Timmi

Allright... I made some verifications... I went out and sat in the sun with a calculator, pen and paper. (I ended up plunking it all into my spreadsheet anyways, but it was too nice outside, I couldn't resist).

Now just a note so you know I'm not trying to start a war here! So everyone, please, stay calm, as this may shock you! If you have a heart condition or are easily offended by alternate points of view, please change the channel now! ;-)

The purpose of this is to examine the feasability of replacing a triple with a double-compact. And because Richard gave us an example of a triple that works really good for him, he gave us the opportunity to do some analysis with a good triple as our baseline.

Some of our criteria shall be to minimize chain flex (crossover), unecessary sifts, and too wide gear gaps when changing from one chainring to the other. (if I read everyone correctly, I think that pretty much encompasses the discussion here).

Please take note that one thing I don't want to get a retort on, is that these are not actual gears. I know! So let's just call it a "multiplication factor", which inserts (in bold) into the gearing as follows:
(Chainring/Sprocket)*wheel circumference=development
It doesn't matter whether you're looking at just PART or the total formula, because it's all multiplication and division... so proportionally it stays the same. So lets reduce things to their simplest form, shall we? That way it applies as much to 20", 24", 26", 700C wheels.

As a side note, since we're on the topic of cranks, I may as well mention the entire formula we would concern ourselves with, if we calculated what this all means to us ERGONOMICALLY. Thus, the formula gets a little bit longer:
PedalStrokeCircumference*ChainRingT/SprocketT*WheelCircumference
(where PedalStrokeCircumference = (CrankarmLength*2*pi) ...I don't substitute "wheelcircumference" with a formula because you SHOULD measure it right with marks on the ground and rider weight reducing the radius at the crush point with the ground - or contact patch if you prefer).
That last formula is to compare SPEED (NOT RPM) your feet are going in relation to the wheel's linear displacement on the ground. Speed as in muscle-twitch speed - important. This is useful when changing from a 170 to a 165mm crankarms for example... you'll feel like you're pushing bigger gears, so watch out. For this reason, you can use smaller gears, and because of reduced circumference, it's easier to spin, and with the increased RPM bring the speed back up to where it was.

The below table has 3 sections, as you can see.
First, is Richard's gearing - this will be our baseline, our starting point.
Second, Compact with an actual SRAM cassette configuration that can be bought in all good bike stores (and on eBay if you like to shop there).
Third, is a proposed configuration, for which not all derailleurs will work. However, I've seen some new rear derailleurs come out in the past year that can handle a gap of 6 teeth quite comfortably.

Triple Crankset, our baseline:
11 - 12 - 14 - 16 - 18 - 21 - 24 - 28 - 32
50T
4.55 4.17 3.57 3.13 2.78 2.38 2.08 1.79 1.56
39T
3.55 3.25 2.79 2.44 2.17 1.86 1.63 1.39 1.22
26T
2.36 2.17 1.86 1.63 1.44 1.24 1.08 0.93 0.81

Compact Double, with standard/stock SRAM cassette:
11 - 13 - 15 - 17 - 20 - 23 - 26 - 30 - 34
50T
4.55 3.85 3.33 2.94 2.50 2.17 1.92 1.67 1.47
34T
3.09 2.62 2.27 2.00 1.70 1.48 1.31 1.13 1.00

Compact Double, with more aggressive spacing in the smaller gears:
11 - 13 - 15 - 17 - 20 - 24 - 28 - 33 - 39
50T
4.55 3.85 3.33 2.94 2.50 2.08 1.79 1.52 1.28
34T
3.09 2.62 2.27 2.00 1.70 1.42 1.21 1.03 0.87

And btw, if your brand of compact is a 50/36 rather than a 50/34, the numbers aren't that far off the above.

Now, looking at Richard's table, it is easy to see why he likes his gearing. When he shifts on the Front, from the Large Chainring, to the Middle Chainring, he is only skipping One Gear. So, instead of landing in the next sequential gear (as if he'd stayed on the same chainring), he is skipping one, thus shifting by two. Not a bad thing when you're meeting up with a hill. Great!

Things get less optimal though, when Richard goes from the Middle Chainring to the Small Chainring. He skips Two Gears (downshifts by three gears). And most of us at that point would certainly shift the back at least one higher to bring that back in line before shifting down again as the ascent continues.
At a glance his largest gap seems between 50x12 and 50x14 at around 0.6 - there is a big sudden jump in percentages, higher than the difference between x11 and x12. Not perfect, but not dramatic either, except we may have one high gear too many, a slot that could be put to better use in the lower gears.

Now, I don't know how many of you are like me, but I just HATE shifting gears in a hill, and the less the better.

In the middle, you see a Stock Compact Double setup (some compacts will be 50-36, others 50-34, so depends on which brand you have - I chose 50-34 for the example because I thought it might be better for the granny-gear).

The gap between 2 gears in this setup goes gradually from 0.7, in the highest gears, to .5, .4 in the middle gears, and .3, .2 in the smallest gears... this is probably a nice progression in percentages. This is good... when Richard is going downhill, he wants wide shifting swings, because he's got gravity helping him out, and he's resting anyways (compared to going uphill), and just having fun... and in the small gears, the gap gets smaller to allow to better fine-tune for those climbs where it's more important.

Now when you shift from, say, 50x23 to 34x23/34x24, you can see that you're also just 2 gears lower. We're coming to a hill and we just shifted down 2 gears a single front shift.

So as you can see, we have about the same gaps. Percentage difference is only slightly higher between gears. We're doing less shifting and landing closer to an ideal gap when shifting on the front.