I have to admit it, the Dimond bike is a head turner.

It looks fast. But the question is, is it fast? When I switched over in the middle of last season from my tried-and-true Specialized Shiv (the original with the nosecone), I took a little bit of a leap of faith that it would be. Sure, there was some wind tunnel data, but this data gets endlessly debated on Internet forums by “experts” of all kinds, so it’s hard to know. Also, it had just been ridden to a convincing victory by pro TJ Tollakson at Ironman Mont Tremblant, but pros are likely to be fast on any bike.

What is important to know is whether I am faster on it than on my old bike. For that, we need data. Luckily, I have lots of data-gathering devices – especially the power meter – and I also like to do the same races either every year or every two years. Call me a creature of habit.

The key in performing an apples-to-apples comparison is to eliminate as many variables as possible. I only had a couple of races on the bike in 2014; the second one was Ironman Arizona. Here’s a comparison between 2014 and 2012:

The overall times aren’t quite right since they’re when I started/stopped the bike computer, but they’re close enough. The heart rate data is almost identical, and the power data is pretty close. The split is a lot slower on the Dimond, though, so at first glance it wouldn’t appear to be a faster bike at all.

However, 2014 was really windy, and 2012 had no wind to speak of, so the conditions weren’t comparable. In 2012 I had the 4th-fastest bike split in the age group, and I did again in 2014, so that might argue that the two performances were equivalent. However, placing is a factor of who shows up, and two trials aren’t enough to claim a trend. So onward we go.

Then comes last weekend’s Ironman 70.3 St. George in southwest Utah, a venue that’s become one of my favorites. I did race here last year on the Shiv, so this race presents an opportunity to generate more comparison data. First, the tale of the tape:

A minute faster in the swim and 4 minutes on the bike (we won’t talk about the run). So the Dimond scores on the initial numbers; however, we have to examine two main variables: was this year’s course faster than last year’s, and how did the efforts compare in terms of watts?

The course was slightly different this year; while it eliminated some strange, slow sections on a bike path where two parts of the course crossed one another, it did add a section on a bike path in Snow Canyon, which we then partially descended and had to execute about the tightest 180-degree turn I’ve ever encountered in a triathlon. Then climb the canyon. The total elevation gain recorded by my Garmin, however, was nearly identical at just under 3300 ft (1000m for the rest of the world).

Another way to look at relative difficulty is to look at the pro times. Since many of the same pros raced this year and last, they’re a good benchmark, since as a group they have the most consistent fitness year over year, and they are racing for a living, which means they’re both supremely motivated to push hard and also able to sustain a higher intensity level than age groupers, in part because their race doesn’t last as long. 🙂 (Joking aside, that’s actually true if you look at the formula for how Training Stress Score is calculated – to achieve the same metabolic cost, measured in TSS points, for a workout, if your duration is less, then your Intensity Factor is higher, actually by its square.) Anyway, if we look at the male pros:

(2014 is on the left and 2015 is on the right.)

What we see is that the same pros year over year were mostly 2 minutes or so slower this year than last year. The male pro race dynamics could have been different, though, so let’s look at the female pros:

Not quite as many repeaters as the males, but of the 6 or so, the gap – other than Heather Wurtele, who was not quite 2 minutes slower – is 3 minutes or more.

That tells me that this was definitely not a faster year, and in fact was probably a bit slower (I’d love to hear from any of the pros about the differences from their perspective between this year and last, especially if I’m missing something key in my analysis).

The other possibility, then, other than the bike being faster is that I worked harder this year, i.e., put out more watts. But that doesn’t appear to be the case:

Compared to 2014, my faster speed in 2015 was achieved at 6W lower normalized power and 10W lower average power – for .8 mph better average speed. I hadn’t worn a heart-rate monitor in 2014, so I can’t compare the efforts along that dimension.

Finally, if we go back to Strava, we can use the segments to compare each year to see whether I rode the course any differently, i.e., did I push the uphills or downhills more in one year than the other?

I’ve just shown the beginning and end pages here (I skipped another couple because I see many eyes glazing over already), but you can see the trend: I did start a little harder this year (2015 is on the right), but after those initial climbs I was faster on pretty much every segment, and at lower wattage. In 2014 I pushed up Snow Canyon a lot harder (I was also probably fitter, having come off of Ironman Los Cabos at the end of March), but even so my splits are very close. I have a few theories:

1. The Dimond definitely seems faster on the flats and downhills. I saw the same phenomenon on training rides, when I started getting Strava segment PRs on routes I ride pretty often. These always came on flat and downhill sections.
2. The fact that my uphill splits were so close (and still mostly better) at lower wattage is somewhat baffling. I don’t weigh less than I did last year, so watts/kg doesn’t explain it. The Dimond itself weighs about a pound less than my Shiv, which could be part of it.
3. Both bikes are equipped with Quarq power meters. They both run the same firmware version, but they are slightly different models, so I suppose there could be a slight difference between the power readings on the two.
4. I did run a different wheel combination this year – Zipp 808 front / Super 9 disc vs. a pair of Reynolds Aero 72 wheels – which likely gave me some of the better speed on the flats and downhills, though likely cost me a little on the uphills due to weight. Tires were, in both cases, Specialized S-Works Turbo 24s with latex tubes, so no difference there.
5. Finally – and this could be a big factor – I could be spending a greater percentage of my ride down in the aerobars than I have in the past. This season I have been making a concerted effort to do my hard efforts on the trainer in the aero position, and I’ve also been doing a weekly stretching class (and reinforcing the exercises at home during the week), so I feel as though I maintain the position more easily. I can’t quantify this as a percentage of my race, but it could certainly explain some of the better speed at lower watts.

To be clear, I do believe I have enough data to say conclusively that the Dimond is a faster bike than the Shiv, at least with me on each of them. What is unclear is how much faster. I rode 4 minutes faster on a course that the pros rode 2-3 minutes slower on, and I did that at 6W less normalized power. So that alone would say that it’s way more than 6W faster. However, some of the other variables I listed above potentially contribute to some of that speed gain, so I don’t feel I can say anything more than “it’s at least 6W faster than the Shiv.”

The rest of my season has a lot of repeat races from last year and/or two years ago, so it will offer plenty of additional data points to see whether this trend holds. I’m happy thus far, though, with what I’ve seen from the combination of Dimond and rider. We’ll try and keep improving the rider as the season progresses.