This season KZM has decided to double our Speed Rating output and include California high school racing in on the fun. Many of you from California may be asking "What's a Speed Rating?" "Why have Speed Ratings?" or even "What's KZM?" So to help clarify we are reposting this article we posted last season that answers all these questions and some.
Put your thinking caps on and read on because we’re about to explain How KZM Speed Ratings are calculated.
“Why even have speed ratings?”
Speed ratings are a great way to compare skiers from around the state. In most high school sports, to determine the top ranked teams or athletes it requires a room full of distinguished former coaches and athletes who see things out on the field the common folk would never see. With their analysis, they debate the performances and produce a list of the top ranked athletes or teams and challenge others to dethrone their discretionary elite.
Thankfully, in racing sports like Nordic skiing, cross-country running, and track there is no need for deliberation. You are as good as your time, and the fastest time wins.
However, in track, there is loads of standardization. From the surface of the track to the distances raced, there is not much variability in one track meet to the next. But what about cross-country? For the most part, every race is 5K, why do we need a speed rating?
Course difficulty and weather conditions have a big effect on racer times. “Racer A” could run a 16:00 5K one weekend and a 17:00 5K the next, on a different course. If “Racer B” runs 16:30 on a 3rd course, who is faster? To compare the two runners, we use a speed rating, which is the runner’s time, compared to all the other runners who have recorded times on that course and we can determine a value that can easily be compared.
Tullyrunners.com Cross-Country speed ratings.
For those familiar with cross-country speed ratings, here is a brief explanation on how they are calculated. You can learn more specifics on their website.
Basically, a speed rating is (1560-adjusted time (in seconds)/3). 1560 are the amount of seconds in 26 minutes. So essentially, the rating is how many seconds faster you were than 26 minutes… divided by 3.
The “adjusted time” in the formula is how Tully accounts for varying conditions. Tully assumes (with years of data and times on specific courses) that the distribution of runners in high school is approximately normal. So with loads of data, Tully overlays this particular race with data from years in the past and takes the difference. [See Fig. 1]
In short; “Racer A” runs 17:00 at an invitational and wins the race. According to Fig 1, the mass of this invitational was about 30 seconds slower than the average of years past. So Racer A gets an adjusted time of 16:30 (a.k.a. 990 seconds). We crunch the numbers: (1560-990)/3 and we get a speed rating of 190. (Dang that’s pretty good, if your familiar with Tully speed ratings).
Tully goes way more in depth with standard conversions specific to each course (for example, Tully knows Spa Park in Section 2 is always fast, so he immediately has to add 70 seconds to everyone’s times). If you want to learn more about Tully Speed Ratings go check it out, its super interesting if you’re a data freak like me!
So why not use this system for KZM speed ratings? It seems pretty easy and quite accurate in ranking runners, why not skiers?
The answer is because this requires years of data we unfortunately don’t have and there is waaaayyy more variability in Nordic racing than running. Consider all of the same variables in running along with varying race lengths (from 1K to 10K and all the not-quite-accurate 3.876Ks in between) and varying race conditions (10k of slush is much slower than 10k on a 20 degree day with hard packed snow). KZM had to come up with a way to compare racers while taking all the variables into account.
How KZM ratings are calculated.
The only way KZM could accurately create a speed rating is to compare athletes to each other. I will now give you the calculation in a sentence then use pictures to better explain it.
KZM speed ratings are racer times compared to the average time of the top 4 racers to participate in the race.
What does that mean? Let me explain with 4 racers:
4 racers toe the line and complete a 5K race (miraculously they finish in 15 second intervals). Their times are recorded and we take the results and calculate the mean season-average and the mean time from the race. We multiply the mean season-average by the mean time and get what is known as the “race-base.” We then divide the racers time by the race-base and multiply by 100 to get the KZM Speed Rating! [see Fig. 2]
Now, obviously, the chances of them finishing in regular intervals is almost impossible, so let’s see what happens when they finish in irregular intervals (more like a real race). [Fig.3]
As you can see, it does not matter the distance, the conditions or the difficulty of the course. If skiers who usually average a particular speed rating can complete the course in a given time, speed rating are calculated based on how you do compared to them.
Now let’s say there’s a fifth racer and Racer D beats Racer C. We must still calculate the Race Base off of the top racers who participated. [Fig. 4]
“So why do some people have a rating over 100?”
Because their time was that much faster than the Race Base time. [Fig. 5]
“What about the first race of the season? Where does KZM get the season-averages to determine speed ratings?”
KZM took last year’s state championship results and gave last season’s distance winner a Speed Rating of 100. All the other Speed Ratings were a proportion of the winner’s time. Then we went back and found as many results as we could from last season and calculated the season ratings that were then used in the Pre-season predictions. In the first race, we used the pre-season predictions as the season averages.
"How does KZM Calculate California's speed ratings for the first race?"
KZM went back and tried to find all the most recent races where NY skiers raced CA skiers and gave them speed ratings based off of NY skiers. This included 2018 JN's and 2019 SR Nats. Then from there, we calculated the first race off of those skiers.
One flaw that we like to call the “Scotty-Effect” is what happens when you have one skier who is head and shoulders above the rest. In the pre-season predictions, Lake Placid’s Scotty Schultz earned a 100 speed rating from the year before. Well ahead of second place, Queensbury’s Brian Beyerbach with 94. With such a large gap, we expect Scotty to dominate the field. Well, when Scotty lost a ski in the first race just before the finish, Brian was able to sneak ahead and claim the victory. Now because Brian (94) beat Scotty (100), Scotty’s rating plummeted and Brian successfully inflated everyone’s speed rating who was in the race (not by much, but a little.) [Fig. 6]
The best way to ensure accuracy is for skiers from different sections to race each other often.
Another down side to this rating system is at the beginning of each season, after taking out the graduates from the year before, who ever has the highest rating gets set as 100. Everyone else's ratings are adjusted in relation to the adjustment of the top rating. This method makes it impossible to compare results season to season, but it is a must, otherwise, the top skiers max speed ratings will slowly decrease over time.
Anyway, despite these few flaw, KZM’s speed ratings are about as accurate as we can calculate in a sport with so many variables. If you have any comments, concerns or questions, please feel free to reach out via email, social media or through the website. KZM is here to promote your sport.
Few More things
Season averages are calculated by taking the average of the season’s races minus the best and worst races. We feel this can help combat the “Scotty-effect” and if you break a pole, for example, you can throw a race out.
Classic and skate averages are calculated based on the average of all skate or classic races (little math involved).
Grade levels are based on info from coaches, parents or the athlete themselves. If we don’t know a grade level we leave it blank. (Although we’d like to find them all by the end of the year so we can predict next season).
It's all for fun. Speed Ratings are designed to compare and start a conversation about ski racing. There are no teams or qualification at stake with these numbers. Please enjoy the product for what it is and not put to much weight on it.
We hope this info clears things up and sheds some light on what the speed rating next to your name means.
Live The Nordic Life.