"EDIT: I'm not saying Kestrel did or didn't do anything shady here. I highly doubt they would have done something so drastic as use deep dish wheels on their bike, and box rims on the others. The cynic in me just thinks that any company will do small things to make their bike come out on top in testing"

Well...for you cynics...I think you're trying to glean too much out of reported data. Given the design elements of this Kestrel, its entirely believable that this bike belongs in the same grouping as the P4, Shiv, etc....and that the apparent gap to the previous generation of bikes (of which the Fuji D6 belongs by virtue of its premier design elements)...is entirely believeable.

So you're probably safe in choosing the 4000 over a P3, just as the graph indicates.

However...even if you had the data from Kestrel of a P4, Shiv, etc...included in this graph...

You'd really have to take examples of those bikes set up specifically for you to a windtunnel to determine which might be faster.


For one thing, AC, Tom A., Steve Hed, John Cobb, and others have all said or at least concurred that comparing wind tunnel results from one event to another is VERY tough to do, if not just downright inadvisable given the numbers of variables that would have to be controlled for to equalize the data sets.

These companies all use many of the same facilities...which are open to whomever is willing to fork over the $ for time in the tunnel. No reputable company is going to risk their reputation in a low margin marketplace by reporting false data.

I guess what I'm trying to say is that most of us don't have the time or inclination of guys like Tom A., AC, etc...to go do in depth field testing. So the next option is to gather as much reported data as you can find...and look for trends. Right now, the trend is for the serious players to be producing bikes that are clearly going a step ahead of the P3C generation. Without paying for my own wind tunnel time with multiple frames...I can't say that the Kestrel is definitively among that group. But I think its a pretty safe move to believe the Kestrel data showing a break with the previous generation in this instance. On the other hand...if the data included a P4 and showed a similar break with that frame....I think we'd all be correct in raising the "Huh?" banner.

Don't let your cynicism lead you to inaction if you're looking for a new bike.

"The cynic in me just thinks that any company will do small things to make their bike come out on top in testing."

So my question is, given Kestrel's statement on the graph that the bikes were spec'd as sold, then if Kestrel's chosen front end configuration is tested more aerodynamic than Cervelo's...is Kestrel doing "small things to make their bike come out on top in testing? Or couldn't you also just say that Kestrel is being smart selling an all-around better piece of gear? In other words...maybe Kestrel is just trying to sell the consumer a better piece of equipment...and gaming the test data isn't the goal of the chosen tested specs. (I have no idea whether Kestrel's spec'd front end is better...just using it to illustrate the greater point).

"BTW, I have to respectfully disagree with Mr. Harad that they have come up with "the standard on testing" in using bikes "as spec'd" for the testing. That would make sense if the buyers weren't allowed to switch wheels, bars, or saddles even when racing. "

Yeah...I agree in principle. But I can see Kestrel's philosophy in that most P3s go out the door with Cervelo's spec'd front end on them. This will be the same for Kestrel 4000s...most buyers will not change out the spec'd item, at least initially. So if they want to use the saddle, bars, brake levers, shifters, components, etc. on the given spec...then fine by me, as long as you state that is the philosophy. On the other hand, many buyers will own race wheels other than the stock wheels on the bike. So it wouldn't really help the average buyer unless they standardize the wheels between the tested bikes to control them out of the data.

That won't satisfy the data dinks like you, Tom...but if I were a manufacturer...I wouldn't be targeting people like you with published data...at least not today in 2009...where the true aero geeks are going to do their own testing and analysis to make a more individual choice.

Is that marketing? I suppose it is in some respects.

Just conjecturing, to be sure, as I have no other data than anyone else here.

Even ignoring wheels, the P3 comes stock with non-aero cranks, pretty low-end aluminum base bar upturned, and significantly upturned extensions, which would all test poorly.

Cranks are not aero. Zipp made a full-on aero version and claimed 8 seconds in a 40K. That's a bit of a stretch.

pretty low-end aluminum base bar upturned, and significantly upturned extensions, which would all test poorly.

Don't kid yourself; the Vision base bar is the best bargain in tri. I've seen results where it equalled or exceeded another brand's flat bar with integrated brake levers. Vision's brake levers are designed to be used on the upturns and are angled accordingly. I've yet to see any prove that cutting off the upturns actually made them faster. The home jobs I've seen traded exposed surface area in one place for exposed surface area in another.

Chad

I think we all need more precision and info from the kestrel guy with pic of the configuration of each bike they tested.

I have no douth that if they really want to be open and honest about this test, they will show it to us. But for now...that test with those info given is worthless....

---

Jonathan Caron / Professional Coach / ironman champions / age group world champions
Jonnyo Coaching
Instargram

Just as an FYI, it was 8 seconds over the next closest AERO crankset, which *I* *assume* was the FSA NeoPro. I don't know the data over say a SRAM Force crankset. Just the facts, ma'am.

---

"Non est ad astra mollis e terris via." - Seneca | rappstar.com | FB - Rappstar Racing | IG - @jordanrapp

"
Post: Can we start a ST.com fund for the testing?
Quote

I'm in for $27. "



I wouldn't waste even that much money on generic testing. If you're totally invested in having the fastest bike, a little patience will turn up enough data to narrow the field down to just a few suspects. Then...

You can all just pitch in $10k for me to take my choices into a tunnel with a well-designed, transparent protocol...and I'll be happy to tell you which one tests fastest FOR ME! ;-)

Barring that...just write P4, 4000, Shiv, Speed Concept, Giant TT, Plasma 3...post them on a dart board...toss a dart...and be done with it.

The rest is in your legs.


PS....if you're paid to ride...none of this applies. Shut up and ride what they give ya! ;-)

The $27 seems so arbitrary that I will have to second it, and volunteer another $27 for testing.

Chad

i agree, the vision tech is the best buy for your money. BUT to cut the upward part and put a good set of aero levers was a kick ass move aerowise. Not even close to the uncut version. I wonder what data you saw.....????

---

Jonathan Caron / Professional Coach / ironman champions / age group world champions
Jonnyo Coaching
Instargram

While I agree that it's hard to fairly compare bikes using manufacturer tests, I think you can using the following assumptions:

1) The P3 in Cervelo's tunnel tests was set up in a favorable position.
2) The 4000 in Kestrel's tunnel tests was set up in a favorable position.
3) A company reporting ~675 grams of drag on a bike at 0 yaw did in fact manage to measure the bike under that much drag in some way.

Putting those facts together, it seems to be a reasonable assumption that the 4000 and P3 are similar bikes aerodynamically. Honestly, however, who cares? If you're looking at a 4000, you're looking at the P4, not the P3. If Kestrel wants their marketing to hold up here, and sell bikes off of it, they need to show it in a legitimately fair, well documented comparison with the P4.

If they want a fancy chart for dealers to use that will increase sales, they've done it.

Cranks/chainrings are not wheels. Even so, 8 seconds over a 40k isn't remotely unreasonable compared to a decent crank, such as Dura-ace. There's ample opportunity for air to go turbulent through all those holes.

Ever looked closely at the FSA NeoPro? It's pretty, but it's not optimum. Eyeball wind tunnel says that the Zipp ones are faster. 8 seconds? Maybe, maybe not. Tough sell, since the data is unsupported from Zipp. Bottom line, I'm almost certain that the Zipps are faster (nobody else has even tried to compete), and I would be almost certain that they save over 8 seconds over a standard chainring; probably significnatly more.

We are getting a little off track on the Kestrel test, but I called Zipp when those came out and was told it was 8 seconds faster than the FSA Neo Pro, and 12 seconds faster than a "typical" non-aero crank. The test was done on a bike with rider pedaling.

BUT to cut the upward part and put a good set of aero levers was a kick ass move aerowise. Not even close to the uncut version.

Did you test that personally? The problem with home testing is that once you've cut them off, they are off.

Chad

no chad... i didnt test it myself. But i sure cut mine when i saw the test results ;)

---

Jonathan Caron / Professional Coach / ironman champions / age group world champions
Jonnyo Coaching
Instargram

Ok, so I may have exaggerated a little bit. I do trust the manufacturer data enough to place each bike into a specific category (in this case, the P4, Kestrel, Shiv, Giant, Trek SpeedConcept, etc. category).
My main thought was that, among the bikes listed, manufacturer data is useless in determining which is "fastest" within the category. (I think we're in agreement here, but I could be wrong. Wouldn't be the first time)
While the Kestrel most likely is a better choice than the P3, I'm skeptical that the size of the gap between the two is entirely accurate.
I'm not in the market for a new bike, but if I were, aero data would not be at the top of my list. I would be more focused on price, comfort, fit, etc.

And with regard to the "small changes" I was referring to in the testing, I didn't mean components so much as just the way the bike is set up. For example, leaving the cables a little bit longer on the competitor's bike, using an extra spacer or two, or testing without a seat (like another company did), or small things of that manner.

hmmm, I had been considering cutting mine soon. If it's as good as you say, I just might do it. Do you have the results for it?

Ahhh, so you have seen something definitive. Is it shareable? :) I have not and would have to be convinced it was significant before I gave up the safety of my upturns. On the other hand, if I had 12 bikes like you do, I'd have a race bike set up that way and keep the upturns on my training bars.

Chad

It will also matter which year P3C they tested... ie with the latest 3T fork?

That was a big difference.

I also believe that they did test with similar wheels, but then the marketing guys said: "we need to show "data" that maximizes the differences between us and the competition"

And this graph was developed.

That is sad...really, really, really sad. If you want to do a fair comparison then the only thing that should differ are proprietary components, e.g., it makes sense to test a Specialized Transition with its special brakes and compare those results to a Cervelo P3 with standard calipers. Otherwise, though, everything should be identical....

FWIW Andy, as shown sans rider, this P3C tested at 689g at 0deg yaw in 2006 at LSWT

so, not sure what wheels they used to get it up to 850g

From the chart :

>>Complete bicycles as they would be received by the customer

So they take a stock P3 with training wheels and compare it to their top of the line bike with race wheels? If they did that I think the entire industry would think Kestrel management were a bunch of lying, cheating, deceptive opportunists. Time will tell.

Yeah, that makes sense as my wife's P3C track bike came in just a little below that at TAMU. It is smaller than your bike and the conditions weren't identical (e.g., we tested it with less aero bars and with pedals attached), but everything still says that the only way a P3C would have as much drag as Kestrel claims is if a parachute were attached. ;-)

Maybe they ran the tunnel @33mph or so? I didn't see the tunnel speed listed there .. in any case this 'data' raises more questions than it answers ...

heck my old alu P3 tested out at around 850 g-f drag @30mph (untared). Full bars, cables, extension, cranks, brakes, pedals @0/180 ....

"the only way a P3C would have as much drag as Kestrel claims is if a parachute were attached."

Or...what might it also say about the A2 tunnel vs LSWT? (Serious question...based on some of your previous comments on tunnels and testing, I am curious what amount of the difference, if any, might you REASONABLY attribute to different facility, conditions, etc. Also because when I read the above, I hear that you don't feel the P3's number is reasonable even if Kestrel is guilty of testing a full-up 4000 against a stock, non-race wheel P3. Hard to determine the amount of hyperbole in your parachute! )

The only wind tunnel that I have ever been to is the one at Texas A&M, but my understanding is that the numbers from different facilities can differ. That said, I have heard nothing but good things about the A2 wind tunnel - in fact, others (I can't recall who) have mentioned that it is the LSWT that may be "outlier". In any case, I can't imagine that the differences between various tunnels is large enough to explain the apparent discrepancy in these data.

As for the parachute comment, that was indeed hyperbole, as based on both the magnitude of the difference and Steve H.'s comments in that interview the most logical explanation would seem to be that they did indeed compare the bikes with different wheels, etc.

A little birdie told me that "things were done very fairly to all bikes in the test. The Kestrel is quite impressive" This little birdie was very close to the action and impartial.