Would it be less twitchy if he moved the saddle back a couple of centimeters and either moved the elbow pads back or swapped to a shorter stem? This would put his weight more over the center of the bike. I do not know how much this would effect power production or comfort. If it did, he could bring the pads up a little to compensate. When I converted the Cervelo S2 to tri bike use, I moved the saddle forward using a zero setback post and used a 2cm shorter stem than I did when it was a road bike. The saddle is certainly not as far forward as on a dedicated tri rig, but the bike still handles ok.

When I add aerobars I tend to have the pads behind the top of the bars a little. To be honest, most people racing could do it on a road bike with clip-ons and be more than aero enough. We tend to look at the FOP on ST and not the MOP and BOP where a majority of the people should be on road bikes for 90% of their riding and just add clip-ons which would be more than enough.

I agree with your point about shorty bars on ITU bikes. If I am going to put clip-ons my road bike, I sacrifice the position in the extensions and leave the saddle and handlebars alone. This results in the middle of my forearms, maybe even closer to my wrists than my elbow, resting on the pads. I used this setup for a gravel triathlon. Not comfortable for an hour, but tolerable for the flat section before the next hill.

With respect to full TT setups for Triathlon, I agree that the base bar can be lower than your hoods position, but not a ton lower. For example when I converted the Cervelo S2, I moved my saddle a couple of centimeters forward and switched to a flat stem that was 2cm shorter. This lowered my "hoods" position on the base horns by about 2cm. The shorter stem also brought the the elbow pads further back (with saddle change now 4cm closer than the ITU shorty bar set up). The reach to the bull horns is about the same as the reach to the hoods on the road bike because the bull horns had more reach than the road bars. This bike works very well in both the extensions and the bull horns. This may because I am strange and ride fairly low with my saddle pretty far back.

I previously owned a Nose Cone Shiv and an Orbea Aletta that I used as a training bike. Both of these bikes were set up with the same pad stack and reach and saddle angle. The Shiv had a bunch of spacers to stack the pads, maybe 8 to 10 cm worth (and I still had at least 10cm of drop from saddle to pads). The base bar position on the Shiv was useless. It felt terrible climbing a hill and terrible out of the saddle. The Aletta had a Profile design T2 aero bar and no drop base bar, leaving the pads about 6cm above the base bar. The bull horn position on this bike was decent. I could climb hills in the bull horns and get out of the saddle if desired. I think it would have been even better had I used an aerobar with the pads closer the base bar, but it was definitely the bike I wanted for a 3 hour hilly ride.

I think most triathletes are best served by bikes that keep the base bars no more than 6cm below the pads. I think they mostly set their bikes up this way, but many use a bunch of spaces and/or upward angled stems to get there. I was wondering if their bikes would handle better and be more aerodynamic if the bikes were designed with more stack from the start, like an aero road bike, and if the default aero bar was one that placed the pads 3 to 4 cm above the base bar instead of 6 to 8 cm.

I understand that most athletes that ride with 8cm or more of drop from saddle to pads need their saddles further forward than me and that most road bikes would handle badly set up like that. But I also see plenty of athletes with elbow pads even with the saddle to maybe 5cm below. Do they really need their saddle moved much forward from their road bike fit position? Would they not be better off on a road bike with clip ons or a full TT cockpit if you want to squeeze out the last few aerodynamic gains?