Agency in D&D - The Roll of the Dice

I decided to answer a question on Twitter about bad rolls in D&D and how they impact the game here as the answer is a bit long. First, a digression.

The first few times I played Catan I despised the game. I rolled badly at the start and I ended up with terrible tiles. Then the rest of the game was watching everyone else build while I sat around doing nothing, as I couldn’t get resources. So I didn’t like the game. We bought several expansions, Seafarers, Cities and Knights, Traders and Barbarians, etc. and each of them had the same problem, too few resources for some players when playing by the rules. Then one night I sat down and combined every set together into one huge game board.

They we played, we decided on a 20 point game because of the expansion, and it was great. There was enough game board available that everyone got to have resources and build, even if some did better than others. There was still agency in the game, I find that’s the big issue.

In D&D, the problem for me is this. I want dice rolls in the game, randomization makes the game world seem immersively real, and it takes some of the burden of decision making off the referee. There is something valuable to achieving a goal, overcoming an obstacle, when you know that you did it independently of help from the ref and despite game mechanics that are independent of everyone. There is a sense of achievement, and it is, in part, meaningful and rewarding because it contrasts with the times that you lose, and the dice don’t work out

So for me, I want to keep dice rolling in the game, I don’t want players to be able to say, “that didn’t happen”, independently of the dice. When you roll them, they should count, otherwise I think their value and impact on the game is lessened. 

However, I also feel the problem of bad rolls, not that any one particular failure is a problem, because there have to be some failures, otherwise the game loses it’s interest. But long strings of failures, particularly when you don’t get to play often, are an issue. 

I addressed this issue in my games when I had “that player”, the one who almost always rolled badly. The inner statistician in me knows that for the most part dice work as advertised, if you roll a d20 then 1’s will come up about 5% of the time, give or take. But nonetheless I have known players who have this issue, for whatever reason. 

Here is what I do in the game to mitigate the issue. 

First, old school D&D recommends that players describe what they do and the referee only rolls if they feel that the outcome is uncertain. This approach is underused as many players expect, and even insist, on having a game mechanic for everything they want to do, or to have their actions end up at a game mechanic. 

But part of the point of a referee is to give the game another option resolution without mechanics or rolls. In this situation, it isn’t the roll that matters, it’s the player's ability to make a convincing case for something. This can take the form of the ref telling them their idea worked, or it can be a modifier for the roll. So for example, one of the PCs in my Friday group is a shipwright. They were looking to book passage on a ship leaving the city. Problem was that they were short on funds. So the PC walked the docks and asked how each of the ships looked. “Which one needs work?” He was a shipwright, so I figured he could tell by doing an inspection, and I told him that the Lion’s Roar was the most run down looking ship. So he spent time checking it out, and asking where the problems were. I told him some details, and when they asked about passage and it came to price, the PC mentioned that he was a shipwright, and that he could fix X, Y, Z etc. on the voyage in exchange for a discount. Now, according to the rules I can roll for this to see how the purser reacted, use a charisma modifier, etc. 

Instead, I decided that the purser would just say yes. It was a good deal for them, and it was good play on the part of the player to ask these questions and leverage their skill. Not everything requires a roll. Once players realize this, they start looking for ways to interact with the environment that don’t rely on the dice. 

So this is one solution for the problem, it won't’ help with combat rolls or saves, but it can help with other rolls, and give that player who can’t roll other forms of agency in the game to achieve things. 

For the combat rolls and saves part, I have two suggestions. 

First, whatever you decide to do, I would NOT recommend that you leave it to the ref to make the decision on when to overrule the dice. I know this seems counterintuitive, but hear me out.

If the ref can change rolls, then, inevitably, players  will ask, “why didn’t she change the roll THAT time”? And that’s a good question. As the ref, I DON’T WANT to have to make that decision, ever. I want the dice to decide for me as often as possible, as this makes my role as objective as it CAN be. Favoritism, intentional or otherwise, can happen as well, which should be obvious. If you rely on the ref to resolve this issue through direct intervention, the burden is on them to be fair, and that’s a big burden to allocate. 

However, I also don’t recommend playing in a game where the players can just say, “That didn’t happen” whenever they want. There are a few reasons for this. First, it takes away some of the immersion and excitement from the game, as you know you can always “roll it back”.

It also means that the dice lose their impact, why bother rolling if the players can just say no? I’ve tried this sort of playstyle with D&D a few times, and it becomes weird really fast. Players can get decision paralysis when they realize they can just make things happen by fiat. 

To be clear, there are games that use cooperative decision making by players and these games are hunky dory. D&D is not one of those games, it was designed, and all iterations of it were designed, for some form of dice/ref resolution with input from players. It wasn’t designed for ‘that didn’t happen’ mechanics. 

So the solution is to integrate the “that didn’t happen” into existing mechanics, rather than putting it on the ref. I think it makes sense to give players a way to influence things, the key is that it can’t be ALL OF THE TIME, it has to be SOME OF THE TIME. Fortunately, as is often the case, the answer is right there in the game, you just have to change your lenses to see it.

In AD&D if you have a high dexterity, when you are surprised you neutralize one segment of surprise per point of DEX RAA. So the dice roll happens, but your DEX can neutralize the result. 

So sometimes it can roll it back. 

A monk in 1e who is hit by a successful “to hit” roll with a missile weapon can roll a saving throw to deflect the missile. That takes an existing die result (the “to hit” roll) and changes it. 

AD&D also has a lot of “passive narrative mechanics”, eg. saving throws and HP. The reason your 5th level fighter doesn’t die from the sword stroke that would kill a common soldier is that they are lucky, agile, etc., this manifests in the game in the form of HP. But HP are there if you choose to use them or not, that’s why they are “passive”, but they are definitely story or narrative mechanics, as they reflect the hero’s ability to survive and progress despite things that would take out a minor character in the story. Saving throws are the same thing they mean you always have a chance of surviving, even in situations that appear hopeless, because you are the hero of the story. Gygax frames saving throws in exactly this way, the hero always has a shot.

So the game already has the answer, just take what is already there and leverage it, so for Bhakashal, the AD&D setting/retroclone I am creating, you divide PC’s HP by 4, and that’s how many of their hit points they can actively add as modifiers to other rolls. In short, it takes a passive luck point mechanic that always works to absorb damage (hit points) and gives it to the PCs so they can spend those points on other things, adding to the “to hit” roll, or adding to the saving throw. Once they are all spent, you can’t spend more until you are back to full HP again.

So say I have 12 HP. I am allowed to spend 3 hp to change rolls. Say I take 2 hp damage from a sword. Then I decide to spend 3 hp to change a saving throw. Now I have 7 hp left and I can’t spend any more hit points until I’m back to 12 HP again. So you need to be thoughtful about when to use them.

The key to making this work and not becoming a case of always ignoring the dice is that HP are a limited resource, and the PCs only get to spend ¼ of their max total HP on this mechanic. So it has a real cost (if you want to spend 2 HP to make that saving throw when you only have say 4 hp left, it’s a tough call), and the resource is limited (if I have 5 of my HP to spend and I spend those 5 hp.

Stars Without Number has a class mechanic that allows you to reroll a failed skill check a limited number of times per day. In Bhakashal anyone can sneak (roll to see if you surprise the enemy) but if a thief fails to surprise the enemy, they get to roll a move silently to try again.

The hack I use for Bhakashal (¼ of your HP are available to change rolls) works easily with any standard D&D game, but you could simply give a player X number of rerolls per session or per game day, as long as it is limited then the player has agency in deciding how to use it. The agency makes it important. If you allow them to reroll any result, the dice lose their efficacy and impacts and the game becomes very different. If you give them limited access but provide some mechanic, it preserves the value of the dice, gives them some agency back as they have to be involved in making the decisions that save them. 

I saw this happen in my last game, one of the PCs refused to spend any of his HP allocation on rolls for the adventure, so when he came to a decisive point, he had them available to him. 

When the adventure was over, I looked at him and said, “That was a good decision, not spending any of your discretionary HP early on, it made the difference in the fight”. He was beaming, as he knew it was because of good play on his part. 

One last thing. The tweet that was part of this original discussion (not from Ajit George, but from Banana Chan) had a line in it that I wanted to address.

“... giving players narrative control in TTRPGs when they fail a roll overlaps with having safe, responsible play”.

I think this is a dangerous equivalency to make. Because in the heated, divisive TTRPG sphere, this quite easily translates into: “Those games which don’t give narrative control to players when they fail a roll are unsafe or irresponsible”. This is of course a possibility, but that remains to be explored, and it seems unnecessarily inflammatory to frame it this way given the way the discourse is around TTRPGS.

It can also elide very quickly into, “referees who do not give players narrative control in TTRPGs when they fail a roll are unsafe and irresponsible referees.” To be clear I am NOT suggesting that this was the implication, but it would be disingenuous to suggest that it will not be made.

I’ve seen it made before, people have suggested here in game design Twitter that the role of the DM is essentially one with a God Complex, that DMs have psychological issues because they play a game where they have such power, that the DM’s roll is inherently adversarial and hostile, and they almost always end up in some sort of psychological evaluation of the role of the ref. 

I don’t think that’s a fruitful avenue to pursue. Instead, I think the focus needs to be on solutions that mitigate the extreme cases (players that consistently roll badly) while preserving the value of those rolls to the game.