With crankarms too long, you may hurt yourself and be less efficient, by throwing your hips around. With crankarms too short, there isn't a huge negative aspect, though you may have a less efficient stroke.
Throw caution to the wind. Go with 171.25. That's where the real sweet spot is.
I just got through reading about this below.
The "logic" of "crank length should be proportional to leg measurements" has been around for a LONG time, and lots of people have turned that "logic" into a formula for determining crank length. Only one problem: the research doesn't support it. One key feature that is often ignored in these discussions is the duration of muscle contraction that is controlled by cadence. It just may be that there is an optimum here, which is why there is a fairly narrow range of cadence for optimum performance. Longer cranks tend to mean lower cadence, moving you out of that optimum range. Crank length has been a point of debate since the introduction of the "safety" bicycle in the late 1800s, and there have been all sorts of fads in that regard.
There is no reliable formula for predicting crank length. There ARE lots of formulas out there, but they are just figments of the imagination of their purveyors. No one has ever done a study that shows how crank length should relate to anything.
You will find no high quality data to support any particular crank length as being better than any other. This is true whether or not you correct for leg length, femur length, etc. On the other hand, you will find lots of anecdotal or low quality data to support all kinds of conclusions, and more theories than you can shake a stick at. A rider's response to changes in crank length is 1) highly individual, 2) dependent on riding style and the event (TT, climbing, crits, track racing, etc.), and 3) most important, highly adaptive. This is why it is so hard to study the effect of crank length.
A 2008 study by Jim Martin, Ph.D., from the University of Utah shows zero correlation between crank length and any performance factors.
Fred Matheny Summary: There have been studies of crankarm length, but the results aren't consistent. Some show that longer cranks provide greater leverage for turning big gears. Some show that shorter cranks foster greater speed via a faster cadence. And some show that crank length is completely individual.
So, longer crankarms aren't a panacea for time trialing. In fact, there are dangers associated with them. The added length makes your knees bend more at the top of pedal strokes and extend more at the bottom -- both of which can lead to biomechanical injuries if you jump from 170 mm to, say, 180 mm.
Also, longer cranks reduce cadence -- and a brisk cadence is the key to good time trialing.
Oddly, I have less knee pain with the slightly longer crank arm. The longer length allows my knees to more closely achieve their natural flexing arc. On the shorter crank, my knees start to complain because they're unable to get quite high and bent enough.If you have no knee issues, go with 172.5.
If you have had knee issues in the past (but not currently), and you think you may develop knee issues down the road (because you will plan to ride a lot more), then go with 170.
When in doubt, go with a shorter one..