Mercedes’ latest W17 aerodynamic package appeared to be a step backwards after Sepang, but the direct comparison carried out in Singapore suggests a more nuanced picture. The newer specification seems to offer greater potential than the Montréal version, particularly in traction and race pace, although it is significantly harder to bring into the correct operating window.
After qualifying in Sepang, Mercedes expected to have gained roughly three tenths with the new package. Instead, Toto Wolff said the team appeared to be “three, four or five tenths behind”, with Kimi Antonelli fourth and George Russell eighth, marking the first time Mercedes had fallen outside the top three by such a large margin. The package initially looked like a regression, which is why Singapore became such an important test.
Singapore gave Mercedes its first direct comparison
At Marina Bay, Mercedes ran two different specifications. Kimi Antonelli used the Sepang package together with a new front wing, while George Russell returned to the Montréal specification used up to Baku.
Antonelli finished third, while Russell recovered to ninth after starting 19th. It was the only weekend in which the two packages were directly compared on the same circuit, providing the clearest indication of which responds better and which might offer the stronger foundation for 2027.
Melbourne: the original Mercedes W17 had no real rival
The launch-specification W17 featured bodywork with more downwash, a higher air intake and more heavily curved wings. In practical terms, the concept directed more airflow towards the floor and rear of the car, where much of the downforce is generated, while the more curved wing elements added further load.
Up to Miami, only an exhaust detail and the brake drum lip were changed. Australia, China, Japan and Miami therefore provide a clean four-race baseline for the original car.
That version of the W17 was comfortably ahead. Over a qualifying lap, Mercedes held an average advantage of six tenths over McLaren, Ferrari and Red Bull, while the race-pace advantage was seven tenths.
Against the midfield median, meaning the teams outside the top four, the margin rose to 1.7 seconds in qualifying and 2.1 seconds in race trim.
During races, Mercedes was around 3 km/h faster than its main rivals on the straights and 2.7 km/h faster in high-speed corners, while tyre life remained comparable.
Where the original W17 was strongest
The original concept strongly favoured medium- and high-speed corners as well as straight-line performance.
In qualifying, Mercedes’ apex speed in medium-speed corners was 4.1 km/h higher than its rivals.
The trade-off came under braking. Across braking zones, Mercedes travelled an average of 28 metres further than its rivals before reaching the apex: 16 metres more before slow corners, 43 before medium-speed turns and 50 before fast corners.
The car therefore spent longer in the braking phase, although that measurement alone does not necessarily identify a weakness.
Slow corners were the only area where the W17 did not clearly outperform its rivals. At the apex in qualifying, it was 0.8 km/h slower, while 120 metres before the apex it was 1.9 km/h faster. Full throttle arrived 6.6 metres later than for the opposition.
Race data looked better, with a 4 km/h advantage 120 metres before the apex and full throttle reached 6.7 metres earlier, but the four tracks produced variations of more than 13 metres, making those differences statistically weak.
Traction was already the key question
The distance between the apex and the point of full throttle is a useful indicator of traction. It shows how early a driver can commit fully to the accelerator after the tightest part of a corner and therefore how effectively the car can put its power down.
At Melbourne, Mercedes was not clearly better than its rivals in this area. That would later become especially relevant because traction is one of the characteristics directly affected by the Sepang package.
Almost everything else across those opening four races represented a broad and consistent advantage.
Montréal package won immediately but rivals began closing in
The first major Mercedes update arrived in Montréal and contained eight listed changes.
The team lowered the outer front-wing elements and connected them to the footplate, redesigned the floor with a reshaped box section and a slotted corner, enlarged the brake ducts for circuit-specific requirements and redistributed the rear fins.
The stated objective was to create more robust airflow structures, increase local downforce and improve the flow entering the diffuser.
In simple terms, Mercedes wanted more cornering load without destabilising the air travelling towards the rear of the car. A robust airflow structure remains more stable when the car pitches or changes direction, preventing downforce from falling away through the middle of a corner.
The diffuser is the final section of the floor, where the airflow expands and generates a significant proportion of the car’s downforce. Improving its inlet conditions can therefore increase its efficiency.
Antonelli won immediately with the Montréal package
The immediate result was victory for Kimi Antonelli in Canada.
From Montréal to Baku there were 11 race weekends. Mercedes was faster than its main rivals in qualifying at 10 of them, with Hungary the only exception, and held an average three-tenth advantage.
In race trim Mercedes was ahead across all 11 weekends, by an average of four tenths.
Compared with Melbourne, however, the advantage had shrunk from six to three tenths in qualifying and from seven to four tenths in the race. Against the midfield, the race advantage remained close to two seconds.
Mercedes stayed ahead, but its rivals developed faster
The main problem was that the opposition improved at a faster rate.
In qualifying relative to the midfield, McLaren gained four tenths, Ferrari four tenths and Red Bull almost six, while Mercedes improved by only around 1.5 tenths.
The Montréal package kept the W17 in front but did not deliver the leap Mercedes needed.
After Canada, only circuit-specific adaptations followed from Monaco through Madrid. Baku was the final race with that configuration, where George Russell won ahead of Max Verstappen by 0.196 seconds.
The strongest characteristics remained tyre degradation, which was 0.014 seconds per lap of tyre age better than the opposition, and straight-line speed, where Mercedes retained an advantage of roughly 2 km/h.
Medium-speed circuits exposed the weakness
The weakness emerged on medium-speed circuits such as Barcelona, Hungary, the Netherlands and Madrid.
There, Mercedes’ qualifying advantage fell to only one tenth, compared with four tenths on high-speed circuits such as Austria, Great Britain, Belgium and Monza, and six tenths at mixed-layout circuits such as Canada and Baku.
The cornering data reinforced that trend.
In high-speed race corners, the apex-speed advantage fell from 2.7 km/h in the Melbourne phase to 0.7 km/h.
In medium-speed qualifying corners, the advantage dropped from 4.1 to 1.0 km/h. In race trim, medium-speed performance held up better, at 2.4 km/h compared with 2.1 previously.
The W17 therefore appeared to have lost part of its advantage in medium- and high-speed corners. Those were precisely the areas Mercedes aimed to address at Sepang.
What changed with the Mercedes Sepang package
The Sepang package contained seven listed changes and focused primarily on the rear of the car.
The floor received a revised box section with vertical elements, a leading edge with downwash features, a slotted corner and reshaped roof, sidewall and diffuser-wing surfaces.
The rear suspension fairings, cake-tin winglets and rear bodywork were also changed, with reduced volume and a steeper ramp.
The floor is the single biggest downforce-generating component because the air passing beneath the car creates a low-pressure zone that pulls it towards the track.
Slots and vertical elements help control that airflow, while suspension fairings, cake-tin winglets around the rear brake assembly and tighter rear bodywork are used to clean up the air heading towards the diffuser and rear wing.
The target was more load in medium- and high-speed corners
The declared objective was to create more local downforce and improve airflow towards the rear of the car, increasing load in specific areas without paying too large a drag penalty.
The main target was medium- and high-speed cornering, where Mercedes’ advantage had diminished during the Montréal phase.
The expected gain was around three tenths. At Sepang, however, the initial result appeared to be the exact opposite.
Singapore added a new Mercedes front wing
In Singapore, Kimi Antonelli also tested a new front wing with redistributed chord across three elements and only two elements active in straight-line mode.
The purpose was to better control the wake generated by the front tyres and make performance more stable across a wider operating range.
In practice, the front wing determines how airflow opens around the front wheels, and a cleaner wake makes the car more predictable.
George Russell remained on the Montréal-Baku configuration, meaning the Singapore comparison measured both the package itself and Antonelli’s new wing together.
The two Mercedes packages are not completely separate concepts
The two specifications are not alternatives in the strictest sense.
The Sepang package is built on top of the Montréal concept and retains its front wing philosophy, while the development focus moves towards the floor and rear of the car.
The expected benefits are improved traction and medium-speed performance. The two possible costs are drag and a narrower operating window.
Mercedes admits the Sepang package has a narrower window
The operating window is the one drawback Mercedes has openly acknowledged.
After Singapore qualifying, Antonelli said that “the window in which this upgrade seems to work is narrower than with the previous specification” and that the car “reacted differently depending on how I drove it”.
The following day, Toto Wolff said the package “is definitely a little more difficult to bring into the right set-up window, but there is clearly something good in it”.
The set-up window refers to the range of settings, including ride height, suspension stiffness and balance, in which the aerodynamic package delivers peak performance. With a wide window, the car can still perform strongly even if the set-up is not perfect. With a narrow one, even small deviations can lead to lost downforce and reduced driver confidence.
For the driver, that translates into unpredictability. For the engineers, it means more testing is required to find the correct operating point.
Sepang raised concerns over drag
Drag was the other possible downside and initially looked like a real problem at Sepang.
Wolff said that “it looks like we have a lot of drag, because this is the first time we are losing so much time on the straights”.
He also highlighted understeer at Turn 2, meaning the front axle did not rotate sufficiently and pushed the car wider than intended. Drag is the aerodynamic force that slows the car through the air, and its cost is most visible in straight-line speed. Teams chasing more downforce will normally accept some reduction in efficiency, but only if the extra cornering performance outweighs the straight-line loss.
What the Sepang and Singapore numbers actually show
With the Montréal package, Mercedes was ahead of its rivals almost everywhere, with Hungary qualifying the main exception. Race pace was effectively equal only in Hungary and Madrid.
At Sepang and Singapore, that pattern changed.
In qualifying, Mercedes was 0.19 seconds behind its rivals at Sepang and 0.06 seconds behind in Singapore.
In Singapore race pace, Antonelli was 0.28 seconds behind, while Russell was 0.6 seconds off.
Part of that is linked to circuit type.
Sepang and Singapore were the only two slow circuits of the season apart from Monaco, and on October 3 Antonelli said that “all year, every time we have gone to circuits like this, with very open asphalt, we have always struggled”.
That means comparisons with Montréal across different circuit layouts can be misleading, making Singapore the most reliable direct test.
Sepang straight-line deficit may have been misleading
The speed figures reveal less than they initially appear to.
In Sepang qualifying, Mercedes was on average 4.3 km/h slower than its rivals in top speed: 6.3 km/h down with Antonelli and 2.3 km/h down with Russell, compared with its usual 1.2 km/h advantage.
However, the normal variation is around 4 km/h.
Similar deficits had already appeared in Australia with the launch-spec car, at minus 5.8 km/h, and at Silverstone with the Montréal version, at minus 4.8 km/h.
In Singapore, the two specifications could not be separated by any of the three maximum-speed measurements.
Tyre degradation and energy use do not separate the packages
Tyre degradation and energy deployment also provide no clear distinction.
With Montréal, Mercedes degraded 0.014 seconds per lap less than its rivals.
In Singapore, Antonelli degraded 0.028 seconds per lap less than the opposition and Russell 0.063 seconds less. The difference between the two drivers remains within the uncertainty because Russell had only 13 laps in clean air. Clipping, where electrical energy runs out towards the end of the straight and the car stops accelerating as strongly, was identical between the two specifications.
Full-throttle acceleration was also effectively equal.
Singapore initially looked bad for the new package
At first glance, Singapore appeared to condemn the newer specification. Russell was fastest in opening practice and second in Sprint Qualifying, holding a 0.57-second advantage over Antonelli, who was seventh and described his car as “very difficult, very unpredictable”.
On Saturday, however, after set-up changes, Antonelli qualified fourth and beat Russell by 0.19 seconds, compared with his usual 0.12-second advantage.
The race comparison favoured Antonelli’s specification
The Grand Prix provides a less clean comparison.
Russell started 19th because of a power-unit component penalty, drove a rebuilt car after crashing out of the Sprint on the opening lap while leading, and completed only 13 laps in clean air.
Antonelli’s race pace was 0.28 seconds behind the main rivals, while Russell was 0.6 seconds off. That creates a three-tenth difference between the two Mercedes drivers compared with their normal one-tenth gap.
The newer Mercedes aerodynamic package therefore appears to have been worth around two tenths per lap in race trim and less than one tenth in qualifying.
Traction produced the clearest signal
The clearest difference appeared in slow-corner traction.
During the race, Antonelli reached full throttle eight metres later than the rivals, while Russell did so 16 metres later. With the Montréal package, Mercedes had previously averaged reaching full throttle three metres earlier than the opposition.
At 120 metres from the apex, Antonelli was 5 km/h faster than Russell after adjusting for their usual performance difference.
In qualifying, Antonelli reached full throttle five metres earlier than the rivals, while Russell was 12 metres later. Both struggled in race conditions, but the Sepang specification lost less without completely solving the weakness. In medium-speed corners, Antonelli reached full throttle between 14 and 18 metres earlier than Russell once their usual gap was taken into account, although the difference remained below the threshold required for a fully conclusive result.
The cost-benefit equation favours Sepang, but only slightly
The expected cost-benefit relationship does appear in the data, although on a relatively small scale.
The expected cost, higher drag, is not visible because both specifications perform equally on the straights.
The expected benefit, more downforce, does not appear in high-speed corners or in apex speed across any corner-speed category, where the differences remain inside normal variation. The gain appears instead on corner exit, clearly in slow corners and with less certainty in medium-speed ones. In practical terms, the new specification accelerates away from slow corners better, but it does not carry more speed through the middle of the turn.
Mercedes reaches a similar conclusion
The team’s interpretation broadly matches the data.
In its October 11 statement, Andrew Shovlin wrote that “by the end of the weekend, Kimi’s package was clearly the faster of the two”.
Antonelli also spoke of “really good progress with our upgrade package”.
On October 13, Wolff said he did not believe the upgrade had yet delivered the performance Mercedes originally expected, but still described it as a step in the right direction. The numbers support that direction without yet proving that a decisive advantage has been established.
Strengths and weaknesses of the Montréal package
The Montréal package has several obvious strengths.
Qualifying and race pace were almost always ahead of the opposition, by averages of three and four tenths respectively.
Tyre degradation was better and the car was around 2 km/h faster on the straights.
The weaknesses were the shrinking margins on medium-speed circuits and in medium- and high-speed corners, alongside an advantage over the midfield that was no longer increasing.
At present, the operating window is the only criterion where Montréal remains clearly superior to Sepang.
Strengths and weaknesses of the Sepang package
The newer package shows its advantages more clearly in detailed metrics than in outright lap times.
In slow corners, the traction deficit is smaller.
In medium-speed corners, Antonelli reaches full throttle 14 to 18 metres earlier than Russell once their normal difference is accounted for.
The main weakness is the narrower operating window, which Mercedes itself has acknowledged.
Traction improved in Singapore but still remained weaker than the opposition.
The drag concern that emerged in Malaysia, meanwhile, remains within the range of normal variation.
Eight of 11 criteria are effectively equal
Across 11 criteria, eight produce no meaningful separation between the two specifications.
Race pace and slow-corner traction favour Sepang.
The operating window favours Montréal.
No measured criterion shows a clearly distinguishable deterioration with the new package, meaning the only confirmed cost comes from the team’s own comments about how difficult it is to set up.
Which Mercedes aero package should be carried into 2027?
The 2027 regulations will change the balance between combustion and electrical power.
Under the FIA-approved revision, internal-combustion power will rise from 400 to 420 kW with five per cent more fuel, while MGU-K output will fall from 350 to 300 kW.
Energy recovery will rise from 350 to 375 kW, producing a 58:42 split.
Combined peak output therefore falls from 750 to 720 kW, while the declared objective is to create qualifying conditions with more sustained full-throttle running.
In simple terms, the combustion engine will do more work and the electrical system less, while 30 kW of maximum combined power disappears.
With less total propulsion available, aerodynamic drag and efficiency become even more important.
In that respect, the Sepang package currently shows no measurable cost compared with Montréal, while also losing less performance in traction.
Sepang currently looks like the more sustainable choice
For that reason, the more sustainable direction appears to be the aerodynamic package introduced at Sepang, with one major reservation: its narrower operating window.
Mercedes will need to study that characteristic extensively as it prepares the next car, because having theoretical aerodynamic potential matters far less if the team cannot access it consistently.
Austin is the test Mercedes still needs
The missing test is Austin, scheduled for October 23-25.
On October 13, Wolff said Mercedes would “probably” use the newer specification there.
Unlike Singapore, where the team had only a single free practice session because of the Sprint format, Austin will provide three practice sessions and therefore much more time to work through the set-up.
Without a second specification running in direct comparison, the question will change.
The key issue will no longer be whether Sepang is faster than Montréal on the same circuit, but whether its traction advantage remains present and whether Mercedes can return to the front of the competitive order as consistently as it did with the older package.
Leave a Reply