with all due respect to those who took a shot at answering the question. I'd like to point out a few oversimplifications. Firstly, the froude number mentioned in the paper only suggests a soft threshhold as it pertains to maximal speed. Most modern naval architects do not even recognize "hull speed" as a a term that has any meaning at all, although paddlers like to throw this around alot. While it is a fair generalization to say that if displacement and max waterline beam are held constant that a longer hull has lower wave making resistance at higher speeds, It is also fair to say that skis are made 21ft long so that they could be fast without being inordinately tippy. If one made a 21ft boat as narrow as a k-1 it would be completely unpaddleable. A shorter boat can be faster if the sectional shapes are more semicircular and parabolic to reduce frictional drag, which comes into play even at high speeds as the boat is accelerated with every stroke. Traditional thin-ship hydrodynamics fails here b/c it is difficult to model the roll, heave, pitch, yaw, and squat effects of a cyclical and reciprocal stroke instead of steady-stade propulsion. Most studies of displacement hulls presume a linear propulsion akin to a model being towed in a basin and miss the point entirely when it comes to high speed paddling. At some point in the stroke there is a bit of hydrodynamic lift applied to the boat that actually lessens the displacement to a degree (not planing). On a downwind run the boat is often in "semi-planing" mode (also a term with no real meaning in hull design) at this point other factors begin to affect speed more than paddle power. Smaller folks will obviously benefit from shorter boats, but I find that any boat over 20ft simply won't accelerate as well as a shorter boat with higher prismatic coefficient and less form drag, windage, and frictional resistance, even though the wave making resistance may be predicted to be less on the longer hull. Also, when the speeds are slower due to current, headwind, etc. the resistance curve is different at slower speeds and frictional resistance begins to dominate. A poster mentioned that surfing downwind may be the only time that a 6.5meter boat makes any sense at all, because the velocities are high enough to shift the resistance curve upward off the chart.
It doesn't surprise me that shorter boats can be surprisingly fast as I can keep up with V12's on flatwater (for a while at least!) in an 4.5 meter ICF wildwater boat that is suprisingly narrow at the waterline, has a higher prismatic coefficient than most skis, effectively no rocker, no rudder for added resistance, and will accelerate much faster than a ski.
That said, K1's are as short as they are b/c the rules limit the to 5.2meters long. This is why they have to be so skinny to achieve the speeds they do. They would certainly be closer to about 19ft if the rules allowed it. To state the obvious, One can make a 19ft ski much faster than a typical elite ski if it is designed to be more "K1 like". The likely tradeoffs being downwind speed and far greater tippiness. Also the nelo 560 for instance is more K1 like in it's outfitting than many ocean ski's making for greater possible application of power. I'm guessing 50-75% of the ski owners in the world are on a ski that is too long for their weight or intended application. Big boys in the big swell are the only people who can benefit from a 21.5ft boat of any kind. Until recently, all of the shorter skis were beamier and aimed squarely at intermediate paddlers. It is great to see someone designing smaller, faster skis for serious speed in other conditions and for smaller folks.
BTW Those ridiculous 23ft boats you see for adventure racing are a joke, they aren't any faster for any purpose and will soon be on the scrap heap.