The Importance of Randomness

[Tortoise]

It's not simply a question of random encounters, though. Take combat, for example. What should be the proper balance between player choices and randomness in determining the outcome of a combat? I'm sure that preferences would run the gamut between the completely random coin toss (or die roll, if you don't want an exactly 50-50 chance of winning or losing) and the almost completely choice-driven chess game. I'm more inclined toward the chess game end of the spectrum myself (because I believe that good player choices should almost certainly guarantee success. Almost ), but I can see how more randomness might appeal to others.

While swingyness from random outcomes is always a factor, it is less so in some editions than others. That is a matter of choice for players and DM when determining what edition to play and what style of game they prefer.

I would never go for the coin toss method for determination since I like player choices to have an effect, but I also do not go for the chess style either (despite enjoying a challenging game of chess occasionally) in D&D. For me the level of randomness that I enjoy has to be risky, but influencable. Good decisions and planning by players can turn a bad situation into a major victory and likewise very bad choices combined with some bad breaks on the dice can make a good situation very ugly. (Players should always have their characters be willing to run away. Not all things they meet have to be beatable.)

Since my current campaign style doesn't have major plot arcs tied to specific characters it doesn't harm the campaign as a whole when a character dies. Compare that to my Ptolus campaign a few years ago where some plotlines would suffer from such a thing. It comes down to the style of campaign being played at the time.

 

[Tortoise]

Let's not pull out the troll card just yet. I feel fairly confident he was not merely trolling, he was making valid points even if they weren't specifically congruent with the perspective of "reader as wandering monster" as you meant it in your previous post. It's often best to assume that someone has simply misunderstood rather than writing them off as blatantly antagonistic. We're on the internetz, and nobody knows what anyone else is really talking about we just keep posting as if we do.

I feel it worth mentioning that if a table is involved, then it's only psuedo-random at best. At any rate, I do make some use of random encounter tables when I feel it is appropriate. I might introduce a bizarre encounter because of the challenge it offers me as a DM and to the players. Creating a reason for the encounter on the fly can be a ton of fun all by itself, but the potential new possibilities it can open up are usually worth it (unless I really muck it up). Generally though, I try to keep the encounters consistent with the current environment because over-use of out-of-place encounters can easily weaken the story and leaves a bad gamey aftertaste. If you rely too heavily on encounter tables you are creating a whole other level of predictability.

Essenti you make a good point and if I was being rash then please accept my apology. It is simply coming from a background where I have dealt with deliberate point dodgers a bit too frequently not to take that style of post that way.

When DMing I will often create my own random encounter/event tables to help flavor an area the way I think it should represent itself. I'm also not against using presented tables from a module or the rule book to give me ideas. A DM that just uses everything at face value isn't using their noggin to run a good game for their players. Thus my example about the two tables for a choice giving me the werewolf or halfling options and me deciding combat wasn't the purpose of the roll. Just because I decided on halflings instead of werewolf didn't mean the halflings would be attacking the party. A creative DM call based on what the random roll offered up.

 

[Harlekin]

I don't think the difference is even remotely as big as you are making it out to be.

The Death of Randomness

If there's anything that can be said for sure about 4e, it's that random events are constrained to the combat board, and even then, the range of random events is highly reduced. A creature has an expected lifespan, an expected damage output, and an expected treasure parcel.

This is true for any edition of D&D ever. Lifespan of a creature in combat and damage output are random variables and have therefore expectations. And IIRC, AD&D monsters not only had treasure types, there was even a table in the DMG that provided the expected yield from a given treasure.
The only difference is that modern editions, especially D&D 4th actually consider these variables in their design rather than leaving game masters to figure it out by themselves.

Encounters are built to minimize the likelihood of a disaster or overwhelming victory,

Again something that existed in all editions. You yourself say below that you balance encounters with more powerful foes by giving players the opportunity to avoid them.

<Some Snippage>

For a certain play style of DnD, this is a good thing. Knowing what's in the box lets DM's build encounters and whole campaigns with a reasonable certainty and a minimum of chaos.

But for other styles of DnD, randomness is absolutely essential. Not just in combat, but in the whole of the world. Randomness frees the Dungeon Master to be an impartial referee. Take wandering monsters, for example. If you decide that in a dungeon there's a 20% chance of encountering a monster every 10 minutes (or hour, or day), you have set a danger level without coming off as a bad guy. Of course, it requires skill to set these danger levels at something the players feel is logical. Walking through a town shouldn't have an 80% chance of being attacked, and walking through sparsely populated woods shouldn't have the checks made every five minutes. When the DM uses random methods to propel the 'story', it frees the DM to describe the situation and watch what happens, and it frees the players from having to second-guess the DM's intentions.

You really think there is any difference whether a GM plans 5 challenging encounters or he plans 10 challenging encounters, but rolls which of the 10 happen? I as a player would assign the same level of agenda to the GM. As the GM can also roughly expect how long it will take the players to explore a certain dungeon, he can even predict how often he will roll on that table.

Random events undermine metagaming. Metagaming requires treating the game world as a statistical analysis, or treating the DM's adventure as a predictable narrative. Humans can only make statistical predictions within very tight constraints, however. With multiple variables, predictions outside of the system become impossible. Instead of wondering why the DM isn't letting my amazing tracker shine, I know that within this town there's a chance that I simply won't encounter any badguys worth tracking. It stops becoming about an antagonism between the DM and the players and now the players have to deal with a new situation.

Unpredictability spurs creativity. If the DM doesn't know what is going to happen from moment to moment, plot creation involves finding links between events, strengthening some, pruning others and letting the players move through the world freely. It frees the DM to keep low investment in any one outcome, because a trick of the dice might undo all her hard work.

Again, the DM decides what's on the random encounter table. If only 5% of all encounters include anything that can be tracked, the GM is being a dick to the amazing tracker, if 30% of all encounters are trackable, it doesn't matter if the DM rolls on a table or assigns them to a specific location.

To summarize, it may be nice for the DM, if he cannot always know what the next encounter will be, but he looses the abillity to have encounters build on each other. From the perspective of the players it makes very little difference. The GM still controls pretty precisely how many encounters and of what difficulty the PCs will encounter. If you add that the GM can arbitrarily modify the setting of the encounter, for example by turning an encounter into a simple sighting of a monster (as you describe below), I think the differences are pretty trivial between 1) planning encounters in a given order and 2) planning encounters and then randomly rolling in which order they happen .

 

[Lanefan]

Randomness makes things less predictable, and as a lover of chaos this is just fine with me!

That said, a DM has to know when to go with the dice and when not to; and that's something often only learned by trial and error.

Lan-"chaotic neutral except chaotic good on Tuesdays"-efan

 

[Essenti]

...
When DMing I will often create my own random encounter/event tables to help flavor an area the way I think it should represent itself. I'm also not against using presented tables from a module or the rule book to give me ideas. A DM that just uses everything at face value isn't using their noggin to run a good game for their players. Thus my example about the two tables for a choice giving me the werewolf or halfling options and me deciding combat wasn't the purpose of the roll. Just because I decided on halflings instead of werewolf didn't mean the halflings would be attacking the party. A creative DM call based on what the random roll offered up.

I might roll on a table and get a result I don't have a clue how to integrate or just don't like, and I have no qualms throwing the dice again. But I'm of the school of thought that the dice are only suggesting a possible outcome, it's my job to construct tension and situations for the players to work through, and I don't always rely on dice to tell me how to do that.

I know there are plenty of people who don't like DM fudging about with the dice, but I have never had any complaints. Then again, I make a lot of rolls behind my screen, many times I'm just rolling to keep the players guessing rather than having any mechanical reason to do so If the tension is already built, I'll let the dice do the talking, it's fun to see how things will play out.

 

[Dannager]

That might be true in some games, but as a DM I have no idea whats going to happen. The PCs might live to become god-kings or die ignobly in a pit in the ground. They might uncover the vast conspiracy or become a part of it. I play to see what happens. If I know what's going to happen, I'd just write it as fiction.

Yes indeed and this is a feature not a bug. If I want to know whats going to happen without any suprises then I'll just watch a movie that I've already seen.

Why would I want to be predictable to myself? I'm not writing a novel here. I like unpredictability. Keeps me on my toes.

Why the heck would I want a predetermined outcome that I control in a game? I'm not interested in telling a predetermined narrative with the occasional dice roll.

Because DMs are not playing a game, in the traditional sense.

DMs do not have a game theory win condition. They do not have a game theory lose condition. Their role is to facilitate the enjoyment of the game - and, I would suggest, they are primarily there to facilitate the players' enjoyment of the game.

Heck, I would argue that the players don't even have traditional win/lose conditions.

As a DM, your goal should be to derive enjoyment through your players' enjoyment of the game. In my eyes, you're not a good DM until you've gotten there. Any argument along the lines of, "If I wanted to control everything I'd just write fiction," misses the point on two counts - first, it ignores the fact that the DM doesn't control the players, and can therefore be surprised by their decisions even while maintaining control over the game world; second, it completely sidelines the players' investment in the game by holding that the only important difference between playing D&D and writing fiction is whether or not you can surprise the DM.

I realize that this is a nuanced position - it makes an effort to acknowledge things like the difference between the game world and the players' sphere of control, the fact that the players have their own hedonistic calculus to satisfy, and the fact that random encounter tables are designed as a tool of convenience rather than a formula to construct an ideal game around - but dammit, it's about time we started trading in nuanced positions instead of, "Embrace the chaos!"

 

[Dannager]

And a good random encounter generator will be based on things like population density, prey distribution, HD distribution and environment.

A good encounter generator will be based on things like, "What will make for an enjoyable play experience?"

In all seriousness, if you created a random encounter generator that accurately mimicked a believable environment, one of two things would be true: a) it would be too dull and tedious to suffice in a D&D-like play environment, or b) everyone except the PCs would already be dead. Perhaps both.

 

[chitzk0i]

When the DM uses random methods to propel the 'story', it frees the DM to describe the situation and watch what happens, and it frees the players from having to second-guess the DM's intentions.

The DM still came up with the location the players are at, the DM still composed the random encounter table, and the DM is still filtering (or not) the results of rolling on that table. The DM's intentions are just as involved here. Just like running a module doesn't absolve the DM of responsibility over the outcome, using a random table doesn't shift the blame to anyone else.

 

[Connorsrpg]

Thank you JonWake for expressing something I have been 'missing' in my 4E games of late. I have been mentioning my love of randomness over many threads for a long time. We LOVE to play this way. It is not just about random encounters and it is certainly NOT about being a lazy DM.

I usually prep my arse off...but how do I do it? I roll a bunch of dice on the random tables available (on the many charts and tools I have on my website) and mash the results into ideas.

I have random charts for Encounters (Creature, Place or Event), Creating Realms (from organisations and tribes through to empires), Random buildings & Businesses, Creating NPCs, etc. I have pooled many sources into one bunch of charts and I use them all the time. It is my favourite way to create stuff. (Oh and you can meet pairs and single creatures on the Creature charts, even if they are not solo or elite).

But I also miss the randomness in game too. I miss the no expectations on how long things should take or how many hits something can last. (A feature always there, but too refined currently). That is why we have Crit and Fumble Charts. We don't care about 'how this will tend to hurt the PCs more'. We embrace it and our charts can be much deadlier than max dam.

WE like a game where you just might die from one bad hit, where, if there is a green dragon in the forest, you just might meet it (or the orcs that have always been there no matter your level), and the players never think I am being unfair. Mind you, those that think random charts are made just for fights??? No, I am sure each DM using them also has ideas for how to use the encounter. (And for the random-loving GM, yes I have a chart for this in the Encounters doc ).

And as for the person asking 'Why would you roll random encounters in front of the players? Well, we have gone one step further - I often get the players to roll for encounters (or draw the card in Savage Worlds). They LOVE it!

Mind you these are the players that flat out refuse to point buy too.

Randomness has a lot to say about running a fun game. (Remember Ray Winninger's First Rule of Dungeoncraft)?

I create all my campaigns randomly. Those interested in doing the same might like to check out the GM Tools of our website too (in sig).

Anyway, Thanks JonWake for saying clearly what I have been trying to express for a long time on a lot of threads.

 

[howandwhy99]

Wandering monsters is always something that has bothered me. Monsters don't "wander" for starters, they live in specific environments and only leave them for absolute necessities(or in the case of more intelligent creatures, boredom or mayhem). When wandering through the forest, the chance that you're going to run into anything more spectacular than a pack of wolves is fairly low. Giants, dragons, oozes, ect... these things live in specific places and are usually fairly keen on staying there.

It's far more likely for YOU to wander into their territory(which is often well marked by primitive signs or natural indications), and in such a case your more nature-savvy character should be able to give you a fair warning of where you should or shouldn't wander in the woods. Even the most unmapped forests are likely to have rumors and stories about them, which a party passing near said location is likely to pick up. It is even more unlikely to be walking down a well-used path and run into anything particularly monstrous, you're more likely to get robbed than attacked by a Hag or something.

First off, I agree with a lot of what you say in the rest of your post, but I don't believe randomness, at least in terms of die rolls, was intended to promote unpredictability. On the contrary, I think it was to express a swath of potentiality with a pragmatically static probability curve. Attack or To Hit rolls are typically understood this way every day. What are the odds of getting an 11 or better on a d20? How about a d12? It's about expressing reliable, predictable probability with a range of results and as few as possible predetermined yeses or nos.

To the above quote, remember Wandering Monsters tables also include % in Lair in their creation. In my estimation WM tables are defined by a particular territory, especially creatures who share a relatively fixed territory (like a dungeon level). A positive WM result means a (still only possible) encounter of creatures in their perceived territory. % in Lair is how often these creatures are at home and not guarding/roaming their territory. Activity Cycles (nocturnal or diurnal, sleep, hibernation, eating, nursing, etc.) all factor into the WM and % in Lair probability determinations of individual monsters as well as their encountered behaviors.

WM checks are based upon territory population density, but a positive result could still wind up with no PC encounter if the WM spots them and successfully evades before the PC has much of a chance. (Most of these monsters have better vision and other senses by far.) So yes, WM results are bumping into creatures in their claimed territories. The exception is emptied territory currently being expanded into by surrounding elements (a.k.a. emptied dungeon levels being refilled).

Civilized paths, as opposed to wilderness trails formed by animals and the like, are for civilized WM table encounters. These are all manner of NPCs with NPC classes, but those results are usually described with varied constancy, like plant life in most wilderness territories. The relevant rolled results are the creatures that prey upon these others, usually more often in less populated regions (basically PC foes like goblins who ambush merchants in the borderland frontier).