Cycling Previews Accuracy

So I’ve spent 2019 writing a fair amount of race previews for professional men’s and women’s cycling. I set out with the aim of doing a preview for each Women’s WorldTour race, which I almost did and a preview for each Men’s Spring Classic, which I did do. Other than that, some previews like Rund Um Köln were just because I like the race!

At the bottom of each preview, I always give a top 3 prediction. Just -a little something to give people who don’t want to read the whole thing an idea of who is likely to do well. There’s times where I feel a little creative, otherwise, I’d just end up picking Peter Sagan or Annemiek Van Vleuten to win every time.

Now the 2019 cycling season is over, I’ve crunched the numbers to see how spot-on (or not) I was!

Men’s Cycling Previews Accuracy

I’ve put together a Google Sheet which anyone can visit, but here’s a short rundown of the facts:

I wrote 9 men’s race previews in 2019 and 16 previews in 2018.

  • I got 3 winners correct – 33%, up from 19% in 2018
  • I didn’t get a single 2nd place correct – 0%, down from 13% in 2018
  • I got 1 3rd place correct – 11%, up from 6% in 2018
  • 3 riders I predicted to finish 2nd, actually won the race. So 66% of the time, the winner came from my top 2. Big improvement from 2018’s 25%
  • 12 of the 27 riders I predicted finished in the top 5 – 44%, 29% in 2018
  • 16 of the 27 riders finished in the top 10 – 59%, 58% in 2018
  • I nearly predicted Amstel Gold perfectly – I got the winner, my 2nd place finished 4th and my 3rd place finished 2nd. In 2018, my predicted winner finished 66th, predicted 2nd finished 4th and predicted 3rd finished 11th.

Women’s Cycling Previews Accuracy

I wrote 17 women’s cycling previews in 2019 and just 2 previews in 2018.

  • I got 6 winners correct – 35%, 0% in 2018.
  • I got 1 2nd place correct – 6%, 17% in 2018.
  • I got 1 3rd place correct – 6%, 17% in 2018.
  • 2 riders I predicted to finished 2nd, actually won the race and another 2 I’d predicted to be 3rd won races too. In 2018, I picked 1 eventual winner in 3rd place, but 4 of the 6 places finished on the podium for 67%.

Summary

Procyclinguk Predictions
Closeness Score compares the predicted place to the actual finishing place and totals the difference up. DNFs are given a default position of 100.

The race average shows the average score for each race – 0 is a perfect score, so -37 is the best score in the table. The Combined Total Score adds the men & women together per year.

On average my 2019 predictions were 32 positions per race more accurate than 2018.

So what I’ve learned is that pretty much 33% of the time, I picked the winner in my 2019 previews. Compared to 2018, that’s an improvement of about 10% or so – improvement!

My margin for error is pretty good too in men’s racing. If I picked someone in my top 2, then they won the race 66% of the time. A massive gain on 2018 of 40%. It’s also best to pretty much ignore the rider I’ve predicted for 3rd place.

With the women, in 59% of 2019 races, the winner of the race has come from the 3 riders I’ve picked. A similar sort of number to the men’s predictions and a big jump up from 2018’s 17%.

So all in all, I’m pretty happy with the big improvement shown in the 2019 race preview results. But the question I now have is are those numbers actually any good in the grand scheme of things?

So there’s an obvious follow-up blog post to this one where I pick some other cycling race preview blogs and websites and see how they compare to my previews for the same races.

Any volunteers or site suggestions to challenge are welcome!

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.