Your description is misleading. The force available to the pad-rim interface is a function of the angle formed between the cable and an imaginary line drawn through the center of the anchor and the pivot. The way you described it makes it sound as though the length of the cable affects the leverage which is incorrect. Or perhaps that the "height" of the cable above the pivots has any effect on leverage. The moment arms (there are two per side) are fixed by the shape of the caliper, neglecting pad wear and vertical adjustability. What's important is the angle at which the cable pulls on the moment arm. However, even assuming the differences are much greater than they look I doubt the loss is as high as 75%. The TRP caliper has been better optimized for this sort of actuation by using a longer upper and shorter lower moment arm and by increasing the angle between them. This latter may be what gives you the impression that the "height" of the cable run above the pivot is important, but that's not really the case.

To restate: the moment arm (lever if you prefer) is fixed for a given pad adjustment when pad wear/compression and caliper flex are neglected. The only thing varying between the setups is the angle at which the force is applied to the moment arm. Don't believe me? Draw the FBD, it shouldn't take more than a few minutes.

With that out of the way, I do have a few ideas for improving leverage with this configuration:

The first is obvious; bend the calipers. This would work great if the likelihood of breaking them wasn't so high. It's probably out as a solution.

The second I'm not sure would work as I've never used the necessary part. Does anyone know if it's possible to use a travel agent backwards? I've not used one to know exactly how it is they work but it seems like they use a sort of virtual cam in the pulley which could probably be re-oriented to do the reverse of what is intended. This would, of course, require the brakes to be adjusted much more closely to the rim, reducing clearance. It would also accelerate pivot wear. So it's not really very elegant either.

Another idea that I think could work is adding a roller between the arms of the caliper. Friction will be a small issue here but probably not too much. It does spoil some of the charm of the whole set-up in that it looks more busy and requires more clearance above the caliper, though not nearly so much as using a straddle. This one could be a winner depending on the application.

Lastly, fabricating "stand offs" to move the cable anchoring points up and in toward each other could be done in such a way that the only net effect is increasing the angle between the moment arms (desirable because it increases the pull angle of the cable) but without elongating the lower arm. The big drawback here is the additional weight and complexity.

After all that you're probably wondering "why bother?" and you're right, but how else should I keep myself occupied? Sometimes you just have to do stuff because it's there.

For those interested in trying, the first order of business should be to set the pad as high as possible and use the lowest profile pads you have, both of those are easy leverage gains.