I disagree with the assessment of the results. Certainly, there were some bullet designs tested that opened too fast and failed to penetrate, but there were others tested that opened to full diameter and still penetrated denim and gel to 19 inches from the LCR -- Speer Gold Dot and Buffalo Bore.
At 1300fps, the Buffalo Bore cartridge is going to have a lot of recoil, but that same Barnes XPB bullet is also loaded by Barnes brand (Remington) factory ammo and by Federal. In those more moderate loadings, they have both achieved 16+ inches of penetration in denim and gel tests with velocities of around 1170 and 1100 fps respectively from a 2" barrel.
The measurement of a successfully opened hollowpoint has to be considered. With a petal-type hollowpoint like a Black Talon, SXT, Gold Dot, XTP, Golden Saber, and especially the Barnes TAC-XP (which is all petals and no lead core), the bullet opens to a maximum diameter, and then if it continues to be driven, the petals are forced back and folded against the base of the bullet. A final measurement may show 0.50" but the bullet actually opened as much as 0.70" in the wound channel, and if a cartridge is designed right, it's opening to maximum diameter where it counts. The Buffalo Bore is "over-expanding" but at 19". Within the AR's test criteria, which they apparently adopted from the FBI, at 12" to 18" the expansion could have been optimal.
As for recoil, yes, .357 has more of it than black-powder pressure .38 Special curios and relics. Whether this kind of recoil interferes with the function and purpose probably depends on the shooter. I've read gun writers dismiss the 9mm LCR for excessive recoil with minimal ballistic gains over .38, but I've also seen shooters clear plate racks in the same amount of time whether using a gun and cartridge combination of of low-recoil or high. I think the "follow-up" shot speed thing is mostly a myth.