When I recently posted my race simulator here, several people expressed interest in seeing pre-race simulator predictions. So, here is the first simulator prediction thread!

THE TRACK

Sochi is a particularly difficult test for the simulator, as there are no previous data whatsoever. This makes it difficult to anticipate the rate of tyre degradation and the time taken to make a pit-stop.

Pit-stops: I was able to estimate a net pit-stop time loss of at least 24 seconds from pit-stops in practice, but then the FIA thought it would be funny to complicate matters by reducing the pit-lane speed limit from 80 km/h to 60 km/h. Given the length of the pit-lane, I think this will increase the time loss by ~7 seconds, for a net time loss of ~31 seconds. This is a very large time loss. By comparison, the net time loss is ~14 seconds in Britain, ~17 seconds in Canada, ~18 seconds in Germany and Belgium, ~21 seconds in Hungary, ~22 seconds in China and Monaco, ~23 seconds in Australia, and ~26 seconds in Italy.

Tyre degradation: For this race, Pirelli are fielding soft and medium tyres, as they did in Australia, Bahrain, China, Belgium, and Hungary. I have been able to estimate some basic characteristics of these compounds (time difference and relative durability) from these previous races. Interestingly, the practice sessions in Sochi suggested a very low rate of degradation on this smooth asphalt surface. Looking at the timing data from each driver’s longest run on prime tyres in FP2, there is indeed no noticeable degradation in lap times within the first 10 laps – graph here.

RACE STRATEGIES

With a shorter or faster pit-lane, we would probably see a two-stop race. However, with the very harsh time penalty associated with pitting and the seemingly low rate of degradation, most drivers should be aiming for a one-stop strategy. The estimated optimal strategy for a driver running their own race in clean air is about 24 laps on softs then 29 laps on mediums. However, racing considerations may cause the first pit-stop to shift earlier. The soft tyres should be slower than fresh mediums after about 20 laps, so I expect some drivers to stop around that mark to prevent an undercut. Drivers who pit too early will risk needing to stop again before the finish, but a well-timed safety car (a real possibility at any street circuit) could help reduce degradation or reduce the time penalty associated with a second stop.

THE RACE LEADERS

For the first time this season, we look to have a genuine three-way race, with Hamilton as favorite. The Williams team have impressed in qualifying before, but this time they have long-run pace to seriously challenge Mercedes. The model predicts a slim time advantage for Hamilton in clean air (2 tenths over Bottas and 3 tenths over Rosberg), but unless he can break free in the early laps, DRS could keep the leading bunch together for much of the race. Moreover, Bottas has the straight-line speed to cause trouble – he was a full 4km/h faster than anyone through the speed-trap in qualifying and 5-6km/h faster than the Mercedes cars.

Predicted probability of winning

Hamilton: 75%

Bottas: 13%

Rosberg: 12%

BATTLE FOR THE NEXT POSITIONS

Button is favorite to finish 4th, with a 39% chance. He also has an outside chance of a podium if trouble strikes any of the leaders (31% chance). However, he will likely face strong competition for 4th place from Kyvat (22% chance), Magnussen (15% chance), Ricciardo (12% chance), and potentially Vergne (4% chance), Alonso (2% chance), Raikkonen (2% chance), and Massa (1% chance).

Alonso’s pace is hard to judge, because he didn't complete any long runs in FP2. Based on Raikkonen's FP2 pace and both drivers’ pace in qualifying, it seems they will require a small miracle to finish higher than 5th.

Vettel was closely matched with Ricciardo in FP2, but was mysteriously slow in qualifying. With several strong challenges from behind and relatively poor straight-line speed, he has only a 27% predicted chance of finishing in the points.

DRIVERS TO WATCH

Kevin Magnussen: Magnussen will start 11th due to a gearbox change, while teammate Button starts 4th. However, qualifying and practice times show that Magnussen should be a serious force in this race if he can stay out of trouble. His predicted pace in clean air on fresh rubber is actually 1 tenth faster than Button, Massa, and Kyvat, 3 tenths faster than Vergne, and 3-6 tenths faster than the Red Bulls and Ferraris. Magnussen was also the second fastest driver through the speed-trap in qualifying, behind only Bottas. The model considers 4th-7th most probable.

Felipe Massa: Massa will start 18th in a car that is probably the second quickest in clean air and the fastest in a straight line. If he can overcome any reliability issues, expect him to cut through the pack. He is not quite able to match Bottas for pace here – the model predicts he is about half a second slower in clean air – so a podium is probably out of the question, but points are likely. The model considers 7th-12th most probable.

Nico Hulkenberg: Like Magnussen, Hulkenberg has a 5-place grid drop and so he will start 17th. However, his long-run pace is matching the Ferraris and Red Bulls, and he has been the kindest of anyone on his tyres this season, running an average of 26.2 laps on prime stints (9% above average) and an average of 22.0 laps on option stints (12% above average). The model predicts that he should be fighting for 9th to 11th, most likely just one place behind Perez who is starting 12th.

POINTS FOR BACKMARKERS?

At this stage in the season, the only thing more certain than Mercedes winning the WCC is Toro Rosso finishing 7th. However, the lower places, especially 9th to 11th, could hinge on a single upset. So what are the chances of scoring points for the minnows at this race?

Lotus: 2% chance

Marussia: 1% chance

Sauber: 7% chance

Caterham: 2% chance

Marussia have almost no chance with only a single car and poor pace for Chilton. Lotus are looking no quicker than Caterham on long runs in Sochi and have been hampered by reliability problems all season. There is little chance of scoring for Caterham or Lotus, although notably Ericsson looks much improved, with most probable positions of 13th to 15th ahead of his teammate.

Sauber are the team with real hopes of an upset to overcome Bianchi's points at Monaco. Had Gutierrez not broken down in Singapore, he had a good chance of fighting for at least 10th. Sutil and Gutierrez have identical predicted pace in clean air on fresh rubber here, so the model predicts a close fight between the Sauber drivers – hopefully with no collisions! In Sutil's favor is the fact that he has gotten better life from his tyres than Gutierrez this season, running an average of 8% further on prime stints and 26% further on option stints than his teammate. In Gutierrez's favor is the fact that Sutil is the most crash-prone driver on the entire grid, with 1 race-ending crash every 5.4 starts! Maldonado has 1 race-ending crash every 10.3 starts for anyone wondering...

For those who know how to code in Matlab and want to tinker with the code, you are welcome to download it here. For public use, please credit http://f1metrics.wordpress.com/