[DrPraetorious]

I believe that you get about twice as many of the "very bad" events with misfortune-3 as with misfortune-2 - but that's just a guess, the thing I was checking was total money.

For whatever reason - lost income, killed population, unrest from barbarian invasions - you make less money on average (a few %, I did enough samples to get to statistical significance) as misfortune-3/order-3 than you do as misfortune-2/order-3/death-1, at least over the medium haul.

I just ran it again to make sure I had the magnitude right:
setting, pop after 10, money after 10,notes
o3d1m2, 29540, 3740
o3d1m2, 29670, 4585, militia
o3d1m2, 28090, 3892
o3d1m2, 29110, 4129

o3m3, 25780, 4164, no lab
o3m3, 24430, 4077, +death gems
o3m3, 29560, 3686, militia
o3m3, 29410, 4073

That's 345 gold less you get on average - or ~8%, which is a bit larger than the value I got before.

I did a larger sample at one point and got about the same result (but statistically significant, 95% confidence interval of 1-11%, IIRC, but can't find my notes). This isn't a perfect methodology (it was better before, had same armies for everyone) - once I have the kinks worked out I'll post something more comprehensive. I only used nations with temperate preference.

This is from setting your initial army on patrol and hosting for 10 turns with autotax. Without the patrollers the effect is presumably more pronounced, since most of the money lost to misfortune is from increased unrest.

This also ignores the semi-logarithmic effect of fortune/misfortune. The number of events you get doesn't scale linearly with empire size. On very large maps, misfortune-3 is therefore more worth it.

Anyway, the risk of losing your temple with misfortune-3 seems to be significant; with misfortune-2, negligible. If you're a high-roller and inclined to take risks, obviously misfortune 3 would be more appealing on that basis as well.