23 July 2017

Froome Wins 4th and We're Nearly Perfect!

After his performance in the Alps and in yesterday's time trial, there was no doubt that Chris Froome would win his fourth Tour de France.  He now has a three-peat (should I send Pat Riley money for using that term???).  Froome didn't win a stage this year, but was clearly the best cyclist.  Staying near the winners in the mountains and in the time trials, Froome was simply better than everyone else.  His Team Sky mates played a large role in his victory.  It doesn't hurt to be supported by a powerful team!

Dutch cyclist Dylan Groenewegen won today's final stage.  How did our model perform today?  We saved our best for last, as you'll see below.
  • Stage 21:  2h 25' 39" (actual), 2h 25' 50" (prediction), 00' 11" slow (0.13% error) 
As tough as it is to predict the mostly-ceremonial final stage, I'm thrilled to end this year's Tour de France with a near-perfect prediction.  How did our model perform overall?  After summing the stage-winning times, I found we were 1.11% slow.  I'll definitely need to spend time thinking about how much athletes and technology have improved since last year.

I never cease to be amazed by elite athletes.  A total of 167 cyclists finished the Tour de France.  I would be hard-pressed to finish a long flat stage during daylight hours.  As for those grueling mountain stages, forget it.  I need more time in the gym!  My model estimates energy burn, i.e. internal energy burn with an average efficiency of about 20% and not just energy output needed to power the bike.  During the entire race about 115,000 Calories could have been burned.  Published cyclists' data may be below that number, but an estimate has to be made of internal energy efficiency.  Our published papers on Tour de France modeling cite sources that are consistent with our energy estimates.  The point is that a LOT of energy is burned during the three-week race.  At 550 Calories apiece, those 115,000 Calories amount to nearly 210 Big Macs.  That averages to 10 Big Macs per stage!  I don't recommend eating Big Macs before cycling, but it does give you some idea of how much energy those cyclists burn each day.  You may have heard that 3500 Calories matches the energy content in a pound of fat.  That's roughly true, but you may have to burn about twice the Calories to get a pound of fat off because of the complicated way the body converts energy.  Either way you think about it, 115,000 Calories represent one or two bowling balls of fat weight.  No wonder elite cyclists stay in such great shape.  Their job is a wonderful form of exercise!

I once again thank rising high-school senior Ryan Wainer from New York for his work this year.  He acquired all the terrain data, which led to a successful set of predictions.  How successful?  We had one bad prediction with Stage 5 (9.24% error) and five good-a-decade-ago-but-want-to-do-better-today predictions in the error range of 4% - 8%.  But that leaves 15 predictions to better than 4%, 11 of which were better than 2%.  Five of those 11 were better than 1%, including our best prediction today.  A nice way to end!

22 July 2017

Froome Seals the Deal!

Chris Froome did what he needed to do in today's individual time trial.  He came in third, just six seconds behind Poland's Maciej Bodnar.  Froome now leads the overall classification by 54 seconds over Colombia's Rigoberto Urán Urán.  Unlike the first individual time trail in Stage 1, today's result was slower than I thought it could be.  We were a bit more than 5% off, as you'll see below.
  • Stage 20:  28' 15" (actual), 26' 42" (prediction), 01' 33" fast (-5.49% error)
One of the first things to look at after this year's Tour de France will be the two individual time trials.  We were short on power in the first time trial and had too much power today.

Tomorrow's final stage will be mostly ceremonial until the big sprinters go for the stage win once they are in Paris.  Froome will probably be seen with some champaign and four fingers up (three in a row!).  The last stage is always tough to predict because of the ceremonial nature of the stage.  We've done well in the past by backing off on power.  Our final prediction is given below.
  • Stage 21:  2h 25' 50" (prediction)
  Chris Froome showed today why he's the best cyclist in the world.

21 July 2017

Hagen Makes Norway Proud!

Norwegian cyclist Edvald Boasson Hagen won today's flat stage, which was the longest stage of this year's Tour de France.  We had a good prediction, as you'll see below.
  • Stage 19:  5h 06' 09" (actual), 5h 11' 09" (prediction), 05' 00" slow (1.62% error)
That makes 12 of 19 stages that we've hit better than 3%.  Strategies on long flat stages are hard to predict because the peloton dictates so much of the pacing.  Boasson Hagen and 19 other cyclists came in within two minutes of the winning time.  Chris Froome and the rest of the peloton came in 12' 27" after Boasson Hagen.  I would therefore claim that our prediction was spot on!

Tomorrow's individual time trial will be the last chance Romain Bardet (23 s back) and Rigoberto Urán Urán (29 s back) have of catching Chris Froome.  Below is our prediction.
  • Stage 20:  26' 42" (prediction)
We were a bit slow on the first individual time trial, which was the opening stage this year.  I'll be anxious to see if we do a little better tomorrow.

20 July 2017

Barguil Puts Us Under 1%!

Warren Barguil made France proud today with his second mountain stage win in this year's Tour de France.  We predicted the two big mountain stages in the Alps really well.  Look below to see how well we did today.
  • Stage 18:  4h 40' 33" (actual), 4h 41' 48" (prediction), 01' 15" slow (0.45% error)
Chris Froome came in fourth today with a group just 20 s behind Barguil.  He has a 23-s lead over Romain Bardet.  The goal for Team Sky in tomorrow's long flat stage will be to keep Froome with the cyclists just behind him.  The individual time trial on Saturday looks like it could be a lot of fun!  Below is our prediction for tomorrow's Stage 19.
  • Stage 19:  5h 11' 09" (prediction)
If Chris Froome is to secure a three-peat, his team will have to have a great day tomorrow.

19 July 2017

Another Prediction Error Under 1%!

Primož Roglič won today's Stage 17 by 73 seconds.  We hit the stage by less than two minutes, as you'll see below.
  • Stage 17:  5h 07' 41" (actual), 5h 09' 36" (prediction), 01' 55" slow (0.62% error)
If riders thought today's stage was grueling, they'll enjoy tomorrow's stage.  It has a delightful Hors catégorie climb to the finish line.  Below is our prediction.
  • Stage 18:  4h 41' 48" (prediction)
Froome now has a 27-s advantage over the two riders behind him.  Based on how he looked today, it seems he is headed for a three-peat.

18 July 2017

Matthews Makes it Two out of Three!

Michael Matthews won his second stage today.  Racing turned out to be fast on the downhill, which is what I mentioned in yesterday's post.  There is no way to predict team strategies!  Below is a comparison between reality and our prediction.
  • Stage 16:  3h 38' 15" (actual), 3h 45' 02" (prediction), 06' 47" slow (3.11% error)
Not a bad error, but yesterday's result spoils me for more sub-1% predictions.  The next two stages in the Alps will likely decide this year's winner.  Our prediction for tomorrow's stage is given below.
  • Stage 17:  5h 09' 36" (prediction)
Monster climbs in tomorrow's stage, but the finish will be a downhill sprint.  Can anyone sneak across the finish line in under five hours?

17 July 2017

Stage 16 Prediction

Below is our prediction for tomorrow's Stage 16.

  • Stage 16:  3h 45' 02" (prediction)
The stage is classified as flat, but it's quite hilly for the first half.  The middle of the stage is mostly downhill.  Depending on racing strategies coming off a rest day, the above prediction could be slow.  I'll be anxious to see if cyclists push themselves to high speeds on the downhills.

16 July 2017

Nearly Perfect Prediction!

There are some stages I watch come to an end and hope the winning cyclist can give just a tiny bit more effort in the last kilometer.  Check out the comparison between Bauke Mollema's winning time with our prediction for today's Stage 15.
  • Stage 15:  4h 41' 47" (actual), 4h 41' 26" (prediction), 00' 21" fast (-0.12% error)
Come on, Bauke!  Just 21 seconds faster today and we reach perfection.  He won by 19 seconds, so he didn't need to rush the final few hundred meters.  But I'll definitely take today's result.   Missing a nearly five-hour stage by 21 seconds is a lot of fun!  I'm glad I didn't make the mistake I made with yesterday's prediction and alter power output.  Our model did its thing today.

Chris Froome remains in yellow.  Nairo Quintana, who I loved watching battle Froome in the mountains in 2015's Tour de France, slipped to #11 in the overall classification, more than six minutes behind Froome.

Tomorrow is a rest day.  Teams will plot strategies for the next day's flat stage and the two, grueling stages in the French Alps that follow.  I'll post our prediction for Stage 16 tomorrow.

15 July 2017

Froome Back in Yellow!

Chris Froome got 25 s on Fabio Aru in today's Stage 14 and turned a 6-s deficit into a 19-s lead on the Italian cyclist.  Michael Matthews won what I consider to be a slow stage today.  Below is a comparison between the winning time and our prediction.
  • Stage 14: 4h 21' 56" (actual), 4h 12' 56" (prediction), 09' 00" fast (-3.44% error)
We were too slow on hilly Stages 5 and 8, so power was upped slightly for today's stage.  We would have been under 1% without the change!  Oh well, that's what makes this so much fun.  We can't predict team strategies, crashes, and weather.  Below is our prediction for tomorrow's hilly stage.
  • Stage 15:  4h 41' 26" (prediction)
I'm not tweaking the power in our model for tomorrow's stage.  Will I regret that with a rest day to follow???

14 July 2017

Barguil Gives Us a Great Pick!

Warren Barguil delivered a second consecutive mountain stage win for France with his impressive ride in today's Stage 13.  Below shows how well we picked this stage.

  • Stage 13:  2h 36' 29" (actual), 2h 34' 22" (prediction), 02' 07" fast (-1.35% error)
That makes six stages predicted to better than 2% and nine of the 13 stages predicted to better than 3%.  Below is our prediction for tomorrow's hilly Stage 14.
  • Stage 14:  4h 12' 56" (prediction)
Fabio Aru holds a slight 6-s lead over Chris Froome for the yellow jersey.

13 July 2017

Bardet Makes France Proud!

Romain Bardet won a fast Stage 12 in the mountains today.  Check out how fast below.
  • Stage 12:  5h 49' 38" (actual), 6h 04' 33" (prediction), 14' 55" slow (4.27% error)
Our error isn't too large, but I've gotten used to hitting the mountain stages a little closer.  Tomorrow's mountain stage isn't even half as long as today's stage, but it has three category-1 climbs and a speedy downhill finish.  Our prediction is given below.
  •  Stage 13:  2h 34' 22" (prediction)
Aru Fabio picked up time on Chris Froome today and wrested the yellow jersey away from the three-time champion.  Will Froome get it back tomorrow?

12 July 2017

Kittel Gets #5 and We're Under 3%!

Marcel Kittel got his fifth stage win of this year's Tour de France.  The guy is flat-out killing the flat stages!  Below is a comparison between Kittel's winning time and our prediction.

  • Stage 11:  4h 34' 27" (actual), 4h 41' 44" (prediction), 07' 17" slow (2.65% error)
Kittel won't win tomorrow's mountain stage.  Our prediction is given below.
  • Stage 12:  6h 04' 33" (prediction)
A rider may come in under six hours, but the stage will tax everyone.  They may need to conserve a little for Stage 13.

11 July 2017

Kittel Gets 4th Stage Win! We Are Under 2%!

Marcel Kittel is dominating the flat stages in this year's Tour de France.  He won his fourth flat stage of this year's race.  I thought someone might come in under four hours.  Kittel was just a minute over that time.  Below is a comparison between Kittel's time and our prediction.

  • Stage 10:  4h 01' 00" (actual), 4h 05' 31" (prediction), 04' 31" slow (1.87% error)
I love seeing an error under 2%!  Tomorrow's flat stage is longer than today's stage.  Our prediction is given below.
  • Stage 11:  4h 41' 44" (prediction)
Can Kittel pick up a fifth stage win?  It will be fun seeing if he can!

10 July 2017

Great Stage 9 Pick and Stage 10 Prediction

I was traveling yesterday, so I missed watching Stage 9 of the Tour de France.  That grueling mountain stage is one I'll have to catch on replay.  Below is a comparison between the actual winning time and our prediction.

  • Stage 9:  5h 07' 22" (actual), 5h 11' 15" (prediction), 03' 53" slow (1.26% error)
I'll definitely take that error!  We nailed the first mountain stage.  Below is our Stage 10 prediction:
  • Stage 10:  4h 05 31" (prediction)
I fully expect at least one cyclist to come in under four hours.  But what will strategies be like?  Will a relatively short flat stage after a rest day be competed for in an all-out manner?  Will cyclists save up for a longer flat stage the next day?  The leaders in the full classification certainly won't be going all-out for the win.  But will their teams reign in breakout cyclists going for glory?  It will be interesting to see what happens after the first rest day.

08 July 2017

One great prediction, one not-so-great prediction ... again!

Our model hit Stage 7 rather well, as shown below.
  • Stage 7:  5h 03' 18" (actual), 4h 56' 08" (prediction), 07' 10" fast (-2.36% error)
I definitely like coming in under 3%!  Stage 8, however, was not so good for us.
  • Stage 8:  4h 30' 29" (actual), 4h 51' 54" (prediction), 21' 25" slow (7.92% error)
I'll be most interested to see why we were a bit slow on the past two hilly stages.  Tomorrow is a travel day and I'll miss watching Stage 9, which contains three brutal climbs.  I'll definitely check it out on replay!

07 July 2017

Back on track!

After a not-so-great Stage 5 prediction, we hit Stage 6 to better than 3%.  Below is the comparison of the actual winning time to our prediction.
  • Stage 6:  5h 05' 34" (actual), 4h 56' 51" (prediction), 08' 43" fast (-2.85% error) 
I'll be traveling again tomorrow, so I'll have to check the online results when time avails itself.

05 July 2017

One great precition, one not-so-great prediction ...

Stage 4 finished for us rather well, but not so well for the riders involved in the controversial crash.  It's a shame Peter Sagan won't be in the rest of the race.  He is a great cyclist to watch, but his elbow cost him a chance at another points title.  I'll also miss seeing Mark Cavendish, who was the unfortunate recipient of Sagan's elbow.

Today's Stage 5 had a grueling climb at the finish, but cyclists completed the stage much faster than we anticipated.  We thought perhaps a little energy would be kept in storage today, but only three of the 193 cyclists were slower than our prediction.

Below are comparisons for the past two stages.
  • Stage 4:  4h 53' 54" (actual), 4h 48' 37" (prediction), 05' 17" fast (-1.80% error)
  • Stage 5:  3h 44' 06" (actual), 4h 04' 49" (prediction), 20' 43" slow (9.24% error)
When I return from traveling, I'll need to look at Stage 5 more closely and determine where we were slow.  The first part of the stage to check will be the final climb.

Below are predictions for the next four stages.  That will take us to the first rest day on Monday.
  • Stage 6:  4h 56' 51" (prediction)
  • Stage 7:  4h 56' 08" (prediction)
  • Stage 8:  4h 51' 54" (prediction)
  • Stage 9:  5h 11' 15" (prediction)
With Chris Froome in the yellow jersey and the big climbs still to come, can anyone keep him from winning his fourth Tour de France???

03 July 2017

So far, so good!

I'm currently in West Virginia while traveling.  I had to sneak a few minutes away from other activities to report on the first three stages of the Tour de France.  Our model is doing well!  I thought we might be a tad slow on the time trial, and we were, but we've nailed Stages 2 and 3.  Below are the comparisons for the first three stages.
  • Stage 1:  0h 16' 04" (actual), 0h 16' 48" (prediction), 44" slow (4.56% error)
  • Stage 2:  4h 37' 06" (actual), 4h 42' 08" (prediction), 05' 02" slow (1.82% error)
  • Stage 3:  5h 07' 19" (actual), 5h 04' 57" (prediction),  02' 22" fast (-0.77% error)
It's great being under 2%, not to mention being under 1%!  I'll try to check in on the next two stages and then get more predictions online while on my travels.  Team Sky certainly looks strong again this year!