Auto-translated
The Lizard doesn't consider many options; thanks to his extensive experience, they are already ingrained in his subconscious. During gameplay, he often doesn't even fully analyze the situations, especially the non-standard ones, and can make mistakes out of nowhere. There's no time to think, forward, forward!! Speed!!
But this doesn't negate the fact that his decisions are generally (in almost 100% of cases) good. His subconscious is working. So, I like the direction of his thinking here.
Basically, what I want to say is this:
Full randomness can always be ideally mitigated not by completely banning randomness, but by using pseudo-randomness. And if done properly, this idea can be further developed in this direction.
That is, the principle remains, but its activation cannot exceed certain critical thresholds - in our case, the obvious ones have already been mentioned:
1) A solo stack does not trigger morale in the first round.
2) Out of five stacks, at most one triggers morale in the first round.
I would suggest thinking further about something more universal, like a maximum of [percentage of morale multiplied by 2 or 1.5] in all other rounds (including human battles).
The path of pseudo-randomness was once applied in Dota and made the game much better.
Wild critical hits stopped occurring one after another, misses and dodges stopped triggering endlessly.
All these <5% probabilities are fun to watch from the outside (and preferably from a great distance), but when it negatively affects someone in tournaments, which, for example, regularly happens in Hearthstone, it actually looks like madness.