How great should the risk of dying be?

[AMFV]

I actually think we are agreeing here. Most of the time players go for greater risk, it is no fun killing goblins at level 20. But then again not everybody is playing D&D and some system are much more lethal and you don't have a bloated HP blanket that keeps you warm and safe.

To be fair, I wasn't discussing D&D, in particular, or in specific. Most of my discussion was aimed at games like Rolemaster or GURPS Fantasy, which have both complex character creation and high lethality. AD&D and older D&D which had higher lethality did not have complex character creation. It's a virtue in certain systems to have certain kinds of risk.

I think that part of the difficulty is that everybody treats risk a little differently, and so their perception of the risk will be different.

No but I was starting GMing for a new group and in my experience player perception of mortality is far more important than actual risk. So at the start I had established that characters can die and that makes the risk seem more real.

Not to me, it wouldn't. Now if it does to your players, that's one thing, and that's fine. But you can't assume that everybody's perceptions and reactions are the same. I know enough statistics and enough game theory to have a reasonable idea of when things are starting to get increasingly fudged in my favor. Not that I'm necessarily opposed to that. In a game where heroic fantasy is the goal, and focusing on my character's heroism is the objective, it's good to be able to act more heroically than I often would. Of course, that means that my character doesn't really get the whole benefits of the heroism though. It's an expectation, rather than something with real risk.

6 months actually, but then again I'm not trying to run an high lethality game, but the system we use is rather deadly so the risk is always there. But as I have mentioned earlier in the post player perception is more important than the actual statistical risk.

Well that depends on how astute your players are to the statistical perceptions. If your players are particularly savvy they'll even notice that you're pulling punches to avoid character death, since statistically it should be happening more often. Again, you're making broad strokes about general perception of all players based on a very small sample size.

Also, part of the virtue of high lethality games is that you as DM will sometimes be surprised by death. Sometimes characters can die in shocking and unexpected ways. It creates a very different tone and atmosphere. If I were playing a game set in a wartime setting, that's the kind of feeling I would want. The same holds true for Horror games. Although it doesn't work for everything.

After decades of playing different systems and settings with various degrees of lethality, I for one have come to the conclusion that high risk of dying has nothing to do with how exciting the game is.

Some have likened this to gambling where you risk losing money, if nothing is at stake the excitement is less. Yes could be if you are playing a combat simulation, where the narrative only exists to lead the party from one fight to another.

But if yo are playing a narrative or a character driven game then players are excited about what happens next...not about how their dice land in next combat

Well it's a different kind of excitement. I can be excited reading a novel or watching a movie. But if there's actual risk to me, then it's much more exciting. You can get some enjoyment out of a narrative like The Fast and the Furious, but it's not the same rush or excitement as you would get from actual drag racing. The stakes do matter. Of course, higher stakes also means that eventually worse things will happen.

Also, high lethality isn't necessarily opposed to narrative development. If anything high lethality inspires a degree of paranoia and nervousness that forces people to consider non-combat options. A game where there's a high risk of death will have less combat than a game where death is reasonably uncommon. Compare D&D 1st Edition to D&D Fourth Edition. OD&D was very very risky, particularly at low levels. Players actively tried to avoid fighting as much possible, sneaking or talking their way out whenever they could. Fourth Edition didn't have as much risk of death, and had many ways to avoid dying, it's basically a straight-up combat simulator.

Not that there's anything wrong with either course, naturally, but I think you're drawing correlations where none exist, and worse drawing some correlations in reverse of what actually tends to happen.