Crisis at Crusader Citadel reminds the GM to make random reaction rolls for two of the NPCs. Taking player characters through that introductory module soon taught me how capricious the reaction table can be. (Rules, p. 27.)

Using the random table, a hero whose charisma offers no reaction modifier has a 10% chance of triggering a response of "violent hostility" from good-minded journalist Jessica Anderson or police detective Victor Broyko.

Granted, the module assures the GM that these NPCs will be professional/cooperative to the player characters just the same. That being the case, the reaction table shouldn't deliver such a high probability of hostile emotions--before negative reaction modifiers might even come into play.

In a situation like this, I'm now tempted to roll 5 + 1d10 on the reaction chart. Standard reactions will range from 6-15: mildly hostile/suspicious to agreeable/interested. Then charisma-based adjustments can increase/decrease reactions as usual.

These narrowed results allow for a more realistic range of emotions among even-keel NPCs (like Anderson and Broyko). Further, clustering the unadjusted reactions toward "neutral" makes characters with positive or negative reaction modifiers stand out that much more.

Another option for determining same-side-of-the-law NPC reactions is to use 2d10 (average score of 11), which allows for the full range of emotions on the chart but takes the chance of "violent hostility" down to 1% (rolling a 2) and limits "very enthusiastic" to 3% of the reactions (rolling 19-20)--before factoring in any reaction modifiers.

The 1d20 method is well-suited for melodramatic situations. But I don't imagine that Broyko gets tempted to throw his coffee mug at 1 in 10 people who walk into his office each day, or that Anderson wants to slap 10% of the people she interviews, even if they have modest charisma scores.

Using S.D. Anderson's "Charisma Counts" table (as many gamers do) helps curtail some of the problems among same-side reactions, but the volatility of the 1d20 method still remains.

Over the years, I've come to enjoy the way randomized reaction results can help flesh out an NPC's personal preferences: Maybe Dr. Dupont isn't a big fan of wings. Maybe curator Avery Bitman feels at ease around robots.

The GM ultimately controls each NPC, but the randomized reactions give me something more to work with--so long as those results fall within reasonable boundaries, instead of landing wildly all over the map.

The 2d10 method is one way to accomplish that--especially among NPCs on the same side of the law as the heroes.

Here's something more.

Battle Above the Earth suggests that the GM subtract 3 points when rolling the reactions that the good townsfolks in that module will have toward unfamiliar heroes. That super-phobia is rather severe, but it shows how NPCs' personalities can predispose them toward certain reactions.

I could see using variations on that model in other scenarios.

The Dawn of DNA is the only module that bungles badly when it comes to charisma adjustments, leaving too much up to chance.

Halfway through the adventure, when the heroes face disaster situations, the module advises: "It is unnecessary to actually play these events out, just assume that each disaster situation will be personally tailored for the abilities of the PC in question and roll 1d20 plus one for each experience level the character has reached over 4, minus one for each level he is under level 4."

Following the module's guidelines, an adjusted roll of 6 and up earns a player reward money for their success. Rolls of 10 and up also deliver charisma bonuses of +1 to +3.

But a low score on that single die roll offers these consequences:

1 or less: Player bungles badly, looks foolish, -1d8 charisma.
2-3: Player fumbles rescue, -1d4 charisma
4-5: Player is unable to avert disaster, -1d2 charisma

In other words, PCs below 9th level should avoid this adventure.