The Gamemaster, often called the GM, is the source of all your fun and all your sorrow in a role-playing game. An adept or experienced GM will make the worlds of the RPG come to life, and present a vision that makes you feel as if you're really there. In other cases, the GM can come off as a total dick. The GM is responsible for describing the game world, playing the role of its inhabitants (NPCs and monsters) and adjudicating the results of your actions.
A lot of people come asking for advice on how to run a role-playing game, but the simple truth of the matter is that a game master is not born; rather, made. Only experience, reading and knowing the group of people you play with will help you become really good. Different GMs have different approaches, some improvise everything, others painstakingly prepare every map, encounter and NPC the players come across. This way of doing things rarely pays off, as players usually hold to long-standing player customs of shrugging off all over your meticulously-planned work, as they decide to take one wrong turn or ignore one person that was supposed to put them on the right track you laid down for them, and wander off in the complete opposite direction. Some GMs counter this by railroading their players, which is generally seen as an douche way of doing things. When being railroaded, the players typically become little more than unwilling spectators to the GM's personal fantasy movie, which usually (read: always) sucks.
If you know what your players want you're one step closer to running a good game. Some just want to kick some goblin arse, others want to get involved in the political intrigue at court, others don't really know what they want. Try to lead them on adventures that involve all the characters and give them all challenges that depend on what they do best. If you have a rogue in the party make sure to have some sneaking or trap-finding to be done, if you have a barbarian be sure there will be opportunity to kick some ass and so on. Talk to your players.
No matter what your GM playstyle is or what your RPG is, there are a number of universal truths that help any GM run his game.
Rule 0:Have fun. That's why you and everyone else is playing. When the game stops being fun, that's a sure sign to take a break for a while and come back another day when your enthusiasm has built up again.
1. Talk to your players. Don't just "tell", otherwise your players can come to resent you as a tyrant or a railroader and be no fun to be around. Two-Way communication helps a lot even when you're in the middle of a long script of text such as clarifying points about the setting or the rules.
2. DON'T be That Guy. It's just so much worse when it is the GM. You are presumably hosting the days event, so be a gracious host and that'll encourage them to come back. (note that doesn't mean providing all the munchies and other entertainment, just be patient with your group)
3. Improvisation is a MAJOR tool in your arsenal, even if you are running a heavily scripted adventure. Since you know the ending to your story you can incorporate player decisions into the overall plot and keep the story moving no matter what they do, which is kind of the point. A good improv can stop the campaign from being derailed and keep the players entertained by the twisting of events.
4. NO is a word you will probably use often, so learn to say it in different ways or come up with alternatives. For example: Sometimes a player will come to you with an awesome character build but unfortunately it may break your game, so finding ways to negotiate him away from it into something more suitable becomes a necessity. That might require some give-and-take on your part to keep the enjoyment factor in, but remember ultimately it's your decision, so don't feel pressured.
Similarly No Beats Yes. This is a problem for very new GMs (and White Wolf storytellers). Every group will have rogue players that try to do some foolhearty or unwise action that the rest of the group disagrees with. As a GM, it's important to allow the players to come to a consensus about the party's actions before listing the horrible consequences of an impulsive player's choice. The players don't ever need to be at each other's throats or considering ejecting a player for their ADHD if you as a GM give them a timing window to talk about group actions.
5. Rolling the dice is an acceptable means of making shit up, even if you actually aren't making it up. Rolling the dice in front of the players (or even behind your GM screen) gives them the impression that a decision is out of your hands and that the events unfolding in front of them are completely even-handed and fair, even when you already knew what the outcome was going to be. Even killer GMs who throw 20 great wyrm red dragons at groups one after the other can generate less complaints if the players thought you rolled for them randomly.
There is an argument for the GM not fluffing dice rolls, because it can lead to a reduction in the impact the players have on the game by negating their mistakes or triumphs. While this argument is valid, and when done egregiously it can ruin player experiences and immersion. It ignores the fundamental position of the GM: they are not a "player" in the same way that everyone else at the table is, they don't build their characters and monsters using the same methods and are not restricted by available resources when preparing encounters. The GM is instead a facilitator for the progression of the narrative; even if that narrative is only "Dungeon: what do?" they still need to make sure the game still moves forward and doesn't get bogged down, that everyone at the table gets their moment to shine, and that the continuity of one session (whether railroaded or not) are carried on to the next session. While this is not a justification for griefing the players, sometimes you simply have to let an unfortunate dice roll slide for the sake of the game, your ability to adapt to unexpected rolls can be dependent on your GM style and the result does not need to necessarily be good or bad, whether your players come back regularly will be the main indication.
6. Work your players for their rewards. it's unlikely to be very fun if players find artefact-level gear all over the place / level up every hour / have the immediate and utter respect of every new NPC they meet (unless that is your particular thing). It's important to always have a carrot to dangle in front of the players and keep drawing them forward. When they catch it, give them another slightly shinier goal to chase, but keep it in small steps and don't overload them with rewards at the end of each adventure. This makes character growth memorable and makes the players more attached to their PCs in the long run.
7 Do the Research, when you use real-world or in-universe analogues that players might aleady be familiar with; such as military rankings, feudal titles, established figures or events (fictional or historical) or even technical wording of theories or philosophies. It's not hard, and you don't need a degree in the subject, but you'll look a fool when your players notice something wrong and could do a quick wikipedia search to point it out. GM fiat only gets you so far and you can start losing credibility if you expect players to buy into an anachronistic campaign setting without a suitable explanation. So basically: Know what you're talking about.
8 Learn Pacing, especially when writing your own material. It sounds obvious, but figure out the difference between Encounters and Adventures then make them discrete things and break them up by having not have the same things happen twice in a row. Fight after fight after fight gets boring, even for hardcore hack-and-slash players; while repetitive adventure themes end up resolving themselves predictably and players stop caring.
An encounter should take no more than a two or three hours, even if it's a combat when the game gets bogged down by turns and dice rolls, any longer and players lose interest. But don't extend them needlessly if the players seem to breeze through too quickly/easily. Adding "reinforcements" if the fight goes too quickly is just lazy, just move on to the next encounter.
Adventures should be a handful of encounters and take around two or three sessions, it should not be the entire campaign setting itself. That means inserting sensible "start" and "end" points with reasonable objectives. If a single adventure takes weeks of gameplay to progress through, your players will forget key facts and objectives as the story goes on and it becomes harder to keep the group on track without forcefully railroading them.
Campaigns can have goals, but are not necessary. If you plan for the party to kill the evil king, don't make that the group's immediate objective from the get go, otherwise you create one really long adventure, instead you should build your way to that episodically with a series of adventures.
9 Don't write the ending before the players get involved. One of the most important rules of war is that No Plan Survives Contact With The Enemy, meaning that people's actions are very difficult to predict with meaningful certianity, and they will very frequently try something you didn't think of, so Not As Planned occurs exponentially more often than Just As Planned. The same is true with GMing.
In the end there is really only one rule, Rule 0, which states: Have fun. Meaning everyone at the table. Make sure everything is moving forward, try to avoid stalling and monotony. If the players are really stuck just throw something at them, even ninjas. Keep things happening and everyone interested.
If people aren't excited or interested it's often better to pull out another game, switch GMs, watch a movie or just WATCH ALL THE PORN (or play FATAL).
...or it could be endless optimisers who constantly want to reroll their characters every time a new splatbook comes out, then get all moody when you don't let them play the character they want to play.
...or it could be the uber-hardcore roleplayers that get extremely loud/overbearing at the table or charge headfirst into trouble with a battlecry "LEEEEEEEEEROOOY" "because that's what their character would do" and make life difficult for other players.
...or it could be the lolrandom Chaotic Stupid PC that rolls dice to decide how their character acts and tries to do standup instead of playing the game.
There is one very handy piece of advice to remember: "The game will continue without them".
That is not actually a threat (though it can be used as such), it's more a reminder to yourself that no matter how bad a player attempts to screw up your game, your game will survive as long as you want it to.
Overpowered PCs compared to the rest of the party can generally be ignored or tarpitted while the rest of the party gets on with the actual campaign / abusersof numerical advantages can be comfortably "noped" behind your GM screen during stat contests / overbearing roleplayers can be executed or put in jail on the spot for stupidity.
The only thing that can actually ruin your game outright is a TPK or you just scream "fuck it" and give up, which you as the GM are always the final arbiter of. If this one player has created a situation where they somehow managed to survive it all, then jolly for them... the game is still over and all of their effort has come to nothing.
As a subset of "difficult" players; there can be unseen difficulties when new players join your group, or return after an extended period of time and expect to play the same character. Furthermore you can often find players (usually Powergamers) who contrive of flimsy excuses to kill/retire their character in order to roll up a new build that they like, essentially introducing a "new" character to the game which is already ongoing. There is nothing inherently wrong with any of this, but it can introduce a few problems to the table:
One problem is the plot rationale for the introduction, particularly at high levels. If the players have been playing for a long time and have established themselves as influential figures, then why is the new guy just as good as (or better than) they are? without having done any of the legwork they they have? and why hasn't he shown his face before now? This is most often covered in Dungeon Master Guides and rulebooks, it generally needs some degree of creativity to make sense of the new introduction and some give-and-take on behalf of the whole group to accommodate the new arrival. The new player/character will be left out of a lot of the group history and the in-jokes, and presumably won't have any of the unique artefacts or unique campaign options that can only be achieved through actual time spent at the table (quite a lot of Prestige Class prerequisites come to mind).
The second issue is the sudden jump of experience their "new" character receive, as any player presented with a rulebook and a specified sum of experience and cash will cherry pick their favourite options and powergame with the system, even if it wasn't their intention in the first place. This is less apparent with groups that haven't been playing long, or with particularly small groups who really need all the help they can get. But in large groups or groups with a lot of history; the introduction of a freshly minted Space Marine or a level 18 CoDzilla that never existed before can really upset the group dynamics. At the start of a campaign, players will write their back stories, then branch their characters as the story progresses, picking up feats, talents or skills based on the circumstances of the plot, such as the ability to swim, drive, speak new languages or have fighters being able to cast low level cantrips, or wizards multi-classing as rogues out of necessity when the group changes or the plot moves forward. Contrast with freshly joined players who will usually build the character first then write the story for that build, utterly neglecting the sacrifices other players needed to make to get to the same point.
While RPG groups are not necessarily fragile ecosystems and players do their best to get on with what their characters do best, some players can come to resent "the new guy", even if the actual player has been with the group from the start. This goes double in rulesets which have persistent meta-characteristic based on infamy, status, honor, or include persistent injuries. Grizzled adventurers with psychological problems, troublesome injuries or particular levels of corruption can find it difficult to keep up to a Paladin fresh out of the packet being played with a lot of enthusiasm.
There are a number of schools of thought on the matter that help resolve this:
Have the player start out at level 1: This allows the player to create a fresh new character and see them grow as the campaign develops, as well as allowing them the opportunity to change their mind over their build when the story moves forward. It can be a harsh option that is only really suitable for low-mid level groups. Otherwise it can make new players feel useless or superfluous when there are large level gaps and big differences in ability. However because the distance between levels usually increases in most RPG rulesets, the new player will catch up relatively quickly often jumping whole levels after a few encounters, but will usually remain behind the rest of the group, which at least recognises the work the existing players have put into the campaign. It should also be the go-to option for brand-new RPG players who don't know what they are doing, since the paperwork of playing a high level character can be daunting for those not familiar with the rules.
Hand in hand with this option is the ability to run separate side-quests for the new player(s) in question. Giving them their own plot advancement that does not unfairly elevate their power level without the benefit of actually earning the experience. This often gives a satisfying experience for everyone involved without complaints.
Have the player start out at the equivalent to the lowest level player: This is perhaps one of the fairest options, especially if members of the group are itinerant and only show up in dribs and drabs between sessions. It provides a basic entry level for the player to meet that is not too far behind the rest of the group with the opportunity to catch up and possibly overtake if they become regulars at the table.
Have the player start out equivalent to the rest of the group, with caveats: Some parties allow a new character to join at an equal level but impose certain conditions on the player's character. Usually providing direct input on the characters backstory and build. This is often based on the party's needs, so if they are missing a key element such as a Rogue or a Cleric, then the new guy is guided down a path set for them, or required to take certain options that are necessary based on the way the campaign has developed. This might wreck a powergamers lv1-lv20 netlist build, but considering that the experience is free when everyone else already did the work, they can't really complain if it allows the GM to more easily slot that character into the world without upsetting the balance of the party and the setting.
There should be no substitute for actual time spent at the table. In all situations: If a character build is dependent on certain external factors such as permissions from in-game entities like commanders or royals or is dependent upon the intervention of divine figures, then they probably shouldn't be hand-waved just so the new player can get his build unless you were also willing to hand-wave these for the existing players who have been there longer. If one player has to play through several sessions of his own specific character arc just to get access to certain options or abilities then you should firmly insist that the new/returning player stop trying to argue that his character spent his downtime doing cooler things than the players.
And this elegan/tg/entlemen is why we never let That Guy GM.
Shut up. Just shut the hell up. You know what? You're dead. You died. Orcus reached through a tear in the abyss and pulped you like an orange. Now get ready to roll a new character, we're playing my erotically chargedMy Little Ponyhomebrew sys-*BLAM*Oh, don't mind me, I just crawled out of the 40K Section.Praise the Emprah.
A description of commonly found play styles of GM and their pros & cons. (hoping that more will be added over time) Note that it's easily possible to be multiple types at once.
Aspiring Author - Hand-crafts his own campaign setting, populating it with "unique" characters, factions and history. The campaign will often grow with the players, taking shape based on the things that they do which can be extremely satisfying.
Pros - Can make for a very unique experience if the GM takes his time with the setting, particularly in creating unexpected/memorable scenarios for the players. Also becomes very difficult to meta-game since players don't necessarily know how the system works (and therefore how to break it)
Cons - If he's not very good, the "uniqueness" of the setting will be contrived / cliche tropes that the players already know, and will get tired of if being sold to them as "different". Also if the GM is not committed to the group, adventures may take a long while to write up and cause the players to forget what they were even playing. Furthermore, can be overprotective of his setting (particularly if he IS an author) and refuse to budge when it comes to harsh decisions or situations regarding the greater universe, though he is GM so fair play to him for that. But be aware that just because someone is excellent at narrative doesn't always make for a good GM; some encounters may drag out for "impact" or "effect" but don't always translate to good flow for a group of players, leaving them bored. Or they mismatch the power level of the opponents to the party because "that's what they would be in-universe".
Canon Defender / Fanboy - Applicable when using an established setting/adventure modules (Forgotten Realms, 40k, Star Wars etc) they usually know their fluff better than most and try to maintain the integrity as much as they possibly can by restricting how much damage the players will inevitably attempt to cause by breaking it. Or by restricting their movements to specific "breakable" portions of the in-game universe where the players cannot cause significant harm. (eg: no visiting Terra since the players will obviously attempt to murder the Emperor)
Pros - The "realism" of the campaign is maintained, so if the players know their lore they remain familiar with the setting no matter how much they try to screw it over. By necessity, he is heavily invested in the setting, meaning that he will be reliably consistent with the players and one of the most enthusiastic of GM types. So if the GM is well versed enough, the session can be just as immersive as those run by The Actor, just keeping a status quo that won't come crashing down around the group.
Cons - Players can feel cheated that they are not interacting with the setting as much as they would like. Meta-gaming is going to happen and arguments will occur with players who think they know the setting better. Also, if a GM is not well versed enough with the setting it WILL devolve into railroading as he won't know how to react to unexpected situations.
Most likely campaign ending - Whenever the owner of the setting/universe stops publishing material.
Formula/Dice GM - What the rulebooks often expect a GM to be. Follows the adventure modules religiously, but unlike the canon defender is more interested in the crunch than the fluff. When shit happens, he will likely refer to whatever table of events / random encounter lists / pre-generated characters that are available. He is in it just as much as the players are; just along for the ride.
Pros - Things tend to go most according to plan since a rule can often be found to cover most situations. Is also the least likely to cause a fuss with the players, since he's playing the game as much as they are. So nothing is personal with him.
Cons - The situations can occur that often feel out of place to the players; such as repetitive random encounters or things not happening in the appropriate locations due to the result of a dice roll. (meeting a merchant in the middle of a dungeon?) though if the players are light-hearted enough this can be source of amusement and amusing material that may be worth a chuckle long after the game is over. Also, he needs to know the rules he's deferring to backwards and forwards, or the game will move at a snail's pace.
Most likely campaign ending - Whenever the RPG publisher stops publishing material
Favored Games: - almost stereotypically Pathfinder or D&D
Turncoat Player - Longtime players (most often Powergamers) who want to give the GM chair a shot, often thinking they can do it better than the previous GM since they have sat through the experience of being a player and want to do it differently. Often with a list of things they want to change like an election manifesto.
Pros - If they can learn restraint, they can be the most sympathetic to the players needs and become one of most adaptable GMs. Also will spend a lot of time with the players developing their characters, crafting well fleshed out back stories and often including an inordinate amount of heirloom items. Additionally if they were a that guy player there's a good chance they'll be more sympathetic to the previous GM and any future ones after walking a mile in their shoes.
Cons - More player experience does not always equal good GM. These GMs can be wishlisters; creating settings that they'd rather play in. Giving the players the "best" gear, "freedom" of gameplay, "unrestricted" access to character options (all words which should give proper GMs headaches) and dish out ridiculous experience and in-situ rewards in an attempt to ingratiate himself with the group and sometimes end up in denial that their sessions have any flaws in them at all. Also, since his campaigns are heavily player focussed, groups may find that encounters become less of a challenge and/or that the universe really is quite hollow and not populated with very interesting NPCs. These campaigns generally have short lifespans, mostly because the players accelerate faster than his ability to come up with new material, or the turncoat gets bored of running the campaign or wants to try a "better" ruleset that requires starting all over again, where he's in his element. Additionally as a lesser con, since he is a former player he may be used to thinking in terms of 'optimization' so he may have his NPC's in combat use every trick in the PC's book, such knowing when to focus targets or optimizing action efficiency and so on. Even if he is just doing it because he's used to thinking in that mindset; this can unintentionally make his games much harder even he's not intentionally trying to be a killer GM.
Most likely campaign ending - Whenever the PCs have been prematurely promoted to Generals, Gods or any other point where it makes no sense to continue.
Alternate Ending - Turncoat players (especially powergamers) can go OCD with their campaign, in that he spends too much time crafting NPCs and items, treating them as if they were his very own PCs, then to take things personally when the players outshine them or defeat them. At this point the Turncoat will likely morph into a railroader where they party gets to sit and watch his NPCs interact with each other.
Favored Games: - Whatever the group was already playing
Pros - Loot. Fun. Good. The 'Looter Shooter' style of video game for a reason and entire genre of games exist around the cycle of 'do quest to find stuff to do quests with'. Transplanting that game play to a table top is a natural idea if done correctly.
Cons - It's easy for a GM like this to go overboard and turn the entire party into a band of Mary Sues, some even going so far as to invent items like a sword of instant monster slaying or an infinite-range missile launcher just to have some form of loot to give the players that they don't already have.
Most likely campaign ending - When the PCs begin weeping over the fact that the world has no more items left to loot, ala Alexander the Great.
Favored Games: - D&D, Pathfinder, Dark Heresy. Anything that can have large amount of customized loot to give out.
The Railroader - Often appears to have the most desirable skillset as a GM; knows his fluff, the ruleset and gets on well with the group but do not be deceived, he is one of the worst types of style. Recognisable for his lack of dice rolling, also for constantly talking to his players and knowing what happens next without referring to books/notes. Often giving them loads of information to work with and dropping hints about what they could be doing next. If he's good, he will often defer to the players and create a scenario on the fly depending upon what their options are.
Pros - One of the smoothest operating GMs, since his plan is constantly in his mind he will quickly have a resolution to most scenarios. Also can be as creative as the aspiring author above, and when telling a good story and is in his interactive element, the players may not ever realise that they have been railroaded at all - which is the sign of a REALLY good GM.
Cons - The players will feel railroaded by his constant hint dropping, and will be punished (sometimes severely) with whatever consequence he has in his twisted mind for not taking those hints. Since there is little to no dice rolling there is often no comeback for the players. This GM will insist (and often genuinely believe) he is being fair and is NOT railroading you since you always had the "option" of following his advice. Also, since it's all mostly in his mind, if he gets an opinion or a vendetta in there against you, you're pretty much screwed over at this point and are just pawns in his little game that he's playing with himself. Additionally if he can't think fast on his feet, going off the rails may flummox him even if he tries to play along to the party's wishes. Still, if he calls a break to figure out where the hell things are going, at least he's trying, so cut him some slack. Also a combination with the Fanboy can occur where the Railroader puts the players on the side of a major conflict in the setting without actually allowing players to influence it meaningfully.
Most likely campaign ending - Rocks fall, everyone dies because you went down a corridor the GM told you not to go down.
The Actor - Basically what the Railroader would be if he weren't in touch with his inner control freak. Like the Author and Railroader, he really wants to tell a story, but like the Turncoat he wants it to be the players'. Generally more interested in the role part of roleplaying, he tends to put lots of effort into making colorful side characters and setting flavor, but tends to regard combat as a side dish rather than the meat and bones of the game, so he frequently improvises and works off a set of indistinct guidelines more than rigid pre-prepared content.
Pros - A very flexible kind of GM. Since he isn't married to a preconceived notion of how the session is going to go, he isn't going to be completely floored unless you go full Henderson. The effort he puts into making the world feel alive can be very immersive, and he tries to make events feel like they matter to the characters instead of just being a story they're blundering through. Prefers to keep the game moving, so he's likely to make a judgement call based on the roll rather than look up the exact rule if he doesn't know it. If you tend to regard constant pointless action scenes and random encounter fights as annoying filler where nothing is happening to advance the story, you and he are going to get along like white rice and soy sauce.
Cons - Very prone to taste mismatch. Since he's an actor first, he tends to skimp on combat and hack-and-slash unless it's important to the story. Don't expect lots of random encounters, and don't expect constant life-or-death struggle with everything trying to kill you. Frequently has problems with managing lots of simulationist rules, and often jury-rigs solutions out of GM Fiat rather than looking up the answers in the book which can be a real problem if he's not consistent about it. If you regard the "standing around talking" part of the game as pointless nonsense fluff between the action and fights, you and he are going to get along like sodium and water.
Most likely campaign ending - Whenever the GM moves house away from your city.
Submissive / Reactionary GM - Not an oxymoron, but rather the polar opposite of the Railroader. Bends over backwards to accommodate the players and their characters, but unlike the Turncoat actually knows how to GM properly by maintaining an interesting storyline and also how to reward the players correctly/fairly for what they do. This style requires a tremendous amount of creativity on the part of the GM even though they can stick to a script / published adventure module. They tend to occur with campaigns involving evil PCs or with novice GMs among more veteran groups where players are given more latitude with their conduct either by design or through the GM's lack of personality / assertiveness.
Pros - If the GM is any good then players can get the full package, the freedom to do whatever they want with the in-game universe and it still remain a challenging and enjoyable experience for them.
Cons - Sub GMs tend to have a short career, either by giving up or switching to a different style as constantly having to come up with off-script consequences to outrageous player behaviour can burn them out creatively. Or due to a lack of enforced discipline/conduct, player groups become bogged down by conflicting personalities and the group loses its cohesion.
Most likely campaign ending - When the players have dueled each other to the death over the staff of godhood and there is only one left standing. (or the campaign gets forgotten about and shelved)
The ComedianFile:Joke-GM.gif- A GM that's not afraid to drop gags and jokes into his story. For this GM, Role playing is an activity to be enjoyed, and by slipping the occasional gag into the world in an attempted to get a chuckle from the players. Expect to slay Vampire cults inspired by Twilight, Dwarfs who act like Dorfs, Have an Octopus disgused as a human however nobody notices thanks to an insane bluff score, ectra.
Pros - Comedians are not above running silly, funny and awesome concepts like the Deffwotch and when everything is running well and the players and the GM click they can provide very entertaining games.
Cons - The players and the GM have to be on the same wavelength. Humor is ultimately subjective and a gag that the GM finds funny may be annoying or infuriating to the other players. Further the number of gags, in jokes, and shout outs weaken the overall structure of the game world and makes immersion harder.
Most likely campaign ending - "The players have saved the Dragon from the princess, have been found guilty of murder, looting, and jay walking, and an ork was the prosecutor, they then run as fast as they can out of the kingdom and go into hiding. The End?"
Favored Games: - Paranoia, HōL, or any "serious" game, like the Warhammer 40,000 Roleplay series, White Wolf games, D&D, where the humor can take on a tone of parody of the normal sort of games those systems/settings expect. A good Comedian can make just about any game into a laugh riot. Perhaps even FATAL*BLAM*EXTRA HERESY!
Killer GMFile:Iamthegamemaster.jpg - This GM very much believes that their job is adversarial. They're here to fuck you up through any means necessary, either through crunch or fluff. Expect save-or-die traps, Level 20 town guards, monsters out of nowhere, and Tomb of Horrors style dungeons. Very little or even non-existent storyline, just challenge after challenge, essentially creating a P&P Dark Souls. Killer GM are also unique in that they are often a secondary role, difficult is after all easiest part of a story to change so any of the above roles can also run a killer GM game. (Reportedly Gary Gygax himself was one of these, which might explain the nature of Old School Roleplaying. many those how were their at dawn of Roleplaying games tend to be the killer as the game was a evolution of war games where PC lives were thought as cheap even to those players, with the point of games was beating the meatgrinder and most long term campaign long character development was enjoying the fact you were able to keep your ant alive long enough that they get god-like powers as a reward.) (as roleplay became more important, the Killer GM is either an old-schooler or old-schooler fan, or a dick). Also if you want to run these kinds of games: for the love of god have a plan B for if your players even as optimized as they are just roll a one and die. Restarting at level 1 not only sucks for them it suck for the st of the party as they now lack a major amount of firepower.
Pros - You want a mechanical challenge? This is the GM for you. The perfect opponent for munchkins and murderhobos, the good kind of Killer GM excels at creating harsh, cruel, punishing, though not impossible, scenarios. If all the players just want to kill stuff and feel smart for avoiding traps and generally enacting badass death-defying stunts, the Killer GM can provide.
Cons - Just as with the Comedian here above, players and GM must be on the same wavelength and accept that things can go utterly, awfully, hilariously right or wrong with little forewarning, sometimes on a single die roll. And the GM must be impartial, or even a little in the party favor if they make an honest mistake in the knife edge balance a killer game requires. A bad Killer GM is just a massive twat with a huge hate-boner for you, your character, and everything you stand for. Hopefully not going as far to just kill you with no saves or warning, but the sort of GM to throw Lvl 20 monsters at a Lvl 1 party and then wonder why you didn't min-max enough to beat them. Like a Railroad GM, except the only destination is your frustrating, inexplicable death. Railroaders are also infamous for metamorphosing into Killer GM's in order to punish players who stray from there*BLAM**BLAM**BLAM* their story.
Most likely campaign ending - Total Party Kill either by defiant, awesome last stand worthy of posting on /tg/, or getting so fucked over you wonder why you even bother playing these stupid games in the first place, which can also be worthy of posting on /tg/.
/d/M - No way we could keep all the horrors of the /d/M on everything but it's own page, sanctioned and hermetically sealed, so its wrongness may never ever spill out into the world. Long story short: Go here.
