The Colors of Magic - Available Now

September 29, 2015

Death Omens

Today's post is the second part of Reinhart's guest post on character death.  See part 1 here.  Also, take a look at his blog, Chaos Engineering.

Last time, I explained the appeal of both random character death and dramatic character death, and how to handle both. Before you read this, you owe it to yourself to read that if you haven’t already. I mentioned then that this isn’t a zero-sum game. In fact, you can actually have both random character death and dramatic character death in the same game if your players are willing to make some interesting compromises.

For that purpose, I’m going to give you a system-agnostic mechanic that I use in many of my own roleplaying games: Death Omens. Death Omens help you mitigate the senselessness of random character death while still giving the dice the power to decide some of the when and how of a character’s death. Plus, they’re also really easy to port into most systems.

All you really need are some index cards and a writing utensil. (What good GM doesn’t have have those?) The idea is to directly telegraph when you’re introducing deadly stakes. You see this sort of game mechanic lately in video games like Destiny.

Adapting a mechanic from a video game doesn’t make your RPG into a video game any more than adapting a character from Game of Thrones makes your RPG into an HBO T&A drama.

A Death Omen is a game-world event that signals to the players that their characters might die here.  It’s a note on an index card labeled DEATH OMEN that describes why the scene is potentially deadly.  It makes it clear in black-and-white that death is on the line.

Now and then the dice might create a situation that would permanently remove a player character from the game. When that happens, if no Death Omen is in play for that player, then the consequences are assigned instead to the stakes for that scene.

You’re presenting clear stakes for your conflict scenes, right?  

Stakes and Deadly Stakes

Every time there’s a conflict, both parties want something, and the two goals are in opposition to some degree.  
  • The ghouls want to eat, and the PCs want to explore the tomb.  Being eaten would certainly stop them from exploring the tomb!  
  • The forest fire wants to burn everything, and the PCs want to escape the forest alive.  Being burned alive would certainly stop them from escaping!
  • The orcs want to protect their camp, and the PCs want to sneak in to rescue some captives.  If the PCs get in, the orcs have failed to protect the camp.
  • The Duke wants to preserve his power, and the PCs want him to arrest the nobleman who killed their mentor.  If the Duke arrests the nobleman at the request of some common adventurers, he undermines his power base.

“Stakes” is a common term in all kinds of games.  In essence, it means “the thing that happens if you screw up.”  In a dramatic conflict between PCs and the opposition in a tabletop RPG, stakes always means something that happens when the PCs fail to achieve their goal, or fail to prevent the opposition from achieving theirs.

Sometimes the stakes are deadly:  The ghouls’ hunger and the forest fire are imminent threats to life.  But the orcs might not want to kill intruders so much as keep them away.  They might be content with a blustery show of force, or a cruel and savage beating that drives intruders off.  Sure, orcs are big strong monsters with huge swords.  They’re deadly; but then so are Dukes.  If the PCs screw up with the Duke, he could just have them hanged for trumped up charges.  The scene’s stakes are only “life or death” when the opposition wants to kill the PCs, first and foremost -- if killing the PCs is the opposition’s goal.

Of the two deadly stakes examples, one is necessarily deadly, and the other is possibly deadly.  In both, the opposition wants to kill the PCs; but in the ghoul scene, there are more ways for the PCs to fail than just getting killed.


  • PC Goal: Get into the tomb
  • Opposition Goal: Eat living flesh

Forest Fire

  • PC Goal: Not get burned to ashes
  • Opposition Goal: Burn everything to ashes

In the ghoul scene, there are three possible “failure” conditions:

  1. Opposition goal succeeds:  The ghouls eat the one or more of the PCs
  2. Opposition goal succeeds:  The ghouls eat some innocent villager or some innocent villager’s cattle
  3. PC goal fails:  The PCs are driven off by the ghouls and cannot get into the tomb

In the forest fire scene, there is only one possible failure condition:

  1. Opposition goal succeeds; PC goal fails:  One or more PCs are burned to ashes

Death Omens that Communicate Stakes

When we talk about Death Omens, we’re talking about something that signals to the PCs that the stakes of the scene have become necessarily deadly.

The forest fire scene will start with a Death Omen.  It will say “You could be burned to death in this fire.”

While that card is in play, PC death can happen any time.  It’s a sword of Damocles.

So... were you, like, born with that?

The ghoul scene might not start with a Death Omen.  But once the PCs get closer to the ghouls than any other edible flesh, the GM will probably hand out a card that says “These ghouls want to kill you and eat your flesh.”  If the PCs try luring them into a pit trap with some cattle, or sacrificing an annoying villager to them, the Death Omen might not come out.

Are Death Omens “plot armor”?  No.  They’re a descriptive tool, not a proscriptive tool.  A scene where the GM didn’t want to use character death as stakes anyway doesn’t get a Death Omen.  There’s a social contract that says “nobody’s PC dies unless there’s a Death Omen on the table.”  But as soon as the GM sees death in the offing, out comes an Omen.

This is the most basic application of Death Omens.  They communicate what the stakes of the scene are in the most basic sense.  In theory, you could use an index card to describe the stakes of every scene, but character death as stakes is unique.  In most tabletop RPGs, a player’s ability to play the game at all ceases if his or her character dies.  That’s a big deal.  That’s worth a special omen.

But there is another way to use Death Omens.  It helps make the game more cinematic, and gives the players more control over their characters’ story arcs.

Shape of Your Doom:  Author Stance Death Omens

If you’re a very narrative group, you might want to use Death Omens in an entirely different way.  This is a distinct variation of the above system.  

During Session Zero (or Character Creation for those who don’t know that term), the GM announces the campaign is going to use Death Omens, and the player gets to choose the conditions in which their character may die.  It lets a player choose something thematic about their character, a type of situation that will always spotlight their character, and how rare life-threatening scenes will be.

They have to pick something as their Death Omen, and whatever that thing is, they can’t be killed unless the scene involves that Omen.  They can even choose something rare or seemingly harmless.  

SEEMINGLY harmless.

Each player writes their character name on a card along with their Death Omen and hands it to the GM.  Whatever it is, the GM will serve up life threatening scenes for the party with this characteristic.  When those scenes happen, the GM will place the appropriate Death Omen card(s) on the table.

For example, if you choose “Fire” as your Death Omen, then you can’t die except in scenes where the GM takes out your “Fire” Death Omen card and puts it on the table.  And the GM will not do that unless fire plays a major role in that scene.  You might be crossing a bridge over lava, fighting a red dragon, fleeing a forest fire, battling a mutant pyrokineticist, or running into a burning building to save a loved one.  

Make sure to foreshadow your Death Omen by having your character talk about his or her fears/enemies/weakness all the time.  Make it a shtick.

If your card isn’t out, you’re not in danger of dying.  But you can still suffer major losses.  Here’s where the challenge comes in.  The GM will need to think of alternative stakes for every scene, other than death.  If just one PC has Fire as a Death Omen, a burning building can’t kill most of the PCs.  But they can be injured, disfigured, poisoned with toxic gasses, knocked unconscious, rescued by and indebted to questionable people, have valuable equipment destroyed, etc.

That should be plenty of food for thought. I appreciate feedback, so please give Death Omens a shot in your games, and tell me how they worked for you.

September 21, 2015

Character Death

Today, I present another guest post from Reinhart at Chaos Engineering. This is a two-parter, so stay tuned next week for more on the topic!

Character death: It happens. I think we can all agree that death has a place in stories in roleplaying games, but there are few topics more contentious in gaming. There are plenty of traditionalists who enjoy how characters in D&D and The Call of Cthulhu can die suddenly and without warning. There’s also strong and vocal support across the roleplaying hobby for treating characters as protagonists who are less likely to die a senseless death. Both sides get vilified too much for what is really just a preference in playstyle.

Given the rancor on the subject, you might suspect that these playstyles are incompatible. The thing is, so long as everyone at the table is playing respectfully and talks about it ahead of time, there’s no reason this disagreement has to be a deal breaker. There’s plenty of room for compromise, and if you’re smart, that compromise doesn’t even have to be a zero-sum game.

First of all, if you’re going to find a workable compromise then you need to understand both sides in this tension. A lot of the disrespect I see around this debate is because neither side truly understands the other.

The Case for Random Death

Why would anyone prefer a random death? Well, it comes from a very basic principle of role-playing with dice: play to see what happens. A lot of players enjoy the tension that comes from being uncertain about what is about to happen to their character. Many GM’s enjoy not being in complete control of every situation, and sometimes they need to feel a little less culpable in the decision to remove player characters from the story.

Additionally, random death isn’t so random.  Some people play RPGs to play a game of make-believe instead of to tell a story.  Part of that difference is “what’s there is there.”  In the make-believe game, if there is a death trap there, it’s there; and if the PCs touch the wrong part of it, they will die.  Make-believe style play favors verisimilitude over narrative.

It’s not everyone’s cup of tea, but there’s nothing wrong with wanting and enjoying this style of gaming either.

The Case for a Dramatic Death

Where’s the excitement in a game where you can’t die until the story demands it? If you’re seriously asking that question then you need to go read up on stakes. Just because character death isn’t a lurking threat, doesn’t mean that anything or everything else isn’t up for grabs. The problem with random death is that outside of some expensive D&D spells, it’s pretty damn final. Of all the things that could happen to a character in story, death is probably one of the least interesting because it generally closes off more possibilities than it creates.  

Just as most GM’s run with some basic scenes, challenges, and plot already prepared in mind, many players create characters with potential themes and plot points they want to explore. As a result when the dice end a character prematurely, there can be disappointment with both the player and the GM.

Don’t be a Slave to the Rules

Obviously the players and the GM can always decide to ignore the rules whenever they become inconvenient, but players can find it jarring to just “take back” a roll or consequence like that. It’s generally better to solve the problem before it even starts. Like most RPG issues, the best time to address this is at session zero.

If you and your players all hate random character death, you should probably learn and use a set of rules that don’t emphasize that element. D&D or Call of Cthulhu are great games, but they both have game mechanics that can kill characters without warning. If you don’t want to play that way then maybe pick a game like Fate to tell your story. If you don’t want to learn an entirely new set of rules then perhaps have the players collaborate on crafting their own house-rules for death and dying.

Negotiate the Stakes

Even once you and your players have decided on the rules that best fit your playstyle, there’s still a lot of leeway given to the GM as to when and how character death happens. When a player character heroically runs into a burning building, and the building collapses on them, you don’t have to roll for damage. Their life or death at that moment is should be determined by what makes sense and what is the most interesting result for everyone at the table.

Good GMs are open about the stakes that the players are risking. Let them know when you feel the scene is warranting life or death decisions. “Do you really want to run back into that building? It looks like it could collapse at any moment. It would probably kill you if you got trapped inside now.”

Likewise, during session zero, each player should be very honest about how survivable they expect their characters to be. They don’t all have to agree to this equally, either. One GM that I’ve played Delta Green with suggests that players volunteer for random death by writing “DNR” at the top of their character sheets. When the tension ramps up, those players can usually look forward to a sudden and interesting demise.

I like the DNR house-rule, but as I discussed before just because a player is against a random death doesn’t mean they’re opposed to a dramatic death. For that reason I like negotiating the terms of character death openly when discussing a player’s goals for a character:
  • How do you see your character ending up in this campaign?
  • What would be a suitable heroic arc?
  • How could it turn tragic?

These questions aren’t just there for me to help direct the story, but also to help the player anticipate and recognize what fates are appropriate for their character. Many players accept an eventual demise better when they’ve already accomplished some of their goals and their doom conforms with how they expected their character to die.

Pacing Your Doom

Even if everyone is alright with the rules and dramatic circumstances that govern their demise, there’s still a certain combination of logistics and art to determining when to kill of a character. If your player characters really can drop out of the story at any moment, make sure that players have a developed replacement character. If you and your player characters like to have developed backstories and lead-ins for your player characters then that means perhaps there should be occasional insertions of the back-up characters into the narrative before they’re eventually needed.

Regardless of whether death in your game is random or dramatically orchestrated, it tends to leave one player with less to do and a lot less to get engaged with. For that reason it’s best to plan for death near the conclusion of a session or adventure. For those of you letting the dice determine this, that means planning and designing your adventure so that more lethal threats are introduced later and near important milestones. Players are generally more accepting of their characters being murdered by the villain of the plot than they are by a random owlbear they tripped over in the woods.

Do it Right

A player’s character is often the only way that the player can interact with the shared imagined world of your RPG.  Few systems allow players to manipulate events or storylines without a character.  So death is a big deal.  Death is “game over - at least for now” for a player.  Using session zero, thinking about what mode of character death you want to include in your game, preparing for the type of game you want to run (and doing it right), and taking care not to waste your players’ time all contribute to solid GM skills -- regardless of how you want to run character death.

This is part 1 of a two-part series. Stay tuned for Part 2: Death Omens - a house rule for nearly any RPG that might satisfy all types of players!

September 14, 2015

RPG Session Pacing

I've done a lot of writing about pacing, but I managed not to include a detailed discussion of the key techniques that contribute to solid session pacing.  I gave some quick tips during the series, and Carlos, one of the authors of the upcoming Asylum RPG wrote a guest post on pacing con games for me but that's the extent of it.  

It deserves more.  And because I love you, dear readers, I have decided to set my little listicle to the haunting lyricism of Yeats’ famous poem, “The Second Coming.” Because sleep deprivation. Or something.

The Five Basic Techniques of Session Pacing

There are five basic techniques that make for good session pacing.
  1. Open with a Bang
  2. Modulate Opportunities and Threats
  3. Prep Several Climatic Events
  4. Build Toward an Eleventh Hour Climax
  5. Close on a Hook

Open with a Bang

Turning and turning in the widening gyre
The falcon cannot hear the falconer;
Things fall apart; the centre cannot hold;
Mere anarchy is loosed upon the world...

Every action-adventure television show opens with a teaser scene, where something exciting happens.  Bond movies do a great job of this.  If you can, start your session off with something fast-paced, exciting, and fun.  Make it connect to the plot if you can.
The benefits of opening with a bang are numerous.  First, it gets the table focused on the game.  It's hard to ignore a bang!  Second, it cuts past the "loose ends" discussions that start most RPG sessions, and can often go on for hours.  The players will have to skip a lot of those and get to the most important ones when they get a chance.  A bang is always an unexpected threat (see below) that must be dealt with now, now, now!
Example:  Let's say you're on session #2 of your hex crawl through some demon-haunted city ruins.  Last session, the PCs made it to the ruins, and the session ended with a big reveal: The “guide” the party brought with them turned out to be a mad wizard in disguise, who ran off when they got inside the city.  So let's open this session with a bang!
"Where we last left off, your guide had just teleported away in a fit of mad cackling.  But he seems to have forgotten his Bag of Holding.  As you move to examine it, four one-hundred-pound iron cobra snake constructs come slithering out.  But that's not the worst of it!  The sudden increase in weight causes the ruined barracks floor to collapse under you, dropping you into a root cellar!  Everybody make reflex saves!"

Modulate Opportunities and Threats

The blood-dimmed tide is loosed, and everywhere
The ceremony of innocence is drowned;
The best lack all conviction, while the worst
Are full of passionate intensity...

I’ve talked about modulating scene types before.  It’s important to change things up frequently to keep interest up.  In the moment to moment of an RPG session, the easiest way to do this is to simply switch back and forth between opportunities and threats.

Every event in your RPG either opens with a threat, or opens with an opportunity.  An opportunity is a chance for the PCs to take action.  A dungeon to explore is an opportunity.  A mystic portal is an opportunity.  A nosy bartender who might have information is an opportunity.  To use Robin D. Laws' beat terms from Hamlet's Hit Points, an opportunity is a scene that opens with a dramatic or procedural upward beat.  That means the audience members (the players) feel an increase in hope for the protagonists (their characters) at the start.
A threat is a scene that opens with a dramatic or procedural downward beat - a scene that starts by making the audience members (the players) feel fear for the protagonists (their characters).  It’s a danger at least initially out of the PCs' control.  An ambush is a threat.  Ominous news about the Dark Lord is a threat.  A sentry guarding the enemy camp is a threat.  The example session-opening-bang! scene opener, above, is a threat.

So here's what you do:  If you just opened a scene with a danger out of the PCs’ control, open the next scene by giving them something enticing and hopeful.  
Example:  We started the session with a threat.  Let's give them an opportunity:  "Aarom the Evoker is very skilled in Spellcraft, so he knows that a Type II Bag of Holding can carry up to about a hundred more pounds of weight than those snakes.  Wonder what else could be in there?  Also, the wall in this cellar has a riddle painted on it in dretch blood that is several weeks old:  

Prep Several Climatic Events

Surely some revelation is at hand;
Surely the Second Coming is at hand.
The Second Coming! Hardly are those words out
When a vast image out of Spiritus Mundi
Troubles my sight: a waste of desert sand...
Most websites that give tips to GMs will tell you not to prep the things that happen.  Instead prep the things that the antagonists do.  This is very sound advice, but there are exceptions.  Big, exciting scenes are better when the GM has done at least a little work ahead of time; and session climax scenes make a lot more sense if the session has been building toward them.  To build toward a cool climax, you need to have a few cool climax ideas in your head to build toward.

Here’s how you do it:  In your session prep, or during the session itself, think of a handful of cool things that could happen as the climax of the upcoming session.  Go for quality over quantity -- three good ones that inspire you to think more about them are better than a quick list of ten ideas with no meat on them.  

If you’ve got a dungeon planned out with everything down to the five foot square, this should be easy:  Just find the two to five coolest things in the dungeon within a session’s reach of the PCs, and imagine how they’d make great climatic endings for this session.  (If there’s not much exciting nearby, your dungeon sucks.  No offense.  Modify it so that there are dramatic and cool things that the PCs might discover within a session’s span.)

Example:  When you looked at your ruined, demon-haunted city, you could see that the PCs could easily get to any one of four points of interest within one session:
  1. The PCs could have a long fight in a big ruined palace, where the palace starts to collapse, and they have to flee for the exits while being pursued by monsters.  The mad wizard has golems that will take a long time to beat, and he’ll open with a fireball that shakes the building and starts the collapse.  
  2. The PCs encounter a demon hive, with minor dretches or other little demons climbing out of slimy egg sacs constantly.  They have a prayer ritual that will get the cleric's god to banish the whole hive and lock the location down from future incursions, but it will take five successful religion checks on the cleric's turns (each a standard action), so they have to hold fast in the middle of the hive under constant onslaught by nearly infinite opponents, without the cleric’s help, until the ritual is complete.
  3. The PCs follow a series of clues left by a Glabrezu to lead adventurers to him.  He's too powerful for them to fight, but he’s slothful and bored to tears here; so he offers them two choices:  First, they can play a game of cards against him, and if they win, he will leave forever and grant them each one wish.  If they lose, he will eat them.  His name is Immortal Wynorax, and he's known to have a tell -- there's even an old song about "Immortal Win" (i.e. Immortal Wyn) that the bard knows.  And the chorus goes, "He whistles when he draws and he snaps his claws | He snaps his claws when he's bluffing."  Play poker with the PCs, and snap your fingers every time you bluff, so there's no way they can lose.  The real drama here is that Glabrezu wishes are always poisoned.
  4. A variant collapsing building:  The PCs are in an old library with 6 tall shelves covered in old, cracked clay tablets.  It's a super hard fight against spined devils or something tough, but the PCs can shove the old rusty copper shelves full of clay tablets over to instantly kill the monsters (or deal 10d6 damage, ref DC 20 for half, to any PC caught under them!).  At the end, they find 1d4 intact magic scrolls (on clay tablets) on each shelf they didn't knock over (but none on the shelves that fell).  The devils don’t think about pushing the shelves over on the PCs.  They want to taste living flesh.

Build Toward an Eleventh Hour Climax

A shape with lion body and the head of a man,
A gaze blank and pitiless as the sun,
Is moving its slow thighs, while all about it
Wind shadows of the indignant desert birds...

When the clock on the wall tells you there’s one hour left before the session is over, lead right into the climax scene you prepared that best fits the route the PCs have taken up to this point.  Tailor the scene so it makes sense as a response to the PCs’ actions.  Then make it big, loud, and exciting.  Ramp up the stakes.

This is not railroading.  You're not leading the players by the nose.  They're choosing the route they want to take, and you're selecting the cool scene that best fits their path, tailoring it to their actions, and unleashing it at just the right time. Conflict takes protagonists and antagonists.  As the GM, you control the antagonists, so you control half of the equation.  The climactic confrontation of the session happens when the protagonists and antagonists clash in a cool, decisive way.
Here's how you do it:  One hour before the end of your session, start a scene that raises the stakes dramatically.  In D&D, it might be as simple as ending your session on a hard, exciting fight.  But the stakes are more than just hit points and 5,000gp diamonds: The goals of the PCs and antagonists are in conflict, and those goals are the stakes.
But the real challenge is literally pacing yourself.  As the players advance along the story at their own pace and discretion, drop in clues and plot elements that are necessary to foreshadow the climactic scene that best fits the direction the players are moving.  If they focus on clues going in a particular direction, reinforce that and give more clues pointing in that direction.

Here, again, we need to look at dungeon games versus non-dungeon games.

Pacing a Dungeon  
In typical dungeon-based games, the path is often literally a path.  The PCs follow strange sounds through Door #1 into hallway A that leads to the East wing of the dungeon, which means they will be in the vicinity of Climatic Event #4.  Once they commit to a path, they usually stick with it.  The challenge is getting them to the climax scene at the right time.  Dungeons tend to be pre-drawn, down to the 5’ square.  
  • If you need to speed things up in a dungeon, you can use secret passages to bypass rooms, tip the PCs off to traps that would slow them down, or make combat encounters go faster by calling the fight sooner.
  • If you need to slow things down in a dungeon, use random encounters (but try to connect them to the plot), beef up combat encounters, add traps to slow the PCs down, or take a snack break.  Sometimes taking 15 minutes to eat some cheetos is a better use of the group’s time than a throwaway wandering monster with no plot relevance.  

Pacing in Non-Dungeon Games
Non-dungeon games (modern horror, mysery, thriller, space opera, etc.) are distinctly different from dungeon games.  
  • Without a mapped dungeon, the PCs can pursue multiple paths of investigation at once, which means you may not know which climax they’re going to hit at the end of the session until just before.
  • It’s a lot more common to split the party in non-dungeon games, especially modern-day RPGs where cell phones are common.  When the climax comes, you have to get the PCs back together.  A good way to do that is to start the climatic scene for the largest PC group, and have the other PCs show up a moment later.
  • If you need to speed things up in non-dungeon games, it’s a lot easier:  There’s not a literal map of the dungeon forcing you to find a way to cut around plot to get to the climax.  Just make sure there’s been enough build-up that the climax makes sense.
  • It’s a lot harder to slow things down, though.  If the PCs know where the russian mob boss is meeting, there’s not a dungeon full of potential traps and monsters in the way.  And traffic on the 405 doesn’t take as long in game time as it does in real life.  So you have to stick a filler encounter in that doesn’t feel like a filler encounter.  That’s tough.  Here are my suggestions:
    • Use an encounter that advances a totally different plot.  If the PCs are tracking toward the climax of the Russian mob plot, give them an encounter that lets them know what’s new in the femme fatale plot.
    • Use a friendly-unfriendly NPC.  Maybe it’s the chief telling you this time you’ve gone too far!  You’re on suspension until IAD can look into this mess!  Turn in your badge and gun!  Maybe it’s the landlord demanding to know why there was a hidden camera in the light fixture, and what are you really doing in these offices and should I be reporting this to the FBI?
    • Use aggressive scene framing, if your group is OK with that technique

Example:  The books in the mad wizard's bag of holding point towards the palace.  The PCs will also notice dretch demons spawning from the old sewer.  There are mysterious riddles that lead from one to another and eventually to the Glabrezu.  And the PCs will recognize off in the distance, a library that might contain ancient secrets that could help them here.
If the PCs choose to follow the chain of riddles, keep giving them demon encounters, athletic challenges as they negotiate crumbling bridges, flooded avenues, and rubble-clogged streets, and riddles painted on walls in dretch blood.  If your sessions usually end at 10:00pm, somewhere around 9:00, the travels lead to the smoky temple where Immortal Wynorax takes his leisure.
If you've drawn the dungeon out in 5' squares, it's still easy to improvise the climax into the 9:00pm slot.  Just make sure to hurry or stretch the 8:00 scene to make it fit one hour.  Say the 8:00 scene was a riddle, but the PCs figured it out too soon.  Easy - just throw in a fight with some dretches (to remind them of the other hooks).  What if they're going too slow?  You wrote 6 riddles, and they only got through 3 of them. They still have to go to the fountain and the lighthouse before they get the clue for the temple.  No problemo!  Just make the last riddle the one that leads to the temple instead of the fountain.

Close on a Hook

The darkness drops again but now I know
That twenty centuries of stony sleep
Were vexed to nightmare by a rocking cradle,
And what rough beast, its hour come round at last,
Slouches towards Bethlehem to be born?

The best ending to a session is an unforgettable ending.  In an RPG, where the audience members play the protagonists, the most unforgettable endings are those that leave the PCs with a question.
After the climax, make something happen or give the PCs some information that leaves them with more questions than answers.
Example:  Let's say, using the example above, that the PCs won the card game, and they refused the Glabrezu's offer to grant their wishes.  Well, almost.  The party rogue (who had 21 Dexterity) wished for more Dexterity.  Naturally, the wish is twisted, so the DM takes the rogue's player aside and discusses options.  When they're done, they've agreed that the rogue gains +1 Dexterity but also shifts to Evil alignment.
The session ends with the GM describing what happens.
"As the Glabrezu vanishes across the barrier between planes, you all sigh with relief, except Banor the Paladin, who suddenly feels a chilling spiritual wind coming off Dogfinger the Rogue.  A quick glance confirms it:  Her aura is evil.  She must have secretly made a wish!  And that's a wrap folks!  You can email me any questions you have about that aura and I'll get back to you before next game."
Here’s another quick tip:  Remember all those climaxes you prepped?  Use them for your end-of-session hook.  So an alternative session ender in the running example could have been like this:

“As Immortal Wynorax vanishes, you all breathe a sigh of relief.  In the chamber you find [treasure parcel.]  As you’re collecting the loot, you notice a few clay tablets.  Arrom the Evoker can tell they come from the great library that once stood in this city before its downfall.  The library must be close to this old temple.”