It was the most memorable thing that happened in the entire campaign – a moment that other players in that group would inevitably bring up when we discussed it, and it stands as one my personal “all-time-best” moments. A spectacular, single round of shooting, that dealt 206 damage to the NPC in a single full attack action, at level 3. There were no boots left to smoke on the floor, it was character defining.

And then he brought him back in the next scene, completely unharmed, and I was pissed. I’m still a little bitter about it, but that was the past. I have similar reasons for disliking GMs who save my character from a death that should have happened – it just really cheapens the experience for me to know that I can’t make a heroic last stand, because I’ll always wake up captured shortly thereafter.

I’ve decided that at my table GM dice are rolled in the open, and the only one responsible for what happens to players is themselves for not running when things turned south. Somewhat harsh, but no players have died in my campaign for a while yet because they know the stakes ahead of time.

Now when I ran D&D 3.5 I fudged when the players didn’t do something that was completely brainless (which did happen a few times). Fortune favors the quick-witted. Then again, I also let them do pretty much anything, if they could say it.

Here’s why:

The act of fudging is an expression of GM fiat, a way of making players into passive witnesses rather than active participants who choose their fate. In a lot of the older RPGs, where there wasn’t much complexity in the tactical and strategic choices in player creation or in opposed challenges, GMs often had to “read the tea leaves” so to speak, to either add or cope with complexities offered by the players. As such, GM fiat was baked into their systems. Players and GM would negotiate what happens next, and the result was decided by the GM, though players could “suggest” an outcome. Of course it was personal – because everything was decided by the GM.

On the other hand, the newer games tend to pay more attention to choices and consequences, and those “suggestions” are now planted firmly in the players’ control. For the most part, once combat is engaged, the “game” is very much the same between the players and the GM. With a balanced encounter, the GM can now afford to take less of an interventionist role in the action – because it really is less personal. Yes, the GM sets up the encounter, but once engaged – the GM should play it out.

When a GM exceeds his authority at the table, and cheats to let someone live who should have died, or cheats because there is a danger of losing a character – he imposes his own ending on the scenario he set up. It’s personal because there can be no expectation of consistency. How do you explain why you cheated to save one player character, but not someone else’s? In striving to avoid a personal grudge, you’ve sewn the seeds for confrontation yourself.

In addition to that, once players catch on to the safety net, it saps the game of tension, and over the long run, player apathy sets in. There’s nothing at stake – there’s no GAME – just story. Players need to view the GM as a fair gamer. As a GM, I make no bones about being fully sympathetic to the players. I want them to succeed as much as they do, and when their characters die, or when something unfortunate happens, that’s part of a game, and it’s part of good drama too. There have been a few times where I was as desparate as the players to pull a losing combat out of the fire – threatening a total party kill – and, by golly, they did it. It was a geeky game rush for everyone there.

If something does go off the rails and I do need to “fudge” something, I stick to what I control. My tactics. My environment. I don’t overwrite the outcome any more than I would tell the players what to do. Maybe a third party enters the fray, or some other random and distracting element. Maybe the bad guys do an inexplicably bad tactical choice (like players don’t?) In such a scenario, it’s just a nudge – enough for a party to disengage, or to scrape by.

My answer is: Yes, if necessary. This is why: I’ve had to pull a few punches over the years when my dice hit an infernal hot streak. I tend to have uncanny luck on the dice going from really poor streaks to incredible prowess runs that boggle the mind. I can handle these kinds of streaks as a player, but what happens to my games if they happen as a GM? I can ruin an entire campaign with one night of natural 20s dishing out mayhem on the PCs. So then, what can I do about it? How about making adjustments to my damage rolls? Is that too far out of bounds? What about reducing the third critical on a PC to a normal hit dishing out a low damage total? I think these things are reasonable actions on the part of the GM in an effort to keep lady luck from wrecking the campaign.

Another fudging might happen if I blew it when I was designing an encounter. If I feel that I blew it with the difficulty of an encounter, then I am likely to fudge things a little bit so as to make things more equitable for the PCs. This is particularly true when I suddenly realize that a critter has some devastating attack that I had forgotten about. Why punish the PCs for my design shortcomings? Thankfully, this has become more infrequent thanks to decades of experience. I have also come to the realization that it is better to prepare an encounter with too little challenge and beef it up on the fly, than to have it be too difficult and try and figure out how to ease the challenge without being silly.

I really try to let the chips fall where they may, but sometimes a little fudge is needed. Just remember that too much fudge will make everyone at the table sick.

IIRC, this rule was meant to make up for the fact that, in the beginning, the setting had no clerics and no resurrection. But there’s no reason you couldn’t use it in the early phases of any campaign, when those services are unavailable or unaffordable.

That being said, if you don’t let the PCs kill the bad guy(s) sooner or later – or if the bad guys don’t stay dead – then their players are going to get frustrated.

1) Nothing says “I love you” like Ressurection.
2) If you don’t leave a body, you can’t prove their dead. Cliffs, big explosions (“There’s no way he could have survived that!”), etc. give you a chance to step out of the mechanics of mortality and bring back the (probably heavily-patched together) bad guy or even PC. As an example, I once had a PC killed by a wandering twisted archon on Mechanus, plane of absolute Law. His death would have been a major blow, as he was a prince whom the campaign involved heavily and the party had no means for ressurection. So he got knocked off a sheer cliff with the death blow, found by modrons (you know, the cute little round mechanical critters) and they rebuilt him, made him faster, stronger, better (than the corpse he nearly was). He became the six-million gold-piece prince! Aside from developing a fear of magnets, and an addiction to lubricating oil, we went on with the story with a chuckle and yet another gaming story for posterity.
3) There’s always a member of the family who could fill their kin’s shoes (see: Green Goblin/ Harry Osbourne) to plague the PCs.

These are only a few options, I’m sure any clever DM could think up plenty more for his/her given situation.

You don’t need to be able to fudge the savings throws of recurring villains to preserve the plot in 4e. Just give them a suitably cinematic at-zero-hitpoints power (like the Tarrasque) to escape with.

It’s written down. It’s a rule. Now it’s not cheating.

Given that, I’ll ride the fence and say “it depends.”

If you’re playing a gritty game, or “just for fun” where the players aren’t too deeply invested in their characters and/or the plot, then I say play ‘em as they land.

But those aren’t usually the games I play in. My regular groups consist of deep story and lots of role-play, where the players have spent good effort fleshing out backgrounds and the GMs have spent even more time integrating those backgrounds into the plot. So having a key NPC or PC die at the “wrong time” can really screw things up — wrecking the fun.

That being said, I’m much more likely, as GM, to fudge to save a PC than to save an NPC and/or a plot point. So the fighter got a lucky crit and smacked the BBEG to oblivion? So the diplomat PC got an auto-success roll on her conversation check and thus the double-agent NPC has to spill his guts — including that key secret that wasn’t supposed to come out until much later? Too bad! That’s why it’s harder (IMHO) to be a GM than a player! You have to think on your feet and adjust.

Maybe the BBEG isn’t REALLY the BBEG, but rather just a front for the true threat. Maybe the double agent is actually (purposefully!) misinformed, and the real secret is much more convoluted than it seems. The trick is being creative and spontaneous enough that you can adjust the story and/or setting for unforeseen events.

P.S.: In one of the games I am a player in, we’re doing heavy role-play and deep plot, yet the GM insists on rolling in the free-and-clear as well as playing with occasional “save or die” mechanics. I’ve discussed it with him offline and he insists that without the fear of death that players won’t play properly — they’ll be “too brave” because they know they can’t die. I found this a bit insulting, but couldn’t argue that softening the fear of death would likely color some decision making.

Most of the time, I’d rather negotiate a bad roll at the table– often a player will come up with a solution that’s worse for the PC, but keeps their character playable.

In games like Primetime Adventures, a big part of the trick is making sure both success and failure are interesting. If your character is cool whatever the dice says, then I’m very happy letting the dice have their input. If the system isn’t good that way, I’m interested in seeing if we can make it so– via house rules, add-ons like the Temptation Dice Martin mentions below, or anything else.

It comes down to trusting that your GM is doing what is necessary to keep the game fun and exciting.