Second remark: I would love to see some percentages also on those scenarios, one that could lead to kaboom, one that could potentially lead to a de-escalation or a kaboom:

• What if the internal political dynamics could lead to a coup that will take the hardliners on power? That could be a likely scenario, that resembles the attempted August coup of 1991.
• What if another military defeat could lead to mass demonstration as Argentina losing in Falklands in 1982? Do you think that this would lead to Putin political end or to a bloody repression and a worsening in the war?

Another point that could potentially lead to kaboom:

• What if the west crosses one of the red lines placed by Putin? Because the actual kaboom scenario is based on the assumption that Russia uses the nuke on Ukraine, but what if the west will "escalate" with more powerful armaments? What are the probabilities for you that Russia would nuke (or use conventional weapons on) Poland distribution centers or NATO bases rather than Ukraine?   

To be honest I would like to understand a little bit more on why and how the parameters of this little model were "tuned". I mean, I am not saying that the model is wrong (because the outcomes are forcibly those) but I would like to understand more on how those probabilities were tuned or at least estimated.

Let's put in this way: to possibly have an hint to calculate and have a reasonable estimate I think we should consider the chicken game scenario. Because at the moment we are in a stage of the war that is a "chicken game", in which neither Ukraine, neither Russia wants to divert from the road that leads to the brink. In addition we should consider one thing (this is not a position that says that NATO is wrong or so, but is a matter of fact), both the opponents want to keep going in the war, based on game theory: if they will sit down to a peace agreement both will earn less than they can earn winning the war, and this scenario was possible only with the western armaments. At the beginning of the invasion, who was more determined for sitting down on a peace table was Ukraine, because Ukr could still take back some territories that were invaded with a negotiated peace agreement.

So my suggestion to try to understand the closeness of the probabilities to the reality is to understand in human behavior the "chicken game" outcomes in other situations, because the human reasoning probably is the same and to understand how much the mankind would go to the brink (or, in Italian "Muoia Sansone e tutti i filistei", from biblic tales of Sanson) avoiding to be a "chicken".

On the impact of WW3, I think that the scenario depends on how extended will be the war. If you consider the seminal studies of Turco and Sagan, I think that the nuclear winter would be far more dangerous, leading the human race to the brink of extinction.