Friday, October 14, 2016

The Consequences of Heuristics

Concluding my remarks from this post and this post, the reader may remember that I ended by saying that a DM ought to combine both preparation ahead of time with on-the-spot decision-making, which I have described as using a heuristic.  The word is unfamiliar, so once again I'll quickly say that a heuristic is a decision that is made from the hip, with minimal evaluation of the facts, heavily weighted by the individual's present mind set, biases and experience.

We've talked about how a heuristic can be a negative thing, based on weak or cliched information.  We've also talked about how a heuristic can enable a DM's rapid response to game play that is flagging and requiring momentum.  To that, I'll add that all DMs must be ready to think fast on their feet, given the amount of information that surrounds game play and the potentially pattern-breaking behaviour of innovative players.  A DM can easily feel overwhelmed by all this and unable to make a patient, well-grounded conclusion.  Heuristic decision-making is, therefore, a prerequisite of DMing.

But I stress that it is also a prerequisite of playing an RPG.  Virtually every decision a player makes during the course of the game is made heuristically, without any warning as to what to expect and without the time to piece together a detailed assessment of the problem at hand.  Even in the case where the players have time to plan an assault on a given lair or descend into a dungeon, where they can equip themselves and gear up mentally for the offensive, they still don't know what they're going to encounter before they do.  The DM can take steps to select what can be prepared and what can be left to a heuristic, but the players don't have that luxury.

This gives the DM a tremendous advantage over the players.  It also speaks to the DM's mindset where game-play is concerned.  Arguably, DMs are the sort of people who are made uncomfortable by heuristic decision-making, who prefer the option of planning things out ahead of time.  They are the sort of people who, when making a heuristic decision with unforseen and unpleasant ramifications, want the option of reworking the setting and experience in order to protect themselves.  Players are necessarily subject to the consequences for their actions, being limited in the amount of power they have in a campaign.  On the whole, players are stronger, more ready for the chain reaction of play, having less issue with the potential failure that can result from rapid-fire decision-making.

Yes, I am saying that DMs have the weaker personality.  Or, if the reader prefers, the more defensive personality.  Very rarely does a DM have to accept the mistake they've make and suck up the consequences.  Like an eel, the DM can almost always shift the liability onto the players ~ who, often without complaint (or unaware), accept the burden and keep on going.  I don't say that they accept the burden quietly ~ not remotely! ~ but they do keep moving forward.

So this brings us back to the decision to play D&D, or more specifically the reason why some choose to DM.  The non-specific answer, "the game challenges me," is made more clear in that the game's challenge is in requiring a string of heuristic, off-the-cuff answers that promotes a feeling of stress and risk, resulting in an increased chemical rush of adrenaline and dopamine, particularly when the risk pays off and the character survives.  Fundamentally, we're all just natural drug addicts.

The DM goes one step further in this process by extending the "game play" to quiet afternoons and evenings spent preparing the game in advance, boosting expectation and allowing the DM to "edge" for a longer period of expectation before the game actually occurs, knowing consciously or subconsciously that whatever happens, the system/structure/campaign can be adjusted to ensure the DM's deficiencies are minimized.

Take the time to write down four or five game-changing decisions that have resulted from the reader's involvement in recent games, then take the time to consider the motivation behind each of those decisions.  In most cases, if the reader is honest, there won't be one.  There will be a strong inclination to invent a motivation, to rationalize the moment the decision was made in some context that will offer logic or reason, but more probably the decision was make heuristically, without any plan or solid thought process in place.  Resist the inclination to rationalize.  Accept that hundreds of hours of game play have created an acceptable experience-based patterned response to in-game events that allows the reader, when participating, to simply "go with the flow" in the same detached way that we would watch a movie or participate in a football game.

Embrace this.  Where a snap decision seems to have caused a misstep or resulted in the downfall of a character or the party, trust that the lesson has been learned and that your judgement next time will be improved.  We know that when we have the time to examine something ahead of time, it may still go wrong.  On the whole, we're doing fairly well to stay alive as long as we have in a situation where we're making split-second guesses against a die that may not support our chances for success.

Regarding DMs who may not be aware that they are unconsciously making decisions that re-route the campaign in order to protect themselves, I can only suggest that we should all stop doing this.  It is very easy to do, it is very tempting to do; it is also loathsome and indefensible.  Be aware, however, that we are doing this.  We have the power to do it and corruption has a nature of slipping past undetected.  So often, it is done without a moment's thought, utterly as a heuristic.  It is only by going back and evaluating our performance, night after night, that we can recognize patterns where our behaviour comes short of legitimacy.


Tim said...

This series has reminded me of computer science again. In that field, heuristics are typically functions or methods which produce a "good enough" solution. You sacrifice something like optimality, accuracy or completeness to achieve a quick and easy solution, without necessarily having the backing of a rigorous proof.

For instance, the traveling salesman problem cannot be solved in f(n) time, where f(n) is some polynomial function of n, the size of the input. But there's a heuristic that says pick the next best step every time, disregarding previous steps, and that will give you a decent approximation of the optimal solution.

This leads me to wonder about what a list of DM heuristics might look like. Most computer scientists with some experience will know a few quick-and-dirty heuristics for particular problems. I suspect the vast open-ended nature of D&D makes DMs feel less certain about what tools are at their disposal in a given situation: personally, I've often felt the same stress while DMing as I have when trying to solve a particularly bizarre proof on a midterm. The time is ticking by and you need to establish what solution will solve the problem quickly and effectively.

Then again, there is no "solution" when role-playing, only a response. But a DM who knows their players well can presumably establish what response they want to elicit at a particular moment, and in applying the heuristic achieve it. Might make good content for a short book: how can a DM determine a good heuristic to use, based on how they want to influence the players?

Alexis Smolensk said...


The anchoring heuristic springs immediately to mind.

Keith S said...

I don't know if tables would be considered an anchoring heuristic or not. But I do wonder how tables fit in to all of this. I know some DMs, in particular the OSR types really dig them. My impression is that Alexis, you use tables a fair amount, no?

I haven't used them much in decades of DMing. I'm wondering if they might be something I should consider more so that my decision-making doesn't become more hidebound than it likely is.

Dani Osterman said...

This series of posts is really interesting, and I'm enjoying reading it. I don't know exactly in what direction you are going with these posts, but a list of DM heuristics, a quick-and-dirty toolkit, seems like an interesting idea.

Alexis Smolensk said...


I don't use tables in adventure-building or choosing my setting. I used to, but I've given up on all that. I now choose encounters and situations based on what would be interesting, given the structure and framework I've already provided for each region and sub-region.

I use tables for combat, of course. For most uncertain things, I use ability checks to determine what the character can do or know. There are less uncertain things, however, as more and more sage abilities helps tie these things down.

I don't use tables for abilities because I feel the character can either do it or the character can't - and I try to make it perfectly clear to the player before the trouble starts what they can do. Often there is a percentage of some kind hooked up to this, but most of the time not; more over, the percentage kicks in when the character goes beyond what is certainly possible into what is maybe possible.

The yes/no systems that I use are complex and varied, mostly to give the players awareness of what they can do while keeping me out of the loop as much as possible ~ that is, it keeps me from having to make a decision that hasn't already been made and set as a precedent sometime in the game's past.

Alexis Smolensk said...


Tables are absolutely anchors - because everything we think to include on the table blots out everything we don't think to put on the table. For example, if I use an encounter table to choose what monster I throw at the party, it eliminates the need for me to be inventive and choose to invent a monster rather than have the dice choose one.

Mike said...

"Regarding DMs who may not be aware that they are unconsciously making decisions that re-route the campaign in order to protect themselves, I can only suggest that we should all stop doing this."

This is why I use tables for reactions on the fly or stocking things on the fly. I've a "stack" from the internet, one because I'm lazy and two it helps prevent everything being seen through my own myopic view.

What is planned or non-random is the reaction modifier say.