Where Market Data Meets Formula 1

As an avid Formula 1 viewer and a fan of the British drivers (in the absence of a Scottish one at this time), I listen intently to the commentary and information associated with the race as it unfolds.

During this year’s Spanish Grand Prix, I was struck not only by the way they described the amount of data collected during a race but also by how that data is processed in real-time and relayed back to the drivers. As I was watching the race, I realized this shares a striking resemblance to the way we at TraditionData captures, processes and distributes data as a financial markets data provider.

The data-centric nature of Formula 1

During a two-hour race, over 1.1 million data points per second are transmitted from the cars to the pit wall, easily breaking into multiple terabyte levels before completion. Across all participants, upward of 150TB of data is sent between the remote race circuit and the F1 Media and Technology Center in Biggin Hill, England, which is then processed by the team’s engineers, pit wall, and race control of the teams back at the factory.

Each car has approximately 300 sensors transmitting the data in real time, which electronically record the performance of the engine, the status of suspensions, gearbox data, fuel status, all temperature readings including tire temperature, g-forces, and actuation of controls by the driver. The transmitter is placed in the car’s sidepod with a cable running to an antenna on the nose. 

Each car also has an onboard storage system that buffers the most recent data, so if the transmission fails, the car keeps retrying until it can be completed. This means that, for example, no data is lost when the car enters the Monaco tunnel: as soon as the communication is lost, the car keeps collecting data for storage on its on-board memory. As it then exits the tunnel, or any blind spot, all data collected during this ‘dark period’ is immediately sent to storage. The data is then decoded and converted into a signal that can be understood by a PC, going through a data server which displays the telemetry channels for the engineers. This is the suite that displays all the wavy lines on the screen, and this information is then relayed back to the drivers in real-time as they are on the track.

One additional factor that is ever present during the race is the driver’s input and feedback on how the car feels, the traction of the tires on the racetrack and anything that may or may not be visible via data e.g., the extent of damage to a part of the car that will have a direct effect on the metrics being fed back in the data. In other words, things that may need a human element to allow full assessment of a given situation.  

TraditionData’s race in the markets

We can draw comparisons between the world of formula 1 and the financial market data industry using Tradition’s Trad-X Interest Rate swap platform.

During our “race”, when trading hours are open, we are receiving constant pricing updates from clients in the market that connect via trading API’s, with price changes within the platform (our circuit) processed within milliseconds. The fine margins matter, just like in Formula 1. Individual price points received are processed through interpolation routines which, over the course of a day, can result in billions of price points being produced and collected. This accumulates into gigabytes of data covering a multitude of products across the Interest Rate Swap universe (outrights, spread-over, switches, butterflies, and more).

The data processed comes from multiple participants (cars) on the platform across multiple trading points in the market (sensors), and these trading points include currency (USD, EUR, GBP), bids, offers, trades, notional amounts and multiple levels of depth across each market segment. The platform receives all data points, drops them into the calculation engines, processes them, returns them back to the platform and redistributes them in real-time back to the participants.

To ensure market stability, our systems cache data locally (car on-board data storage systems) before transmitting to the TraditionData backbone, so that in the event of an internal IT failure no market data is lost. This means that when systems are restored, data is restored.

In addition to data being distributed and consumable via an API, the data is also presented on a trading screen (track side telemetry displays) for clients or traders who wish to physically see the price movements on screen.

Finally, in the vast majority of OTC derivative markets, voice brokers play a key role alongside the data. Yes price inputs, price movements in reaction to those inputs, depth of the Order Book market etc. are all visible and measurable via the data. However, things like true size (not shown in a standard order in an order book) of counterparties interests, sentiment as to a trader’s opinion on external factors, or how an upcoming event may affect the market are not always directly understood purely from data. Broker (driver) input by watching, listening and paying attention to the voice traders at the end of the phone also plays a key part in the operational functionality, and therefore success of the different participants (cars).

Although the content may differ on the buy-side, the speed, the diversity of data points, the circle of events from the car/broker/trader, through processing environments, back to the “business-end” of the respective environments in real-time where accuracy and latency are critical and defining factors in the success of the operation the workflow is strikingly similar.

As an interdealer broker, we may work in a very niche part of the world’s markets, but the application of what we do and how we do it when considered in raw data processing activities has parallels that aren’t obvious but, in a way, are equally cool. 

After all this comparison with Formula 1, I’m now off to buy myself a cool helmet to wear at my desk… because why not? They do.