Fail to prepare, prepare to fail. Max Verstappen has won eight Formula 1 grand prix so far in 2024 and each of those victories came after eight billion simulations were run by Red Bull to best predict scenarios ahead of time.
Utilising the partnership with its title sponsor, Red Bull works with Oracle Cloud technology to crunch the numbers to help with strategy, tyre choices and pitstop calls - starting two weeks in advance of every race.
Around four billion ‘Monte Carlo’ simulations are run ahead of a race weekend – nothing to do with the Monaco Grand Prix but instead a principle used by computer simulators, introduced during the Second World War when physicists working on the probability of atomic bomb chain reactions named their simulations after the principality and its number of casinos.
“What I should say is we run four billion simulations before we arrive at track,” explained Jack Harington, partnerships global lead at Red Bull, with the Las Vegas Grand Prix next to get the simulation treatment.
“Then we take the data from FP1 and FP2 and introduce those into the simulations and we will run those four billion simulations again. So in fact, it is about eight billion simulations overall.
“It's small tweaks each time, so you're not changing everything in one go. With Las Vegas, it's interesting because we didn't have the historical data, so having the ability to run so many simulations means we will be better informed, even if we haven't got experience in the circuit. So it really plays well into the strength of running on the cloud.”
As well as being a title sponsor, Oracle is now embedded into the fabric of Red Bull as senior director of product management Taylor Newill explained.
“Red Bull Racing utilises a number of Oracle Cloud technologies across their entire business,” he said.
“Number one, though, are high performance computing technologies, we make large computers that can do millions and billions of physics simulations, to simulate things like combustion engines, to help designers build next generation engines.
“Some of the technologies that Red Bull uses are what we call our ARM core processors. So we're using high performance computing for engine simulation. We use ARM computing for really fast, very high number of simulations for things like race strategy.
“From the Oracle Cloud side, we give them a lot of compute every weekend so that they can they spin up these machines. They do the simulations. They spin the machines down, so it allows them anywhere in the world to access this hardware.”
Such predictive technologies are now commonplace in F1 but with Oracle Cloud, Red Bull has been able to increase its simulations by 25%, with positive financial implications too.
“Whereas before we had to invest in physical stuff to be able to run these, it's also much more cost efficient for us as well. So we're looking at things like the cost cap, being able to run these simulations as and when we need them, access to huge amounts of process," added Harington.
“It's a real performance benefit to us. Ultimately, it's no lie when I say that every race-winning strategy we've made since 2021, has been made on OCI (Oracle Cloud Infrastructure). It is a true partnership; it isn’t a sponsorship.”
Red Bull team principal Christian Horner echoes those sentiments, praising Oracle’s input to Verstappen’s all-conquering title campaign in 2023.
“Oracle Cloud played a critical role in helping our team deliver the most dominant performance in F1 history, and I can’t wait to see what we accomplish next,” he said.
Regardless of person-power on hand, no team could come close to processing the sheer amount of data produced by eight billion simulations.
“Sadly we can’t look at eight billion individual results for obvious reasons,” Red Bull senior race strategy engineer, Stephen Knowles, told Autosport.
“We tend to aggregate the results from each simulation, grouped together by common characteristics – the number of pitstops, tyre compounds used, tyre degradation, car pace, pit laps, safety cars etc - and then use the number of championship points scored to compare the merits of each strategic decision.
“This approach allows us to produce some very clear visualisations of points sensitivity to things we can control, such as pit laps. Clearly we can’t know the future, so we tend to focus on the things that we can measure to tell us which of the billions of scenarios we simulate are most relevant for the racing situation we find ourselves in on Sunday.
“Quite often, as we are able to run so many simulations, they highlight some really interesting and unconventional ideas for strategy. Monaco is a great example, where the difficulty in overtaking makes some pretty extreme stuff possible.
“In these situations, we will look at some of those individual simulated races in detail so we can understand whether there is an opportunity to do something really different.”
While never able to truly predict the outcome on track, the system clearly works. Verstappen is now on the cusp of claiming his fourth successive drivers' championship and could seal the crown in Las Vegas - a scenario that does not require Oracle and Red Bull to run the numbers, with the odds much shorter than eight billion to one.
