The statistic - sDRS (Simple Driver Rating System)
What it is - sDRS is loosely based on the Elo Rating system, but has some major differences to account for the multiple competitors in motorsport racing. The system tells us how strong a particular driver is during a season and how close the competition is to each other. It can be used to predict which driver will be the strongest heading into a race and also accounts for past year's performance.
The calculation: sDRS for Formula One is calculated fairly simply. All new drivers are initialized with 1500 points (which also happens to the long term average of all driver's ratings). Points are awarded based on finishing position after each race as follows:
The formula for calculating the new sDRS score is:
NEW_sDRS = OLD_sDRS + POSITION_POINTS
The nature of this scoring means that the system, like Elo, is zero sum. This means for every point one driver loses, another one gains. This serves to keep the system balanced and easy to judge throughout seasons, no matter the scoring system in use. sDRS is also a much better predictor for mid to back field drivers who don't get much love from Formula One's top ten scoring system. It can show the best of the rest just as easily as it can show the best overall.
sDRS rewards consistency over one off races, too. A driver may get a big jump in points for a good result one week, but if he doesn't back it up after that, his rating will fall back down. This avoids a major pitfall of Formula One scoring in that one driver could get a lucky result and sit in a good points position all year long. A driver will always gain points for being better than average (tenth or eleventh place), and always lose points for a result worse than average.
Speaking of average, the majority of the field will have an sDRS score of 1400-1700. 1500 is considered an average driver, less than 1400 is a bad driver (worthy candidate for replacement next season in most cases), and greater than 1700 is a great driver. A score above 1850 is the sign of a world class driver.
At the end of each season, driver's scores are reverted to the mean by one half. This accounts for the unpredictability of car performance each year. With car redevelopment each season, it's highly likely that a team who was successful one season could be well of the pace this year. Reverting to the mean corrects for this and makes the ratings more accurate at the start of a season. Think McLaren from 2014 to 2015.
Telling the relative strength of the field for a given season is a strength of sDRS. The point spread between the lowest and highest rated drivers tells us how closely matched the field is. A spread of less than 200 points is incredibly close and usually doesn't occur in Formula One. 300-400 points is what an average Formula One field looks like and more than 500 usually means there is a powerhouse team dominating most of the season.
Drawbacks - Like any rating system, sDRS is not perfect and has its quirks. For starters, sDRS doesn't know which track the series is going to next or how many races a driver has competed in. So for example, a driver's great historical record at a track isn't accounted for in the calculations before that race. All sDRS knows is that a driver has been good, bad, or average this year. The second problem is that new drivers will usually have unpredictable ratings for the first five races of the season. It takes sDRS a few races to center in on their real score and see if their early results were a fluke or a sign of things to come.
It's a simple rating system, meaning it uses very limited variables to give a score. This is beneficial in a lot of cases as the more saturated a predictive statistic becomes the more likely it is to fall victim to too much noise - which screws up the data. sDRS does a great job of showing the strength of drivers relative to each other, but shouldn't be used as an all or nothing guide by any means.
Ratings will be updated after each week of the Formula One season.
by: Drew B., Founder