You also have to factor in that in certain spike windows with the right classes you can have up to 37% additional crit rate (15% Chain Stratagem for 15s on one target, Litany at 15% for 20s, 5% AST using Spread Spear over 20s, AoE, BRD passively giving 2% via songs, near 100% uptime, these are additive, not multiplicative) stacked on the whole party over a short but still substantial period of 15s, 52% if an AST caters to you specifically by giving you an enhanced Spear. This is admittedly only 15s, about 6-7 single target GCDs out of every six minutes, but this still happens twice per most fights, under potion timings no less, and these buffs still get used independently from one another in-between. That's the problem with a hard cap. Without the cap you're able to overcap crit rate with raid buffs in specific scenarios without factoring in gear at all, which is why I initially said to lower the rate to 30% at 5 stacks. It's still unreasonable as you've noted, but puts you on par with where you wanted Risk to be in power while still mostly working within the system we've got.

This is also why I find Bard particularly frustrating, to be quite honest, because they get to basically triple the length of these ridiculous windows for themselves thanks to DoT snapshotting and it plays heavily into their performance. Access to these buffs can make or breaks the classes that are that dependent on crit, which worries me about your concept (though really as long as you don't do what BRD does and extend the benefit of the window somehow, then it's fine). Yet this kind of synergy is also why there can't be a hard cap implemented in the game at all. Because most of it is intentional.

The alternative involves a juggling act between these raid buffs to ensure they never overlap while maximizing their individual benefits under what's attainable with gear, which SE does not want to enforce in their design. I get the idea that you might want a melee DPS that can crit so hard that they don't need these buffs, but SE does not like that approach to balancing classes. It's not good that this can be done and abused heavily on Bard specifically but SE would also not want your alternative, where it makes you weaker with certain classes instead. It would make these raid buffs needlessly frustrating to coordinate around for everyone and specifically provide nothing to your kit. It's essentially the Bard problem in reverse, an anti-synergy as opposed to a synergy. Thankfully, even though they were dumb enough to let Piercing and this sort of stacking exist for so long, they have gone on record to say that the former makes both of the ranged DPS too good and are starting to look at it, finally, so I hope that means they'll address it (and hopefully skill speed being garbage too) next expansion.

I think you could honestly accomplish a similar feeling to the current Risk concept with a built-in baseline proc rate tied to Risk specifically that randomly boosts the potencies of certain skills by 30-40% rather than relying on a stat that already exists. It feels clunky, but it would allow you to add traits to tune all of your kit around it based on level, and wouldn't affect gear priorities at all. You would still be able to create that hard hitting but inconsistent feeling and focus more on leaving more room for skill speed as a result too if you wanted.

I'm fine with the concept you have in mind for the momentum changes too. Just keep in mind that the most efficient option will almost always be to simply spam the highest potency per momentum ability or GCD during each 5 risk stack window as much as possible, regardless of what that is. A combo potency on the saps/chains could fix this if they're oGCD, but feels a little counter-intuitive putting a combo chain on an oGCD.

And really if you want to address the defence issue, just make it only apply to physical defence or more to it than magic defence. There was a time in ARR when DRG was simply kicked from raids due to not having enough of the latter on their gear. That, especially when combined with old BfB, lead to unavoidable one-shots even under stoneskin in some cases. It helped start the 'tank the floor' meme just as much as the old jump animation locks did. Last I recall it had something to do with the mitigation formula, making direct boosts to the both defence stats far weaker than they seemed and the penalties even stronger. Maiming gear had to be buffed on the mdef side to match Monk, but they still have higher physical defence built in compared to Striking gear, so it should work out fine at 20% there.