but that’s what could have unfolded. Based on his projected performance levels, Clark would have likely remained capable of taking race wins and even potentially titles in a strong car up to about 1977 (aged 41). By 1980 (aged 44, the same age at which Brabham retired from Formula 1), he would have been well past his best, and if we imagine he decided to just continue on racing for the hell of it, he likely would have become the worst driver on the F1 grid as late as 1990 (aged 54). An absurd, but illuminating, result.
This forecast assumes Clark would have smoothly adapted to the advent of wings, which were just beginning to sprout in his final season, followed by the relentless piling on of downforce throughout the 1970s. But given how other drivers adapted to this change, there’s little reason to think Clark could not have done the same. After all, he was one of the most versatile drivers in history, excelling in all manner of single seaters, sports cars, and touring cars. Let us not forget that Clark was involved in the nascent development of wings for Formula 1 before they appeared in races. Inspired by his own experience with a Vollstedt IndyCar that featured small wings, he encouraged Lotus to begin experimenting with their own.
As far as team selection goes, it’s difficult to know what course Clark’s career could have taken. He had reached revered driver status by 1968, meaning were he to approach any team for a seat, they would gladly have taken him. Lotus, however, remained the great innovators and generally the team to beat at that time. Ferrari were nowhere in F1 in 1968-1969 due to devoting resources to the World Sportscar Championship, so a move there probably would not have been seriously considered earlier than 1970.
Let’s suppose in this hypothetical that Clark remained loyal to Chapman and Lotus. The first order of business would be Clark winning the 1968 season, which he had started with a race win before his untimely death. By the model’s estimate, Clark would have stormed to this title by 15 points (back when a victory was 9 points) ahead of teammate Hill. In this timeline, Hill would therefore win only one championship.
Following Clark’s death, Chapman appointed Jochen Rindt for the 1969 season. At that time, Rindt was an extremely promising young driver. He had won 5 of the 10 races in Formula 2 in both the 1967 and 1968 seasons. He had also outclassed Jack Brabham in the highly unreliable 1968 Brabham-Repco, taking two poles and two podiums, while Brabham could qualify no higher than 4th.
In the event that Clark survived and won the 1968 championship, Chapman would have needed to choose between the safe option of retaining Hill (now 39) or picking up Rindt. I will assume Chapman chose the ambitious option and fielded Clark and Rindt as his two drivers, putting Hill into an earlier retirement.
With Rindt still finding his feet, the model sees Clark as the top performing Lotus driver in 1969. But neither driver would have been any match for Stewart in his dominant Matra MS80. This was a difficult season for Lotus. Chapman had planned to transition from the Lotus 49 to the new four-wheel drive Lotus 63, but the latter proved less competitive than the former.
In 1970, Lotus had a very competitive package again. Stewart was performing near the top of his game, but the new March 701 was a difficult car to drive and fell behind in development as the season wore on, effectively extinguishing his hopes of defending the championship. To compound matters, the new Tyrrell 001 chassis, to which Stewart switched in the final three rounds, was incredibly unreliable.
Instead, there was a renewed threat from Ferrari, who switched focus back from sportscars to F1 and had arguably the season’s best car. Meanwhile, Rindt delivered what is rated by the model as one of the all-time best season performances, winning all 5 of the races in which he did not have a mechanical DNF. By the model’s estimation, this would have been enough to outperform Clark in a typical season for the Scot. The outcome of the championship would therefore rest on Monza 1970, where Rindt was tragically killed by a brake failure and thus did not participate in the last four races. In the hypothetical, Clark wins the championship by virtue of these events. If Rindt had instead survived in this timeline and participated in the last four races, he would be the predicted champion.
For 1971, having lost his two star drivers (Rindt and Clark), Chapman promoted the inexperienced but very talented Fittipaldi to team leader, with Wisell as a clear number two driver. With a strong car, and without any similarly brilliant drivers near the top of their game, Stewart took an easy title.
In this hypothetical timeline, we can imagine Clark paired with Fittipaldi instead. By the model’s estimation, the performance of the Lotus 72C/D was actually well poised with the performance of the Tyrrell 001. We could have thus seen something close to a fair fight between Clark and Stewart.
The model predicts a narrow victory to Stewart, in what would likely have been a championship featuring some classic head-to-head duels between the two close friends and two of the sport’s all-time greats.
By 1972, Fittipaldi was performing at his best, and the Lotus 72D was more than a match for Stewart’s Tyrrell. By the model’s estimation, Clark would still have been a stronger performer than Fittipaldi, yet in the simulation Fittipaldi comes out narrowly ahead on points to win the title. Why is this? The answer is simply car reliability. Fittipaldi had 3 mechanical DNFs in 12 starts, whereas his teammates had 6 mechanical DNFs in 12 starts. Since the model simulation considers the team’s overall reliability, Clark would lose the championship given an average run of reliability, due to Fittipaldi’s relatively fortunate run (cf. Hamilton losing the title to Rosberg in 2016). Of course, we could get into the question of whether Fittipaldi’s weaker teammates were getting the same mechanical service as he was, but that’s further than I want to go with this.
The 1973 season was Jackie Stewart’s swansong, to the surprise of many. He retired at age 34, with three championships and sitting top of the all-time leaderboard for total wins. At Lotus, it was a close fight for 2nd and 3rd in the championship. Fittipaldi beat Peterson by 3 points, but had two fewer mechanical DNFs, giving Peterson the higher performance rating based on points per counting race.
In the model simulation, Clark is predicted to take the fight right to the wire with Stewart. Note, however, that this is under the assumption that Stewart still skipped the final race following the death of his teammate, Francois Cevert. With the championship still in the balance, Stewart might have made a different decision.
Beyond 1973, the Lotus team tailed off in performance due to the abortive Lotus 76 project, to the extent that even a driver of Clark’s abilities would not be predicted to challenge for titles. By the time they became dominant again with the mastery of ground effects on the Lotus 79, Clark would have been past his best. It’s plausible that Clark would have simply retired at the end of 1973, aged 37, joining his compatriot Stewart.
In conclusion, had he avoided that fateful Formula 2 race on a wet day at Hockenheim and remained at the Lotus team from 1968-1973, Clark would almost certainly have won the championship in 1968, may also have won in 1970, and would have been at least in the championship hunt in 1971, 1972, and 1973. The results in some of those years hinge on car reliability, which was a very significant factor in the shorter championship campaigns of that era. In the model’s simulated scenario, Clark finishes his career as a four-time champion, with four 2nd places in the championship, as listed in the table to the left. A stellar career, much more befitting of Clark’s incredible talents than the two titles he ultimately achieved.
Lost American talents:
Donohue & Revson
Since the 1990s, cross-overs between the top US single-seater series (e.g., IndyCar) and Formula 1 have been generally unfavorable for the US side. Emerson Fittipaldi and Mario Andretti remained championship contenders in CART well into their late 40s, reflecting a lack of comparably talented younger drivers in the series. Mansell was an immediate success when he moved across to the CART series in 1993, whereas Michael Andretti was hopelessly off the pace when he moved in the other direction. Alex Zanardi was a fairly weak F1 driver during both stints, but was quite successful in CART. The only obviously great export from IndyCar to F1 in the 1990s was Jacques Villeneuve.
As I showed in last year’s junior driver analysis, today’s IndyCar grid is overall a much weaker field than today’s Formula 1 grid. If we go back to the 1960s-1970s, the balance of power is not so evident. Mario Andretti and Dan Gurney were both very strong F1 drivers who moved frequently between US and European series. Phil Hill also won a championship, albeit with very favorable machinery. In addition, there are two decorated US drivers who arrived in F1 late in their careers and were both killed in F1 cars. They are Pete Revson and Mark Donohue. Some believe they could have been among the top F1 drivers of the era.
Donohue was an extremely versatile and intelligent driver, who excelled in sportscars, sedans, prototypes, and single-seaters. His book, The Unfair Advantage, which chronicles every car he ever raced, is a classic part of motorsports literature. He dominated the Can-Am series in the early 1970s (a time when Can-Am cars approached and sometimes exceeded F1 pace) in partnership with Porsche and Penske. He also won the inaugural International Race of Champions, winning 3 out of 4 races in a field that boasted Fittipaldi, Hulme, Unser, Foyt, Revson, Petty, and Pearson.
Donohue didn’t debut in F1 until 1971, when he was 34. He started only one race that season and finished 3rd — the next driver to score a podium on debut was Jacques Villeneuve in 1996. Donohue attempted a return to F1 in 1974, but died aged 38 of head injuries following a crash caused by a tyre failure during the 1975 season.
Revson first raced in F1 in 1964, but didn’t attempt a full season until 1972, by which time he was aged 33. In the meantime, he was a top competitor in North American series, including Can-Am and Trans-Am. In his first full season of F1, he was competitive with his ex-champion teammate Hulme. In his second full season, he won two races and outscored Hulme. Due to a strained relationship with McLaren team boss Mayer, and Emerson Fittipaldi joining the team, Revson moved to the Shadow team for 1974. He was killed early in the season by an accident caused by a front suspension failure.
How strong could Revson and Donohue have been if they had committed to F1 earlier in their careers? In the hypothetical below, I have imagined Revson continuing in F1 from his debut in 1964. In Donohue’s case, I have imagined him debuting in F1 in 1968 (aged 31), at the time when he was instead debuting in the top single-seater series in the US.
Both drivers’ projections have considerable uncertainty; especially Donohue’s, since he completed only 10 counting races in F1 (i.e., races without a non-driver DNF). Nonetheless, the projections indicate that both drivers would likely have been among the top 5-10 competitors on the grid. In Donohue’s case, it’s even likely that he would have been one of the sport’s best drivers in the early 1970s. In the right car, he very feasibly could have been an F1 champion. Revson was probably not quite at the same level of talent as Donohue, but nonetheless would have been a very strong competitor had he committed to F1 earlier.
Robert Kubica: the injury and the potential comeback
Kubica’s short but spectacular career in Formula 1 had the hallmarks of future success. Across 2006-2009 he was almost perfectly matched with the more experienced Heidfeld: the net tally was 29-28 to Kubica in qualifying, 25-24 to Heidfeld in races (excluding mechanical DNFs), and 150-137 to Heidfeld in points. In 2010, Kubica seemingly took his performance to a new level, establishing himself as a potential challenger to the sport’s elite. He dragged the Renault into positions it shouldn’t have been and some of his qualifying laps from that season remain among the finest I have seen. The model ranks him the 3rd best performing driver in 2010, the highest ranking in his career. He was perhaps on the cusp of something even greater.
Sadly, it was not to be. Kubica suffered a horrific injury in a rally accident before the 2011 season, leaving many unresolved questions. How would we have performed at Ferrari, where he was signed to drive in 2012? Alternatively, could he have taken the 2012 Lotus to the title, had he raced in Raikkonen’s place?
In the following years, Kubica demonstrated he had lost little of his speed, with an impressive switch to rally, motivated partly by his physical limitations in the tight confines of a single-seater cockpit. He won WRC-2 and then competed in the World Rally Championship. At the top level of rally, he showed occasional glimpses of speed, but was too inconsistent to achieve strong results. Recently, the Kubica story took a new twist, as Kubica prepared for a potential Formula 1 comeback. He narrowly missed out on a seat at Williams for 2018 to Sirotkin, but remains currently on the hunt for an open seat. If Kubica could recapture his 2010 form, he would be a welcome addition for almost any team. But realistically, even discounting effects of the profound injuries on performance, how competitive could a driver be after 7 years out of F1 racing?
These are questions we can attempt to answer with the f1metrics model. The graph below charts two predicted trajectories for Kubica. The green curve shows Kubica’s most probable trajectory if he had never experienced the 2011 injury and continued racing from 2011 onwards.
F1 driver performances tend not to systematically improve after four full seasons of experience, or after the age of ~26. In other words, Kubica was likely operating near the height of his powers in 2010 (the model actually thinks he had better than average form that year) and would most likely have maintained a similar performance level for up to a decade. In the most optimistic forecast (top of the green region), he might have challenged the very elite drivers (e.g., Hamilton, Alonso) for top performer in certain seasons. In the more probable forecast, he would have been a top driver, but not an all-time great.
The red curve shows how Kubica would be expected to perform following a 2018 comeback to the sport, based on average rates of improvement with experience across all the fitted data from F1 history, including past driver comebacks. Based on model fits to historical data, by the time Kubica would likely reach his maximum potential (around 2020-2022) he would be aged 37 and contending with age-related decline, meaning he likely wouldn’t ever quite attain his 2010 high point again.
As beautiful a story as a Kubica comeback would be for the sport, the model predictions aren’t overwhelmingly positive about this scenario. Note, however, that different drivers are known to improve at different rates and also undergo age-related decline at different rates, so there is a wide band of uncertainty on this prediction. Additionally, in these comeback predictions, there is no way of objectively estimating the performance impact of Kubica’s injury. It could certainly not improve his performance, but the degree of deficit remains in question. The red curve is thus a simulated best-case scenario, assuming zero impact of the injury on his performance.
Would Kubica have performed better or worse than the current Williams drivers, Sirotkin and Stroll, if given the 2018 seat? A quick glance at last year’s driver rankings shows Stroll at 7.05 ppr, but that is potentially misleading. As I noted at the time, Stroll’s ppr was inflated by that outlier 3rd place in Baku, where he was on course to be beaten by Massa prior to Massa’s car failing. If the 3rd place is replaced with a mechanical DNF, Stroll’s 2018 performance rating drops to 5.6 ppr, which is probably more reflective of reality. Under that assumption, the 2018 Williams drivers are currently rated at 5.3 ppr (Sirotkin) and 6.3 ppr (Stroll). Referring to Kubica’s projected red curve above, we can see that the first point is at 6.3 ppr, rising to 7.0 ppr in the second season. These results suggest that Kubica would have performed at a similar level to Stroll and better than Sirotkin, with scope to improve from there. Beyond 2018, the comparison is complicated by Stroll and Sirotkin also improving with age and experience, but taking those factors into account still leaves Kubica with a probable small edge over Sirotkin up to 2022.
Now, let us return to the hypothetical scenario of Kubica never sustaining his injury — an oft-considered F1 hypothetical. Where would that have taken him?
First, we can assume that in 2011 Kubica would have raced for Renault (Lotus Renault GP) as planned, taking the seat that Nick Heidfeld and Bruno Senna actually filled. This was a difficult season for the team, as the R31 proved uncompetitive. A novel forward-exhaust concept did not pay off. Consequently, the team switched focus early to 2012 development, which ultimately paid dividends for them.
The model sees Kubica as likely the 4th best performing driver in 2011, narrowly ahead of Hamilton, who experienced his career-worst season. In the simulated season, Kubica scores an impressive 116 points for 7th in the championship (compared to Heidfeld’s 34 points and Senna’s 2 points with half a season each) — elevating the team to 151 points, nearly enough to overhaul Mercedes’ 158 points for 4th in the World Constructors’ Championship.
For 2012, Kubica had a signed agreement to join Ferrari, ousting Felipe Massa. This would have placed Kubica in a potential championship fight for the first time, but also in the unenviable position of facing Alonso as a teammate.
In the model’s simulation of the 2012 season, Kubica scores 171 points — significantly more than the 121 points Massa scored, but not enough to drag the unwieldy F2012 into championship contention. Alonso still loses the title narrowly to Vettel, but note that the model does not implement team orders. While Massa was able to provide a boost to Alonso’s chances in the final two rounds of 2012, the rest of the season he was largely a non-entity, unable to extract enough from the car to take points from Vettel or directly aid Alonso. A faster driver in the second Ferrari would have opened additional strategic options and could have selectively taken points from Vettel but not Alonso — perhaps enough to change the title outcome.
There’s an additional wrinkle to this story, which is the 2012 Lotus-Renault. The E21 (the result of Renault’s early switch to 2012 development) was a highly competitive car, fast enough to take Grosjean to 8th in the championship in his first full season, despite crashing out of 6 of the 20 races and being banned from another. If, for whatever reason, Ferrari had not taken the option on Kubica, he may have started the season for Lotus.
It is worth speculating whether a quicker driver could have challenged for the title in the E21, particularly considering that Raikkonen was on return from a 2-year break and was soundly beaten when facing Alonso and Vettel in subsequent seasons. The number of races where Lotus came tantalizingly close to a stronger result adds substance to this argument.
Fact: If we imagine a driver who finished each race 10 seconds quicker than Raikkonen in 2012, they would have won the championship (273 points to 269 points).
On the other hand, to Raikkonen’s credit, he was extremely effective in managing the 2012 Pirelli tyres (often sacrificing qualifying performance for a better race strategy), and finished every race.
Looking at the first column (Driver performance) in the 2012 simulation above, the model does not see Raikkonen as 2012’s strongest performing driver, but it does rate his performance highly: 4th best, and very similar to Kubica’s predicted performance. In other words, the model thinks that Kubica would have scored a very similar points tally to Raikkonen, had he been in the Lotus seat that year instead.
Overall, these predictions support the prevailing view of Kubica as a potential top driver who would have challenged anyone for wins in the right equipment, had his career not been so dreadfully curtailed. But it is a step short of putting Kubica among the generation’s all-time greats.