No way do the hawks win that argument.
Imagine you're the engineer involved with launching a nuke. In times of peace even if you get a phishing call to launch a nuke, you'd have second thoughts and probably take a couple more steps to confirm. In war time, you're already primed for the possibility of the nuke being used and it takes far less for a malicious actor to end the world as we know it.
Same for all levels of decision makers. It really just takes one trigger-happy maniac to make a wrong decision somewhere.
One reason I think this to be true is exactly the risks that you mentioned - the more hot the war gets, the more bodies, the more likely mistakes are made.
In a hot war where multiple countries are involved you can't expect the fighting to just stop once you hit a border. Even when you have the upper hand, the temptation to carry through once you are already engaged in a conflict is real.
I agree that NATO would be very reluctant to so much as fire an artillery shell over the Russian border, but Russia might well forfeit a great deal of personnel and equipment in such a move, anyway. Having troops & equipment abroad is a major liability if you pick a fight with a country (or coalition) that badly outclasses you.
Let's say afterwards Russia nukes Western Europe/the US. What then? That's the thing with MAD, you have to be mad to toy with the possibility.
What would make him desperate enough to sacrifice the entire current and future of Russia? Losing an invasion of Estonia?
When you are 70 years old, spending half your life revered as a great leader of a nation, you would be desperate not have your historical legacy ruined by losing a war and possibly your status and pride in your final years.
I'm definitely of the opinion that this would be an unlikely scenario, but then, are you willing to bet the future of the planet on a hunch that a single person that you don't know personally is actually not unhinged and desperate?
I don't know about you, but I'm personally going to avoid betting on anything that has a mere 1% chance of resulting in nuclear winter.
The scenario that we have been discussing is NATO's response to a Russian invasion of a NATO member such as Estonia. What are NATO's options?
1. Full scale (offensive) war. They attacked us, we will end this in Moscow 2. Repel the attack, take no aggressive action beyond what is required to come to a stalemate 3. Do nothing
I have argued elsewhere with you that I don't see 1. as being at all likely due to NATO's goals, defensive posture, and military superiority.
Doing nothing would only embolden Russia and destroy any pretense of an alliance at all. If not in Estonia, when would NATO intervene? Poland? Germany?
Thus, repel the attack and take no aggressive action. De-escalate and give options for Russia to save face.
We don't have to bet that a single person is not unhinged enough to press the button. There are command and control limitations purpose-built into even Russia's nuclear arsenal. Putin also wouldn't want his historical legacy ruined by the thorough destruction of the present and future of Russia. I'd go further to say that he knows all this, and wouldn't invade a NATO member in the first place.
I hope this is not how nuclear strategy and tactics is prepared. If sane people are in charge, retaliations will be proportional and exchange will be limited.
Russia obliterates Katowice, one of several countries obliterates Volgograd. What is the next move? You just ended the lives of a million of your citizens. You just stop there?