This is a very interesting post from John Henry, explaining how iRacing’s ’strength of field’ system works. I had assumed it just tried to match up drivers with similar ability but it seems that it is designed to give you opportunities match your skills in a variety of situations, where you may be the fastest, or slowest, driver in the field.
“There is a hell of a lot of talent here. One the best things about this service is that as your iRating matures you will find yourself racing people with similar skills. Not everyone in every race will have similar skills, though, even with mature iRatings. Let me explain.
One of the cool aspects of the algorithm used to determine who races whom is that it doesn’t just group people linearly.
Last Dec 24 I was in a race with an iRating of 1293. The top driver in that race came in with an iRating of 1841 and won by 2.5 seconds over a driver rated 1327. The second driver actually gained more iRating integers because the top driver was expected to win. The winner gained 65 and the second place driver gained 67. But the point is a driver rated 1841 drove in a 1293-rated race. The worst rated driver was 705.
What does this mean?
Well, the winner Jim Kepford was placed into a field where he should have won and he did. But not by much. He was certainly the big dog in that field. His iRating rose from 1841 to 1906. In his next race a week later he was place in a field where he should be an average driver (as most drivers find in a particular race). The field was rated 1856 before the race and after because iRating integers are simply exchanged. He was rated 1906 and predicted to finish 7th (he was in car #7) in a field of 16 cars. He finished 4th – 13 seconds behind the leader and that drove his iRating to 1957.
4 days later his 1957 iRating put him into a race rated 2346. Even with his new iRating up to 1957 he was predicted to finish 14th in this field of 17 cars. He had to face the big dog in that race, Adam Haufrect, who was rated 3754.
Jim managed to finish 7th and his iRating rose to 1992. The lowest rated driver in the race was rated 1732. Although he was predicted to finish 17th, he managed to finish 10th. That was Erin Nagy.
Erin was placed in a field too tough for him while Adam was in a field that was too easy – one might say. But this is done by design. If a race rated 2346 had all drivers rated between 2300 and 2400, we don’t think drivers will learn as much as they would with a driver rated 3700, another rated 3200 etc. And we think a 1700 rated driver will learn something in a field like this rather than running 100% of the time with 1700 rated drivers.
In Erin’s next race he came in rated 1740 and the field was rated 1112. In the previous race he was favored to finish 17th but finished 10th 49 seconds off the pace. And in this race he was favored to win. He did so by 18 seconds.
So there is an example where in one race a driver is predicted to finish 17th and four days later in his next race he is predicted to finish 1st. We think this is better than being predicted to finish mid-field every race – although generally you will be placed in a race closer to midfield. We want each driver to once in a while be one of the best drivers in a field, once in a while one of the worst. We don’t want races necessarily always full of drivers evenly matched. That makes passing pretty difficult and is unrealistic as far as what goes on in the real world.
BTW, Erin scored 60 points finishing 10th in that strong field and 81 points winning in the weaker field. An average driver in iRacing is going to score approximately 50 points per race no matter what the strength of field is.
So this is probably controversial, but it is how we do it. I hope I have been somewhat clear. “