The DM's Rubric - X2 Castle Amber as an Example

In my last post, I said that X2 Castle Amber made me a better DM, but I did not fully explain why.

What makes a good DM and good players is understanding the assignment. There is a reason that meme exists. Most role-playing games give players and referees specific roles, often resolved through specific die rolls. Understanding which choices matter, which rolls apply, and what the consequences are is critical to fun and successful play.

I used to be a teacher, and one of the hardest lessons to learn was how to create a good rubric.

X2 Castle Amber made me a better Dungeon Master because it forced me to understand the difference between player agency and railroading, and how inconsistent expectations around choices and die rolls undermine good play. In teaching terms, Castle Amber shows what happens when a game’s rubric changes without warning. Learning to recognize that made me a better DM in every game I now run.

A rubric, as I like to define it, is this:

“A scoring guide that clearly defines the criteria, expectations, and levels of achievement for an assignment.”

In role-playing games, this means understanding how different roles and different die rolls are meant to work. When those expectations are clear and consistent, play improves.

I have touched on learning before. The White Box Set teaches gameplay through tangible examples. My five-star review of the 2000 Dungeons and Dragons movie is about how not to run a campaign (or a movie). My earlier X2 post describes a real learning experience at the table.

The first thing a DM needs to learn is what is and is not a railroad.

In my Keep on the Borderlands series, I ran the same end scenario three times. I prepared over one hundred monsters for a large fight. Two runs ended in total party kills. One did not. The difference was player agency. In the successful session, the players did the unexpected. I did not force them into a fight simply because I had prepared one. 

Players do not know or care what the DM prepared. If they show agency, they should not be pushed into a predetermined outcome. The thief might back away. The wizard might find a clever solution. The cleric might use magic. The fighter might decide the fight is not worth the cost. That is not avoiding play. That is play.

From the DM’s perspective, this should be a success. The players are engaged. It's a consequence of having great players, not a failure to anticipate what is needed or desired. 

Yes, it is frustrating to prepare material that does not get used. Too bad. That is part of the job. Having those monsters ready does not mean they must appear exactly as planned. Presenting the same material in a different way is not railroading.

If the party disguises themselves as enemies and talks their way into the leader’s tent, only a few of those creatures might ever be used and probably not in a fight. If they encounter the group of 100 creatures in smaller pieces and defeat them through magic, logic, or trickery, that is also not a railroad. In each case, the party made meaningful choices despite what the DM planned. 

De-escalating a railroad situation is not railroading.

X2 Castle Amber works differently. It presents a series of changing criteria and expectations. It uses alternating rubrics, and structurally it is a railroad. The players are trapped and pushed from scenario to scenario like a movie. The fun comes from recognizing the railroad and finding the exits. The module describes only one exit, but players are savvy and smart. They might come up with 3 exits. 

This only works if the players are competitive and willing to play that kind of game. If they are not, the DM should not run it. The same warning applies to “you wake up in a prison,” “the king summons you,” or even “you meet in a tavern.” Any of these can become a railroad if handled poorly.

At this point, you are getting spoilers for a 44-year-old module. I don't feel bad, but if you don't own this, perhaps you should stop reading here and buy it at DriveThruRPG.  

Consider the boxing match in X2. It is a straightforward sequence of attack rolls with the option to quit. The rules are clear and the odds are fair. The very next encounter, the dining room, is completely different. Survival depends on a chain of choices and saving throws. A saving throw is not the same as a combat roll.

An attack roll rarely kills a character outright. A saving throw often represents a single moment of survival or death. In the dining room, players are given chances to avoid those saves, but they are not told that those choices matter. The consequences are not clear. If the DM presents this poorly, the players may never realize they had a choice at all.

From a teaching perspective, combat is a series of connected decisions that lead to random outcomes. Each result feeds into the next choice. The character has agency.

Dice are uncertainty. Don't roll them
if everyone is certain. 

A saving throw is one roll with no follow-up. X2 makes this worse by mixing saves that grant benefits on failure, events with no rolls at all, and standard save-or-die effects. The rules change from scene to scene. When players face many such challenges in a row, survival becomes unlikely, not because of poor decisions, but because of constant uncertainty.

This reminds me of another lesson about rubrics.

In school, passing is often set at sixty-five percent. That may not sound impressive, but context matters. On a spelling test of seven to twelve words, that threshold makes sense. It balances difficulty, memory limits, but not the fairness. 

Problems arise when teachers scale assessments without adjusting expectations. A twenty-word spelling test with a ninety-five percent passing requirement allows only one mistake. That is unreasonable. It also confuses failure with consequence.

I remember having to write misspelled words ten times each. That was not failure. It was reinforcement. I was not retested, but I learned the words. That is a consequence, and it is good teaching.

The passing bar stays at sixty-five percent because some people have advantages. Some know spelling rules. Some do not. Knowing when to apply “I before E” is like knowing what the Deck of Many Things is before drawing from it. The situation is stacked whether you realize it or not, and there is nothing hard and fast about applying rules of thumb. "I before E" is often wrong and a Deck of Many Things is usually a deal from the bottom. 

In role-playing games, failure and consequence are often treated as the same thing. In real life, they are not. Surviving Castle Amber’s infamous meal, where the rules and consequences change from roll to roll, is hard. It can work, but only if players understand the choices they are making.

I have already scripted out the meal and the boxing match to conform to how I should have done the meal years ago and to match how I really handled the boxing match. One is what I wished I had done and the other will be a retelling of a good experience. I hope you roll with the creative drama. There will be spoilers warnings on the dramatic turn in my posts. And hopefully some sage advice. 

To survive Castle Amber and enjoy it, both the DM and the players need agency at the table. Once you understand what choices are available, you can make decisions that lead to success as a player, a character, and a DM.

I hope you will follow my future series on Castle Amber. I will be running it solo so I can explain the choices I make from both sides of the DM screen.

And now for the overt commercial: 

I use Necrotic Gnome's Old-School Essentials but picked up the boxed sets from a Kickstarter. You can approximate this with two titles: 

The Referee's Tome and The Player's Tome. 

I hope that I can replace my original D&D books someday. 

You can get the Basic PDF from DriveThruRPG, and they offer both The Expert Book and B2 as print-on-demand. I cannot tell you how nuts that makes me. Why offer parts 2 and 3 as POD but not part one? Pull it together, WotC. You do crazy stuff. 

And for that matter, if they had the BECMI titles in POD, I'd own those, too. But alas, WotC. 

"Passion is inversely proportional to information had."

The above quote is from Gregory Benford. And it's a truism for RPGs. While I build beautiful worlds in my notes, very little of that makes it through to the players. I might know precisely why a gang of whatevers are doing whatever they are doing, but the players are satisfied with the idea that they are merely jerks. 

It works. Players like to have that room to grow, and they can't grow if smothered with too much B.S. 

There is nothing better than the party discovering some sort of detail that just works for them, but there are many cases where they have no opportunity to gain such information without a data dump. Some things just go to the grave with the player's antagonists. It's fine. 

But sometimes, I like to give information. For example, I hope that every player knows how to use the to-hit tables and can calculate their own bonuses or minuses. It makes my game easier. In fact, I often have the players throw dice for even the monsters. It cuts down on paperwork, but sometimes it is an opportunity to give them a hint about something outside of combat. 

For example, if two equal-level fighters are side by side, shooting arrows at a target and both roll the same number, both should hit or miss the target. However, this is a good place to drop a hint about other stuff. Obviously, two great fighting men should know how good they are. For example, someone might have a cursed weapon or a magic weapon. The target may have some magical device that only applies under certain circumstances like once per round. Once the party is aware of some weirdness, they can start ruling stuff out by logic, just like the real world. 

It's probably magic. 

It saves on the "+1 magic sword" crap. 

There are times to hide some rolls, such as surprise or hiding in shadows. But even those rolls can give information. 

One of my favorite tricks is when the party is surprised, I'll drop a die out of sight and say, "You hear a noise." Surprise is a surprise, there is very little you can do to mitigate it due to the mechanics. However, it isn't very fun to be surprised. By making that announcement and letting the party act accordingly, I am cranking up the pressure AND pushing agency to the party. It creates an environment of anxiety while allowing for possible (slight) mitigation PLUS it allows the players to set a standard of expectation that can easily be read. 

For example, if a party thinks they are in an ambush situation, they may try to arrange themselves in such a way as to defend high-value players like Clerics and Magic-Users with meatshield Fighters and Rangers. On the other hand, if they never do this, you can set a different dynamic where those players are captured or incapacitated and the party is looking at a hostage situation rather than a TPK. It's up to the DM to receive the party's intentions or style and react accordingly. 

One of my favorite experiences was a Thief who decided to sneak up to the walls of a fortification for a little recon. The whole party seemed to support the idea. I rolled for his hide in the shadows and move silently attempts. Each time, I rolled amazingly well. No one saw or heard anything. They were such good rolls that I showed the player the results. Obviously, these should have been secret, but they were so perfect so I decided to show her. 

Then, disaster. The player of the Barbarian was having a little sidebar with another player when he suddenly realized stuff was happening and asked, "What's happening?" 

Once the party explained the plan, the Barbarian nodded sagely and bellowed, "Look out! I can see you!" 

Well of course you can see him. He isn't hiding from you, you twit. 

The sneaking Thief got this "Oh, shit" look on her face. I leaned over and showed her that the dice indicated she was still not visible to the people on the castle wall. 

To add to the merriment, I decided that the Barbarian's actions would be taken literally. The lookout on the wall answered: "Oh geez," and stepped back out of sight. 

"How about now?" asked the lookout. 

The party was gobsmacked. I gave them a few minutes to work out a plan. The Barbarian was drooling dumb and for once, his actual ability score matched the player's actions. The party adapted to the situation and everyone climbed the wall while the Barbarian offered unhelpful tips to the lookouts. No one intended this possibility, but damn it was fun. 

You can't hide everything all of the time, but you also can't data dump on the players too much. Even if it is mechanical in nature. Also, you shouldn't try too hard to hide certain bits of data. 

As a DM, you build a scenario, a story if you will, but you can't know how it will be received and interpreted. Information from the DM to the players is a fluid thing. You are effectively trying to merge the player's fictional actions with the player's visceral need for information. The DM needs to decide from the get-go what information is worth hiding and what is not.