2015 model-based driver rankings

Last year, I presented model-based rankings for all the 2014 season drivers. These rankings are derived from a mathematical model, described here, that statistically estimates the strength of driver and team performances in each year, as well as overall career performances for drivers. It achieves this by finding the best fit to all race result data from 1950 to the present. As new races are completed, model rankings are continually updated.

Like all models, this one isn’t perfect, and I’m in the process of revising the model to include additional factors, such as driver age (more on age effects below). Nevertheless, it provides a unique and self-consistent way of ranking performances in a sport where it is notoriously difficult to compare drivers between different teams, and where we are usually left to bicker over the vagaries of subjective analysis.

2015 driver rankings

The metric the model uses to assess performance is the number of points scored per “counting race” across a season. Non-counting races include mechanical DNFs and technical disqualifications. While this is a fairly robust metric for assessing performance across a driver’s career, it can occasionally be unrepresentative in a single season if a driver is particularly unlucky (e.g., most of their best performances occurred during non-counting races). Such issues are discussed in the individual driver entries below.

Driver rankings are quantified by the adjusted points per race (ppr) up to a maximum of 10 ppr, which represents how the model thinks each driver would have performed in completely equal cars.

At the top of each driver entry, I have added a graphic showing the driver’s record relative to their teammate in points, races (excluding non-counting races such as mechanical DNFs, DNSs, and technical DSQs), and qualifying (excluding sessions where a driver was beaten because they set no time due to a mechanical failure).

The Toro Rosso Drivers

Before we get to the list, note that the two Toro Rosso drivers are unranked. This is because they have only driven alongside each other in Formula 1, so they cannot be compared to other drivers via teammate links. For the same reason, the 2015 Toro Rosso is unranked. If I assume the performance of the Toro Rosso car was on average halfway between Lotus and Sauber, the model inserts Sainz at 15th in the list of 21 drivers and Verstappen at 5th. This puts Verstappen in contention for the model’s rookie of the year title, alongside one other new driver, discussed below.

Sainz and Verstappen make for a fascinating comparison, given their different backgrounds. Sainz raced for five years in single-seaters, including one and a half seasons in Formula Renault 3.5. Verstappen spent only one year in single-seaters before jumping to Formula 1 — a new record for the shortest duration between first car race and Formula 1 debut. These junior career differences are summarized by my junior career metric. Sainz’s junior achievement score of 104 places him around the middle of the F1 field, but his excitement score (a measure of how quickly a driver progressed) of 12 is relatively low. Verstappen’s junior achievement score of 54 is low due to his brief career, but his excitement score of 27 is very high.

At the beginning of the season, Sainz and Verstappen seemed difficult to separate in qualifying and race sessions. Given Sainz’s prior experience in powerful cars, this seemed a bad omen for him. If Verstappen’s talent was sufficient to compensate for less experience, one would extrapolate an advantage once both drivers got to grips with Formula 1. So it proved, with Verstappen emerging as the more exciting of the two prospects in the latter half of the season, although Sainz’s progress was not helped by a string of mechanical failures. On a qualifying or race comparison across the season, the pair were tied. The big difference was in the points haul; when Verstappen was good he was utterly brilliant. Given this is still a developing battle, a second year together should give a more decisive picture of their relative abilities, although we will still remain limited in ranking them relative to the rest of the grid.

19. Pastor Maldonado, 4.35 ppr

2015 marked Maldonado’s fifth season in the sport, and his most underwhelming season so far. Last year, he often kept Grosjean honest, tying him 5-5 in races, but trailing 8-2 in points and 11-4 in qualifying. This year, the gap extended. Maldonado’s efforts were not helped by a string of hapless incidents, which some may consider karmic, but it was a disappointing performance however one views it.

In many ways, Maldonado’s career continues to resemble Vittorio Brambilla’s. Both could be quick on their day, but they share a violent, accident-prone style that undermines serious success. Both also share a single grand prix victory in their second Formula 1 season; Brambilla famously spun and crashed as he took the checkered flag.

Brambilla’s damaged car on his victory lap at Austria 1975.

Brambilla’s career dried up after his fifth season, with only sporadic drives thereafter. With the help of PDVSA money, Maldonado will start his sixth season in 2016. How long the funding can continue is unclear.

With Grosjean to be replaced by Palmer — a driver with a similar junior record to Maldonado — and a likely reversion to Renault engines, Lotus seem in very bad shape for 2016. If not for a desperate need to pay their bills in their short term, it is doubtful either Maldonado or Palmer would have acquired a seat. If FOM and Renault can resolve their differences, it will likely be a transitional year before a more committed works effort for 2017.

18. Will Stevens, 4.43 ppr

Early in his career, Stevens was managed by Brundle and Blundell, and even tipped as “the next Lewis Hamilton” by British press. His junior results since then have been underwhelming. He won only 5 races in his 7 years of single-seaters, never finished top three in any series, and was outqualified by teammate Kevin Magnussen by three-quarters of a second on average when they were rookies in Formula Renault 3.5. His junior achievement score is the lowest of any Formula 1 driver since Yuji Ide.

While Stevens’ results never merited consideration in Formula 1, he clearly played the other side of the game very well, maintaining connections within the sport and acquiring key sponsors for Manor to leverage a drive. For most of the season, Stevens was alongside Roberto Merhi, making by far the weakest driver line-up on the grid. While Merhi’s junior career was also weak by Formula 1 driver standards, Stevens’ early season advantage was surprising. Across the year he beat Merhi convincingly in qualifying, aided by a weight advantage of 10-15 kg. In races, the pair were closely matched. Rossi proved a more difficult opponent overall.

17. Nico Hulkenberg, 4.76 ppr

Testament to Hulkenberg’s enormous talent is the fact that he has lasted five (going on six) years in Formula 1’s midfield despite bringing virtually no sponsor money to his teams. Between his incredible junior career and his recent Le Mans victory, it seems Hulkenberg is unbeatable in anything other than a Formula 1 car. In Formula 1, the results continue to fall short of high but gradually lowering expectations, with 2015 the most disappointing year so far. For years now, Hulkenberg’s name has been raised in conversation whenever a top seat is available. His results against Perez may have finally ended that discussion. In 2015, Hulkenberg was consistently outmatched by his teammate, with the exception of qualifying. The overall tally for Hulkenberg-Perez is now almost equal: 154-137 in points, 16-16 in races, and 23-15 in qualifying.

16. Roberto Merhi, 4.91 ppr

With substantially less backing than Stevens, a spotty junior CV, and a poor start to the season, Merhi’s seat always seemed at risk. By the end of the year, he was able to almost level the tally with Stevens, and the model has ranked him slightly ahead of Stevens by virtue of his better finishing positions. I think it is doubtful that we will see Merhi in Formula 1 again.

15. Alexander Rossi, 5.17 ppr

Rossi’s debut at Manor pitted him against one of the weakest drivers in Formula 1, but with the added challenge of jumping into the seat mid-season. Failing to beat Stevens would have been career-ending, and Rossi performed very respectably in the circumstances. Rossi’s results now at least merit consideration for a full-season drive. Whether he will get one at Manor in 2016 remains to be seen, as driver selection at this end of the grid usually depends greatly on factors other than talent. For the seat he will be in competition with Manor’s development driver Jordan King, son of Manor’s 2015 chairman and former Sainsbury’s CEO Justin King.

14. Marcus Ericsson, 5.28 ppr

Ericsson was the lucky recipient of a race seat in the great Sauber driver lottery of 2015. Between them, Ericsson and Nasr are thought to bring around $40-50 million to the ailing Sauber team. Relative to most drivers with this level of financial backing, Nasr is very talented and Ericsson is about average. In his rookie year, Ericsson was beaten 5-2 in races and 10-4 in qualifying by the more experienced Kobayashi. This year, Ericsson was convincingly beaten by his rookie teammate. Both drivers will continue at Sauber into 2016, but Ericsson must do more to justify a longer term drive in Formula 1.

13. Felipe Massa, 5.75 ppr

Last year, I demonstrated that Massa and Bottas were much more closely matched than the points tally suggested. I also concluded that “a second year of Massa vs. Bottas will give us a better idea of where these two relatively stand.”  With an additional season of data, it now seems clear that Bottas’s advantage over Massa is persistent but small. The net advantage to Bottas is 322-255 in points, 17-16 in races, and 22-14 in qualifying. Overall, the model rated Massa’s season as solid, if unspectacular, and puts him very close to his teammate in the list. With Ferrari moving into touching distance of Mercedes and Red Bull receding, Williams were left in a relatively uncontested third. Between them, the Williams drivers finished in the range 4th to 6th a total of 19 times (10 times for Bottas and 9 for Massa).

12. Kimi Raikkonen, 5.79 ppr

Raikkonen’s 2014 struggles were often attributed to the F14T’s tendency to understeer on turn-in, contrary to Raikkonen’s preference for a pointy front end. After voicing great satisfaction with the F15T’s handling characteristics, Raikkonen delivered a similarly muted performance again. To put the result in perspective, Raikkonen’s points haul in the occasionally win-worthy and usually podium-worthy F15T was smaller than Alonso managed in last year’s disastrously uncompetitive F14T.

To be fair to Raikkonen, facing Alonso and Vettel is no easy task. The model thinks that most drivers in the midfield currently (including names suggested as Raikkonen replacements) would have been beaten by a similar margin across 2014-2015. However, one naturally expects more from a past champion who commands a champion’s salary. As the graph below of Raikkonen’s career performances illustrates, he has hit a nadir similar to his woeful 2008 season. The difference is that Raikkonen is now seemingly on a steady decline and he is 36 years old.

Since Ferrari have renewed Raikkonen’s contract for 2016, in 12 months we should be able to definitively conclude whether this is a terminal age-related decline in Raikkonen’s abilities, or just another temporary dip in performance.

11. Sergio Perez, 5.87 ppr

Perez’s career has followed an unusual path. After a semi-successful junior career, including 2nd in his second season of GP2, he joined the Ferrari Driver Academy, who found him a Sauber seat for 2011. In his first year he was outmatched by Kobayashi, but turned the tables in 2012. In May 2012 a bombshell dropped as Hamilton signed with Mercedes. McLaren were left needing a replacement driver, but were out of step with key contracts in the driver market. Their most advanced junior, Kevin Magnussen, was also not yet ready for Formula 1, following a quick but erratic rookie season in Formula Renault 3.5. McLaren was realistically left with a choice between Perez or the Force India drivers, di Resta and Hulkenberg. At that moment, Perez’s stock was high with three podiums to his name, while a clear number one at Force India was yet to emerge — di Resta dominated the early season before Hulkenberg turned things around.

Perez thus found himself with a golden opportunity to impress at a top team alongside a world champion teammate. Instead, we got a convincing demonstration that he was not quite at Button’s level, being beaten 11-5 in races. Unimpressed by Perez’s results and work ethic, and with Magnussen ready for promotion, McLaren dropped Perez like a hot potato. Having severed ties with Ferrari to join McLaren, this seemed to spell the end of Perez’s top-team ambitions. A year at Force India being beaten by Hulkenberg certainly did Perez no favors.

In 2015, Perez looked a new man, genuinely outperforming his teammate. Both Force India drivers’ careers are at a critical point now. If either driver can really impress in 2016, they have a small chance of getting back on Ferrari’s radar as they look for a driver to replace Raikkonen. If not, they could find it hard to sustain a seat into 2017. Driver potential is a valuable commodity in the Formula 1 driver market, and there’s usually little of it left after six years in the sport.

10. Romain Grosjean, 5.95 ppr

Last year, Grosjean asserted his superiority over Maldonado. This year, he straight up destroyed him. While Grosjean can be justifiably pleased by such a one-sided score-line in the teammate battle, successes for Lotus were limited. The season high-point was a thoroughly deserved podium at Belgium with the bailiffs at the door. Despite an obvious improvement in pace with the Mercedes powerunit, Lotus continually hemorrhaged points through a combination of poorly timed mechanical issues and crashes. For his part, Grosjean made errors at Japan and Russia, and was the victim of first lap collisions at Britain and Italy.

2016 will see Grosjean move to Haas. There he will have the opportunity to assert himself against Gutierrez, a teammate with a relatively weak record in Formula 1. It is an interesting move, with an eye to scoring Raikkonen’s Ferrari seat in 2017 — the only likely opening at a top team for an unaffiliated driver in the next two years. For that seat he will face competition from several drivers. If Ferrari want a new hot-shoe, Verstappen and Nasr are both appealing choices. If they want the best driver available, it is probably Ricciardo. However, Ferrari have traditionally run a clear number two driver, who is just quick enough to support the lead driver without taking many points from them. That, along with a presumed objection from Vettel, likely rules out Ricciardo. By the model’s estimation, Grosjean and Bottas are both excellent candidates for the number two role.

The Haas team’s performance in 2016 will test the viability of their unusual team structure. Their relatively small group of ~200 personnel is spread across three countries (USA, UK, Italy). Their projected team budget for next year is only $100 million, putting them just ahead of Manor, as shown in the graph below of estimated 2015 team budgets and the 2016 projected Haas budget.

Around 70 of the Haas team personnel were hired temporarily from Ferrari until November 1, 2015. Other teams viewed this is as Ferrari exploiting a loop-hole, which the FIA has now taken measures to close. Taking advantage of the fact that the Haas team is not subject to the usual restrictions on wind-tunnels and CFD simulations until they officially enter the sport next year, Ferrari’s loaned personnel were effectively free to sandbox aerodynamic concepts for 2016 and beyond, taking whatever knowledge they glean back to Ferrari. Interestingly, the Haas team chose not to convert their state-of-the-art “Windshear” rolling-road wind-tunnel from 100% scale to the FIA-mandated 60% scale. This suggests they may also be using their temporary freedom to run 100% Formula 1 chassis models, and perhaps even correlating the data with the 60% models at Maranello.

While Haas undoubtedly face a significant challenge next year, and are very unlikely to escape the lower midfield, their alliance with Ferrari should at least make them more competitive than the five last teams to join the sport: Caterham, Lotus (the other one), Virgin, HRT, and Marussia. Additionally, Lotus/Renault are facing a tough year with a step back to Renault engines, and continued financial woes. While Grosjean (or Vergne) would have been a sensational lead driver for a French works team, his gamble on Haas seems well motivated.

9. Valtteri Bottas, 6.14 ppr

Last year, Bottas appeared at 4th in the model’s season driver rankings. As I noted at the time, this was partly an artifact of Massa’s poor luck in 2014, as a more detailed analysis of the data suggested only a small performance difference. 2015 gave a fairer assessment of the pair. By the model’s reckoning the Williams was comfortably the third best car, and both Bottas and Massa performed about to par. The table below shows the predicted number of points each team would have scored in 2015 (excluding Toro Rosso) if they each had equally good drivers.

Based on this list, Ferrari (428 points) and McLaren (27 points) punched well above their weight, helped by Vettel and Button/Alonso, respectively. Force India (136 points), Lotus (78 points), and Sauber (36 points) are rated the greatest underperformers, due primarily to the points opportunities squandered by Hulkenberg, Maldonado, and Ericsson.

Alongside Grosjean, Bottas seems to have achieved a level slightly below the leading group of drivers, and slightly ahead of the Force India duo. Bottas is certainly good enough to win races if he lands in the right car, but the model doesn’t currently see him as a driver capable of dominating the next era. That will perhaps be left to Red Bull’s drivers and some of this year’s rookies.

8. Daniel Ricciardo, 6.72 ppr

Besides McLaren, Red Bull endured probably the most frustrating season of any team. All year they were plagued by reliability and power issues, with Renault’s late upgrade setting them even further behind. In the early season, they struggled to even outpace their low-budget sister team, suggesting their problems were not limited to the power unit.

In this difficult position, both Red Bull drivers were pushing hard to make up positions, and they often overstepped the mark. It was not the confident, inch-perfect racecraft we came to expect from Ricciardo in 2014. Overambitious moves contributed to or wholly caused collisions at Malaysia, Monaco, Hungary, Britain, and Japan. Kvyat meanwhile bumped into Perez in Austria, was penalized for exceeding track limits in Hungary, and had big shunts in Japan and the USA.

Across the season, Kvyat outscored Ricciardo by 3 points. In such a close match-up, the raw points tally can be misleading. Ricciardo’s mechanical failures at Silverstone (likely 5th), Belgium (likely 4th), and Russia (likely 4th) cost around 40 net points to Kvyat. Add a low points finish in Brazil (8th) to give 44 net points. Kvyat lost likely points places in Australia (7th) and China (10th), plus a likely 4th in Singapore due to a poorly timed Virtual Safety Car and bungled pit-stop, and a likely 7th in Abu Dhabi due to ERS problems. Paying back all these results to both drivers would result in an overall points tally of 130-105 to Ricciardo.

Alternatively, we can assess who should have finished ahead in each race without misfortunes in qualifying or races.

Overall, this analysis confirms that it was an extremely close match-up at Red Bull, but it definitely slightly favored Ricciardo. Although the model gives a slight edge to Kvyat, a more detailed assessment suggests their rankings for 2015 should really be swapped.

7. Nico Rosberg, 6.76 ppr

After close challenges to Hamilton in 2013 and 2014, Rosberg began the year as something of a spent force. After Singapore, he trailed Hamilton 12-1 in qualifying, having been a close match for Hamilton in qualifying previously, with an 18-18 record across 2013-2014.

In the final stages of the season, Rosberg seemed to drastically improve relative to Hamilton, beating him 5-0 in qualifying after Singapore. It is unclear whether this improvement was due to extrinsic factors, normal fluctuations in form, or the diminished pressure as the championship was effectively settled.

Overall, Hamilton now leads Rosberg 954-810 in points, 27-20 in races, and 30-25 in qualifying. If Mercedes remain the team to beat in 2016, Hamilton should remain the title favorite. But a continued good run of form for Rosberg, or a few unlucky DNFs for Hamilton, could easily tip the balance.

6. Daniil Kvyat, 6.81 ppr

Going into 2015, very few pundits were game to predict that Kvyat could outscore Ricciardo. While Kvyat’s rookie season against Vergne was extremely impressive, Ricciardo had just dismantled quadruple-champion Vettel, staking a strong claim to being among the Alonso-Hamilton-Vettel club. Looking back to my own comments last year, I was intentionally vague but gave Kvyat a fighting chance, “If Kvyat can achieve the kind of second-year progression many seem to be expecting of Magnussen, then Ricciardo had better watch out.”

The analysis in Ricciardo’s entry above confirms the subjective impression that Kvyat got a little lucky in outscoring his teammate across the season. Nevertheless, this was one of the closest teammate match-ups on the grid, with the advantage shifting between the drivers from one track to the next. On raw pace, Ricciardo’s advantage was clear, with an 11-7 result in qualifying and better pace in some of the races where he scuppered his own chances through mistakes.

Right now, Red Bull have probably the third strongest driver line-up on the grid, behind McLaren and Mercedes. They will not overlook a chance to strengthen that line-up and bring fresh junior talents into the queue behind. Ricciardo may not be losing sleep over Kvyat outscoring him, but 2016 will be a serious test of both drivers’ mettle, and I would not be surprised to see sparks flying between them, especially given the implied threat to their seats from Verstappen.

5. Felipe Nasr, 6.98 ppr

Nasr had a phenomenal rookie year that was largely overlooked in favor of Verstappen’s swashbuckling brilliance. As noted in the Toro Rosso entry, the model rates the season performances of Nasr and Verstappen very similarly, under the assumption that the Toro Rosso was slightly better than the Sauber. The cars were very different beasts. Toro Rosso were chronically down on power, but their James Key designed chassis was often credited as one of the year’s best, although its pointy front-end clearly took great feel to master. The Sauber had a strong power-unit but seemed to suffer from an underdeveloped chassis.

Nasr’s junior career was extremely impressive before GP2, including the British F3 championship ahead of Kevin Magnussen in only his third full year of single-seaters. From there he jumped straight to GP2, showing initial promise with 10th in the championship as a relatively inexperienced rookie. However, the next two years of GP2 showed slower progress than was anticipated, dampening enthusiasm for his Formula 1 chances. This year, Nasr has silenced the critics.

In assessing Nasr’s year it is essential to consider the fact that his teammate was not a rookie. Since the 2008-2009 testing restrictions, rookies have had consistently poor results against non-rookies (even relatively weak ones), as the table below shows.

Colors indicate superior (green) or inferior (red) results compared to the rookie’s teammate. *Kobayashi started two races in 2009.

While Ericsson is far from the strongest driver on the grid, Nasr had an impressive edge over him in races for most of the year. In stark contrast to the 2014 Sauber drivers, Nasr stepped up his game whenever points were on offer, scoring 75% of the team’s points. When compared to the results of rookies against non-rookies since the testing restrictions of 2008 and 2009, Nasr’s results are very strong.

Importantly, Nasr brings an estimated $24 million of support from Banco do Brasil. Between this and his strong driving performance, Nasr seems a very attractive option for teams in the upper midfield, and possibly Williams, where he was reserve driver in 2014. If he can continue his progress into 2016, a move up the grid in 2017 seems plausible.

4. Jenson Button, 7.39 ppr

Button emerges from 2015 with an impressive statistic to his name. He is the only driver to have outscored either Hamilton or Alonso across a season as teammates*. That fact probably provides only small relief for a season spent in a hopelessly uncompetitive car. Moreover, it’s not a statistic with great value, since there were very few weekends in 2015 where both McLaren drivers were free from technical issues. To really gain some insight into this driver match-up, it is imperative to do a detailed race-by-race analysis, as I did for the Red Bull drivers above.

Overall, Alonso was clearly the better performer, but the analysis also shows that Button gave Alonso a far stronger fight than any past teammates, besides Hamilton and perhaps Trulli. While the car may be a disaster, McLaren currently have what the model considers the strongest driver pairing on the grid. This creates a quandary for driver selection, because on the sidelines McLaren have Vandoorne, a driver with an incredible junior record who now holds the GP2 season records for most wins, most feature race wins, most points, and biggest percentage points lead.

With decreased earnings from prize money and sponsorship on the horizon, and a huge hill to climb in performance, McLaren have found it difficult to justify buying a seat for either Vandoorne or Magnussen in 2016. Consequently, they have lost Magnussen (another bright talent), and it would not be a huge surprise to see a frustrated Vandoorne leave if another big team asks for his services. Ferrari have potential options for placing juniors on the grid (e.g., via Haas or Sauber) and are currently rebuilding their junior academy in the wake of losing future-star Bianchi and failing to get the expected results from Marciello in GP2. The very talented Charles Leclerc is rumored to be one of their targets. Might they try to poach Vandoorne?

McLaren may be spoiled for choice with drivers at present, but their current champions do not have unlimited shelf-life. Button will be 36 in January. Alonso is 18 months younger, but commands a crucial role as the driver who could potentially drag the car a few extra places up the grid once it is remotely competitive. At what point will their skills begin to deteriorate? In other sports, it is typical for athletes to have a peak plateau around the ages of 25-35, but this varies with the demands of the sport. Reaction times decrease by age 30, whereas physical endurance performance stays near peak levels to about age 35. Playing some sports can take a physical or cognitive toll, such as contact sports with frequent injuries or concussions. In other sports, experience gained over time compensates for decreased physical performance. Below is an analysis, redrawn from “Bridging Different Eras in Sports” by Berry et al., and normalizing the data to place them on the same axes.

Ice hockey and baseball show a very similar age curve, with an increase in performance up to about 27, followed by a progressive slow decline. In golf, the peak occurs much later and has a more stable plateau. Formula 1 is an interesting case to compare, since it requires a combination of fast reflexes, decision-making, deep experience, and physical endurance.

I did an analysis of the effects of age on model-estimated driver performance using all the drivers from 1950-2015 in my database. I fit a linear model to average changes in performance with age for each individual driver. The average age effects are plotted below.

On the y-axis is the predicted difference in performance compared to the peak performance (at age 35). To help interpret the numerical values, a difference in performance of 0.3 is about the same as the average career difference between Hamilton and Bottas. The peak period appears to be relatively late on average, running from about 28 to 37, with rapid decline thereafter. These are average effects and individuals may progress very differently, but this at least gives some indication of when age-related decline begins for Formula 1 drivers.

*Alonso finished 2004 with more points than Trulli, but Trulli was ahead of Alonso 46-45 before he was sacked by Renault.

3. Lewis Hamilton, 7.91 ppr

2015 was a year of immense career success for Hamilton. He overtook Senna and Vettel to take 3rd in all-time grand prix wins, and he joined an exclusive group of ten drivers with at least three titles. Overall, the model considers 2015 Hamilton’s best year yet, narrowly ahead of 2014 and 2007, concurring with Hamilton’s own assessment that it was the best year of his career. Although the car was completely dominant — by end of season the W06 has climbed the list of most dominant cars in history from 7th to 3rd — Hamilton saw off the challenge from teammate Rosberg with impressive ease.

On his day, Hamilton looked untouchable and able to outpace Rosberg with consummate ease. But there is room for improvement too. Hamilton’s error-strewn race at Hungary, following a clumsy mistake at Silverstone, evoked memories of 2011. Spain and Monaco created a blot on the early season record; the former down to Rosberg’s better performance, the latter down to a combined bad call by team and driver that sabotaged Hamilton’s clearly superior pace. From Japan, the teammate battle also switched decisively in Rosberg’s favor.

The key question now is whether Mercedes can maintain their sizeable advantage into 2016. Two changes to the technical regulations next year have the potential to change the team hierarchy. The first is the mandated change to engine exhaust systems, which will effectively allow teams to make some changes without spending tokens. The second is allowing teams to choose between three elected tyre compounds at each race weekend, which could allow other teams to more easily find their ideal operating range. At a higher level, two more factors will influence the Mercedes vs. Ferrari battle. The first will be how cleverly Ferrari have used the Haas loop-hole to their advantage. The second will be how soon the teams shift their focus to development for the extensive changes in 2017.

 2. Fernando Alonso, 7.93 ppr

In 2015, McLaren-Honda set tragicomic records for grid penalties, while Alonso’s former team made significant inroads into Mercedes’ advantage. At this stage in his career, however, the only record Alonso claims to care about is another world title. Whether Ferrari or McLaren-Honda can overhaul Mercedes in Alonso’s few remaining years remains quite unclear, especially with a likely shuffle of the hierarchy in 2017 when cars will undergo their greatest dimension changes since 2009.

During 2015, Alonso put in many extraordinary drives, including the brilliant 5th place at Hungary. But by Alonso’s reckoning, it was not his strongest year.

“I did some good laps here and there. I remember Japan in Q2. The Austin race, I think it was good in difficult conditions, and I felt confident with the car and I could push. But apart from these two moments I don’t think that I had the best season, so definitely I need to improve for next year. When you are running at the front and you have more motivation that helps, of course. But I’m on standby, let’s say, in economy mode, to have full energy next year.”

The model agrees that 2015 saw a downturn in Alonso’s form, and this was especially the case in the latter stages of the season. Indeed, it is the first year since 2007 that the model has awarded the championship to a driver other than Alonso. The graph below shows the model’s year-by-year ratings of the current generation’s top three drivers.

For almost anyone else on the grid, Alonso’s 2015 would be a stunning performance, but his standards of performance are extremely lofty. A natural question to ask is whether this is a temporary dip in form or the first sign of age-related decline. A quadratic fit to Alonso’s yearly performance data suggests he is now beyond his absolute peak, but given a competitive car he is still more than capable of delivering the goods for a few more years. The question then is how quickly can McLaren-Honda deliver?

The rate of progress for different teams across 2014-2015 can be estimated by studying qualifying times at tracks that had dry sessions and the same tyre compounds in both years (Bahrain, Spain, Monaco, Canada, Austria, Hungary, Singapore, Japan, and Abu Dhabi). Below is the median percentage time difference from the 2015 Mercedes for each 2014 and 2015 team on their best timed lap in each session.

Marussia’s times are omitted from the graphic. They were slower at all comparison tracks in 2015 than in 2014, falling from 5.2% behind to 6.6% behind. This reflected their inability to develop a new aerodynamic package for the new nose regulations, as well as the loss of their strong lead driver, Jules Bianchi.

Ferrari achieved a gain of 1.10%, and a relative gain on Mercedes of 0.55%. Sauber also made a large step forwards, despite limited aerodynamic upgrades over the winter, confirming Ferrari’s huge power gains. On the other hand, Red Bull gained only 0.29% from 2014, with a relative loss to the Mercedes team of 0.36%. Red Bull were actually slower at 4 of the 9 comparison tracks in 2015 than in 2014. Lotus made a gain of 0.94% thanks in part to changing suppliers from Renault to Mercedes. Their relative gain compared to others would suggest the change of engine was worth around 0.5%.

Despite McLaren making major steps on their chassis and aerodynamics from the disastrous 2014 car, they lost a net 0.70% on their 2014 car’s pace. The obvious culprit is the Honda engine. More detailed analyses, based on GPS data, suggest around 70% of McLaren’s 2.7% time deficit to Mercedes stems from the power unit, which would account for 1.9% total. Even a gain of the same magnitude as Ferrari’s this year would optimistically put them around 5th in the constructors’ championship next year.

1. Sebastian Vettel, 8.30 ppr

2015 was in many ways a make or break year for Vettel’s reputation. While it is difficult to simply overlook Vettel’s four titles and the obvious quality of his driving in those years, the validity of his place among the all-time greats is a question that keeps coming up. It arises for three reasons.

• Vettel’s junior career was fairly pedestrian by the standards of drivers who go on to win world drivers’ championships. This included losing the F3 Euro Series title to Paul di Resta in 2006, when both drivers were in their second full season in the series.
• Vettel’s four titles all came in clearly the best car, and at least half of Formula 1 success is about being in the right place at the right time. By the model’s assessment, only 11 drivers in Formula 1 history have spent at least four seasons in the best car.
• Webber was relatively old when he faced Vettel, and Vettel was soundly beaten by Ricciardo last year. To quote a recent statement by Hamilton, “I have a lot of respect for him, but it’s difficult to assess how good he really is. He’s never been in a team with someone like Fernando Alonso, but always with people like Mark Webber, who was not on his level, and Kimi Raikkonen, who is no longer at the peak of his performance.”

Solid counter-arguments exist for each of these points. Junior success is not always strongly predictive of Formula 1 success. Vettel has certainly had incredible cars at his disposal and relatively weak teammates, but the margin by which he beat Webber and Raikkonen is still informative. As for 2014, it’s still unclear just how good Ricciardo is, and large dips in driver form are unusual but not unheard of. The model currently rates Vettel’s 2014 dip the 13th largest mid-career dip in history, based on the change in a driver’s performance relative to the years immediately before and after.

Vettel’s 2015 performance should put most questions surrounding his quality to rest. With two notable exceptions (Bahrain and Mexico), Vettel looked mightily impressive as he occasionally took the fight to Mercedes and usually stayed comfortably out of reach of the Williams cars that hounded his embattled teammate. This season marks the first time that the model has awarded Vettel a drivers’ title, putting him in a group of just 22 drivers awarded titles by the model to date.

Since we will likely never see Alonso and Vettel as teammates, it is informative to see them against the same teammate in two consecutive years. Moreover, it’s an interesting hypothetical for many fans to wonder how 2015 would have looked, had Alonso continued his contract with Ferrari. On most statistics, Vettel and Alonso beat Raikkonen by staggering margins, but based on their expected points scoring rates the model still rates Alonso’s 2014 adjusted performance (8.79 ppr) significantly better than Vettel’s 2015 adjusted performance (8.30).

As another way of quantifying their relative performances, I studied the timing data. In qualifying, I calculated the percentage difference between the Ferrari drivers’ best times, only including times set in sessions where both drivers ran and had no mechanical issues. For each race, I calculated the percentage difference between the Ferrari drivers across all laps that both drivers completed, excluding lap 1 and safety car periods. I then calculated the median percentage differences between the Ferrari teammates across each season and the interquartile ranges. Note that it is important to use percentiles here instead of mean and standard deviation, because the sample is not normally distributed and has unrepresentative outliers that dominate the mean (e.g., the races at Monaco 2014 and USA 2015, where Raikkonen crashed and rejoined, losing lots of time).

The results of this analysis are graphed below.

Median (thick line) and interquartile range (box) of the time advantages held by Vettel (blue) and Alonso (red) over Raikkonen (dashed line at zero) during their years as teammates.

From this we can conclude that both Vettel and Alonso had enormous advantages over Raikkonen, being quicker than him well over 75% of the time. Their advantages tended to be slightly larger in qualifying than in races, which is consistent with Raikkonen’s reputation as a better racer than qualifier. The analysis also shows a large relative time advantage to Alonso over Vettel, if we take Raikkonen’s performance level to be relatively consistent across these two consecutive years.

On this basis we should conclude that the Alonso of 2014 could have gotten slightly closer to Mercedes, but there was never any beating such a dominant team. Moreover, the model’s assessment of 2015 driver performances implies nobody would have done a better job this season in Vettel’s position. He was the star of the season. If Ferrari can continue their rate of relative progress into 2016 (and it’s a big if), they should be capable of taking more regular race wins. At that point, you could not count Vettel out from a fifth championship.

The 2016 form guide

Driver performances can be hard to predict in advance, due to fluctuations from one year to the next. Vettel’s return to form from 2014 to 2015 is one obvious example. Nonetheless, we can look at the average performance of each driver over the past 3 years to get a general sense of the driver hierarchy at the present moment. I performed that calculation for all drivers who were active for at least 2 of the last 3 seasons, with the hierarchy graphed below.

As you would expect, the ordering is similar to the 2014 and 2015 driver rankings. A notable absentee from the 2016 grid is Vergne, who is currently ranked 5th in the hierarchy, just behind Ricciardo and just ahead of Rosberg, Button, and Kvyat. Raikkonen is beginning to plummet into the pack, after being ranked the 5th best performer of 2012 and the 4th best performer of 2013. With another year like 2015, he will likely sink behind Grosjean and Bottas, both hungry for his Ferrari seat. Perez has now just inched ahead of Hulkenberg. Meanwhile, Gutierrez has the lowest rank of the current group, not helped by one of his two assessment years being a rookie year. He will get another chance to redeem himself in 2016 against a good benchmark teammate in Grosjean.

While the focus here is on drivers, the sport’s entertainment level in 2016 will be largely dependent on the team hierarchy. Can McLaren-Honda deliver Alonso and Button a competitive car? Can Ferrari turn the championship into a two-horse race? Can Red Bull solve their engine woes and get on terms with Williams? And can Toro Rosso become a serious force once their excellent chassis is allied with Ferrari power? These are questions we will revisit next year.

Tags: 2015 season, Fernando Alonso, Formula 1, Lewis Hamilton, mathematical model, Mercedes, rating method, Red Bull, Sebastian Vettel, Williams