I originally posted this question to learn the purpose of esc caps, and if changing the stock cap would have some kind of advantage. Thanks to replies and especially to Howard Cano for a lot of useful/helpful info, here are the results of my experiment. The stock cap was a single Chinese, 2200uf, 25v, 105* cap. After some Google research and Howard Cano's input, I used four Panasonic 560uf, 35v, 2470mah ripple current, .018Ω impedance, 105*, 5000 hour caps wired in parallel. I was limited in physical size to 10mm diameter, because I wanted to mount 2 on each side of the centrally mounted esc in the triangles of my b4. I wired 2 pairs of 2 caps in parallel to pc boards. I used 18ga wire (instead of the stock 22ga wire), and wired the 2 sets in parallel to the esc.

Unfortunately, most of my results were inconclusive because it is very difficult to test a variety of parameters in controlled conditions in our hobby. Motor and esc temps (114* and 123* respectively) were about the same with the capacitor "bank" vs the single factory cap. However, ambient air temp was a little higher, and the track had more bite, so I think everything is actually running cooler, but I can't empirically prove it. Punch is very subjective, but visually I don't think there was any difference ( despite Murffdog's Octane booster claims.). I was able to confirm that 3 sets of batteries after 6 minute qualifiers, all had 8.0 to 7.98 volts left with the capacitor bank. That compared to 7.81-7.85 volts with the stock cap. Also, the batteries only took 997mah to full charge.

All in all, I think the car runs a little cooler, and the battery has more charge left after a 6 minute race. It was a cheap experiment. The caps were only 70¢ each, 2 pc boards 74¢, plus the wire. For those wondering, the car is essentially a B4.2, 17.5 motor w/12.5mm high torque rotor, max timing, SP Reventon Pro using Stock 0 software, and a 5.13 FDR.

Here is a pic of the installation.