Bonds for Occult Antiheroes

I’m sure something like this has been done before, but I got the idea watching Hemlock Grove. It’s a mechanic for games where the players are meant to be fairly unheroic, selfish individuals that are forced by circumstances to become protagonists even though they’d rather while away their days filled with angst. You know, your John Constantines, Angels, Nick Knights, Duncan Macleods, and such. They’re only heroic in the circumstances because they’re not actively villainous and they have a handful of things that are important enough to them to step up and defend.

Bonds

Each player character starts the game with five bonds. These are five nouns important enough that the character will go out of her way to defend them. Three of them have to be people (friends, family members, or just people the character idealizes and wants to protect, but not other PCs). The other two can be additional people or places, objects, or ideals. If they aren’t people, there has to be some defined way the thing could be destroyed. For example:

• A building could burn down. A secret lair could be exposed.
• An item could be stolen. A treasure could be destroyed.
• A loyalty could be betrayed. An ideal could be proven false.

Essentially, each bond must have a clear way it could be destroyed, killed, or otherwise rendered permanently unavailable to the character. The more ways that this could happen, the better (the point is that the GM is going to threaten them regularly, so making them only have a limited angle of attack will either make it repetitive or the GM will ignore it, making it worthless).

A player may only have five bonds at a time. If one is destroyed, a new one can be purchased with one experience point (multiplied by whatever value payouts are multiplied by, see below) and justification for why this thing is now important to the character.

Threatening Bonds

A GM will frequently threaten the bonds of the characters. Threatening a bond adds 1 experience point to it. The GM must follow several rules:

• A bond may only be threatened with sufficient warning that there’s a chance to save it (at least at the beginning of the scene where the bond is in danger).
• A GM may not destroy a bond without threatening it.
• If a GM threatens several of a character’s bonds at once (such that it is likely that saving one will doom the other without extreme success), he must pay an additional experience point per each bond threatened (e.g., if three are threatened, each bond threatened gets 3 exp placed on it).
• Bonds can only be threatened if the owner of the bond must take difficult action to save it. If the bond is not really in danger, such that the owner’s inaction would not result in its destruction, it is not worth an exp.

The GM should try to threaten at least one bond per player per session.

Destruction, Retirement, and Revenge

If a bond is destroyed, the player gains all the experience points currently placed on it. Essentially, the player must protect the bond at least once to do more than just recoup the cost of purchasing the bond, and protecting one several times creates greater exp profit.

The exp is multiplied by whatever factor makes sense for the system (e.g., if the system expects players to earn 10 exp per session, and the GM only plans to threaten an average of one bond per player per session, it should be multiplied by 10). This may be the game’s primary (or only) source of experience points.

If the bond gains a total of five or more exp, the player may choose to retire it with story justification. A bond to a character may mean that the character moves away from the area and out of danger, or just gets empowered sufficiently to no longer be in greater danger than the PC (sometimes, this just means informing your friend why he’s been targeted by all these crazy things recently). The character may leave the place to no longer keep it in danger, or just may somehow protect it so it’s no longer targeted. An item may be placed somewhere safe so it’s not in constant risk. A retired ideal means that the character has internalized it sufficiently that it’s no longer at risk of being disproved.

A retired bond gives half its exp value to the player, rounded down (i.e., you get paid more for the angst of loss than fully protecting the bond; a player that retires a bond has grown fond enough of it to sacrifice a bunch of exp to keep it safe). It usually leaves the story to live happily ever after. If there are brief visits from the bond later, it should never be in any particular danger unless the players choose to keep pulling it back in (or some other player decides to take it as a bond…).

When a bond is destroyed, instead of accepting the experience immediately, the player may choose to declare revenge. The bond changes to “Revenge for the [death/loss/etc.] of [the bond]” and cannot be replaced until the revenge is completed or abandoned. The player may abandon the revenge at any time and gain the original experience value of the bond.

For every session that the player character expends effort toward fulfilling the revenge (investigating to find the killer, paying back the killer in kind, etc.), the bond gains an additional experience point (to a maximum of double the original value of the bond). When the revenge is finally consummated (by killing or otherwise ruining the person or organization most responsible for the destruction of the bond), the bond is cleared and pays out its full accumulated value.

Non-Deadly Destruction

Players may specify bonds, particularly to people, in a way that means that death is not the only way to destroy them. Generally, this is something like maintaining the innocence/ignorance of the subject. Your friend finding out your secret (which will cause a permanent rift in the friendship), or your sibling being turned into a monster like you may be almost as terrible for you as being killed.

Large-Scale Threats

When a threat targets a region large enough that it might destroy multiple bonds, it doesn’t count as a threat to those bonds until there’s only a short time left to save them. For example, if the players find out several hours in advance that there’s going to be a city-wide death ritual, but they could just call loved ones and tell them to evacuate with plenty of time to spare, that’s not a threat worthy of an exp. If the players deliberately dawdle until there’s no way the bonds could escape without stopping the threat, or don’t even find out about it until it’s too late to escape, then it does count as a threat to all of the bonds (it might still not count for multiple exp on each bond, since saving one doesn’t necessarily make it harder to save another if one success averts the crisis). In general, GMs should be careful about large-scale threats (perhaps saving them for arc finales where the giant exp payout is intended).

As Aspects

These bonds can double as Aspects in a Fate game (and may replace them entirely). In that case, you can obviously invoke the Aspect when the bond is being threatened. All threats to a bond are also Compels, but the GM can Compel the bond without threatening it (for situations where the bond is in trouble, or will get the PC in trouble, without actually being in mortal danger).

Ongoing Flashbacks

Most advice for flashbacks in RPGs tries, unsurprisingly, to replicate how flashbacks are used in most media: as the occasional one scene that can appear whenever it’s relevant, or sometimes a whole episode devoted to explaining a crucial issue. However, pioneered by Lost (or at least that’s the first place I saw it) and now used in a slightly different way in Arrow, another option is the ongoing flashback, where up to half the time is set in the past. In Lost, this was a second story giving more background to a character whose choices were central to the episode, but each episode could have a completely different flashback and there was no particular order. Arrow, on the other hand, show something far more gamable: the flashbacks are in a linear order and are effectively a second ongoing plotline that happens to be in the past rather than another location. The past plotline tends to conveniently parallel whatever’s going on in the present thematically and introduces any facts and abilities the main character’s theoretically known all along but weren’t relevant until now.

This could be a huge having-your-cake-and-eating-it-too idea for games where the protagonists are meant to start ultra competent with minimal advancement featuring players that like to spend exp.

How I’d do this kind of thing (using Fate lingo, but could really work with anything skill-based) is:

• The players start in the present with a full pyramid of skills, but only a bare minimum of aspects, stunts, and powers. Effectively, their options are going to increase over time, but not their power level.
• The players start in the past with a greatly reduced pyramid of skills, and probably zero aspects, stunts, or powers. They’re going to learn everything in the flashbacks.
• When a player is ready to spend exp (or just get an increase on a fixed schedule), he or she tells the GM in advance of the session. Part of that session’s flashback involves picking up the new trait (it’s up to the player to justify why it never seemed relevant to use it until now).
• The characters in the flashback gain skill points at a fairly accelerated rate. If the player raises a skill in the past higher than that of the present version, it’s suggested that the player use the normal rules for flipping skill ranks to make sure they continue to match up.
• The player is free to use tricks he or she thinks will be on the final list in the present in situations where they don’t matter to the rules (e.g., to show off) to drive home the idea that the present character knows everything the character in the past knows, just hasn’t figured that it’s relevant yet.

The GM, in setting up these sessions, should do a few things:

• Plan the advancement path to parallel the course of the chronicle. Once the flashback versions of the PCs have the same skill pyramid as the present versions, it’s getting really close to time for the end of the past to become the beginning of the present, and wrap up the arc. This could be a complete finale, or just a timeskip to even more badass versions of the characters later that have new flashback moments.
• The events of the past storyline should be somewhat flexible in your mind, as they should stay thematically related to whatever is happening in the present. If the present winds up with the players going after someone that is theoretically an old foe, you want leeway to bend the flashbacks to show when they first met him. If something in the present is showcasing a failure of fatherhood, the flashbacks can call out one of the PCs’ own relationships with father figures.
• In the flashbacks, the PCs are obviously in no danger of dying (unless there’s room for a surprise reveal that one of them is a clone with the original’s memories or something). But you can raise the stakes by having a rousing cast of NPCs that the players would like to keep alive. You can even run whole flashback arcs that largely involve protecting an NPC, and if the NPC survives and the players liked her, she soon after appears in the present timeline showing up to help out and reward the players for helping her in the past. You might also build to threats in the present by having flashbacks focus on how much information they were learning in the past: a flashback failure may result in the players having less information and fewer assets in the fight against the present threat.
• Ideally, the PCs have been working together for some time (though you may start off the flashbacks with a “you all meet in a tavern” moment) so you don’t have to split the party in the flashbacks. If the story or character concepts absolutely demand that the PCs were mostly or entirely solo in the past, try flipping focal episodes. Each session, another PC’s past is what’s relevant to the present issue (and that’s the PC that gets to buy new stuff), and the other players are handed lightly sketched supporting NPCs to portray in the flashback. Make sure to give each player a roughly equal number of focal episodes.
• In an actual session, borrowing from TV act structures is a good idea. That is, be on the lookout for a surprise beat to flip between past and present scenes, particularly:
  • Something that might become more potent for being drawn out (“and then a bunch of guys with guns kick in the door… and… flashback”)
  • Something that is directly relevant to flashing back (“the assassin pulls off his hood to reveal… Captain Stone” “Who? Wait, the random captain who was piloting our plane? We don’t really know him.” “Flashback! On the plane to your destination, you hear over the intercom, ‘This is the captain. I’m getting some unexpected contacts on the radar. What did you people get me into!? Oh hell, missile lock, hold on…’”).
• Make sure your story is sufficiently about secrets revealed and tight-lipped protagonists that the whole mechanic continues to feel relevant. If you’re not sure it works for a whole campaign, consider just doing it for periodic one-off episodes where someone’s past is extremely relevant. This is a lot more like just the way every RPG suggests to do flashbacks, but at least alternating regularly between flashback and present between scenes preserves some of what’s different about this format.

 

Skill Based: Scaling Exp into Scaling Time

As I’ve mentioned before, I’m not a fan of skill-based game systems that use a flat-cost system in character creation but a scaling-cost system for advancement. That is, something that costs 2 points per level in creation might cost current level x 2 with exp. It effectively penalizes players for not min-maxing at character creation, as it’s way cheaper in the long run if you start with several useful traits maxed out and neglect median ranks in traits.

What the scaling costs are meant to do is create a more simulationist curve to advancement: of course it’s harder to master a skill than to pick up the basics, so the first rank costs a small fraction of moving from the penultimate to ultimate rank. But even with that stated goal, in the latest White Wolf game I’ve been playing, we’ve started to notice flaws. We have a five-member party where everyone started out with a high degree of skill in at least a couple of areas that compliment one another, so we can usually field at least one person against any given challenge. Since there’s only the lowest pressure to shore up our weaker traits, it only makes sense to save up our exp until we can buy something really flashy (generally the next highest rank in our powers). The GM is stuck in a weird situation where our group can pretty easily roll over most of the opponents in his setting book, because even though those NPCs technically have tons more exp than us, they’ve diversified it across a bunch of traits. We’re magical idiot savants, fantastically skilled at a couple of meaningful areas and worthless at anything else. And the system makes that a great idea.

And, as a secondary concern (which is a problem with virtually every game that doesn’t require training time), we’ve gone from newly awakened nobodies to magical powerhouses in only a couple months of game time. There’s no reason for us to take a break longer than it takes us to heal up, recover mana, and refill willpower. So even though the GM would like us to spend some time in magical study, it doesn’t really make any sense for us to do so.

That’s a long intro to explain the background of the system below. It’s designed to:

• Minimize the differences between buying something at chargen and during play
• Encourage players to diversify spending rather than just buying the flashiest traits
• Enforce a “realistic” time frame on learning skills

Fixed Costs

The following charts are tuned to new WoD (new level) and old WoD (current level), but should be applicable with minor modification to any game that uses a level multiplier for exp costs.

New WoD

New Level x	5 dots*	10 dots
1	3	6
2	5	11
3	8	17
4	10	22
5	13	28
6	15	33
7	18	39
8	20	44

* The fifth dot costs double as in character creation.

Old WoD

Current Level x*	5 dots	10 dots
1	2	5
2	5	9
3	7	14
4	9	19
5	12	23
6	14	28
7	16	33
8	18	37

* The costs are based on a fixed cost for the first dot equal to about 150% the cost of the second dot (e.g., skills cost 3 points for the first dot in oWoD and 2 for the second).

Examples

In oWoD, attributes cost current level x 4 for a five-dot progression. Looking at the chart above, they now costs 9 points per dot. Similarly, Willpower costs current level x 1 for a ten-dot progression. On the chart, that becomes 5 points per dot.

In nWoD, attributes cost new level x 5 for a five-dot progression, so they now cost 13 points per dot (and the fifth dot costs 26). Meanwhile, Willpower becomes a flat 8 points per dot, so that remains unchanged.

Enforced Time

Standard Method

Characters can spend one exp per week per trait. If they don’t have enough saved exp on hand at the end of the week for all the traits they want, they didn’t learn anything that week and don’t get to “buy the week back” when more exp is gained. But if they have lots of exp, they can be working on several traits at once. This is effectively paying for the trait on layaway: when the last point of exp goes into the trait, it is increased on the character sheet immediately.

The GM may additionally want to give out bonus training that is effectively extra exp that can ignore the time restrictions. This will generally be something the PCs were focused on during the adventure and could justify learning faster due to on-the-job training. For example, in an adventure where everyone learned a ton about the occult, instead of 4 general exp the GM might give out 2 general exp and 2 exp that went straight into the Occult skill (and the players could spend another point of exp into Occult for their regular weekly increases).

Players will likely either want a character sheet with room next to every trait to track exp spent, or a scratch sheet to keep track of which traits are being worked on. The GM will likely want to pick a day of the game week that’s exp day, and remember to call it out at the table (“It’s Sunday morning, spend your exp!”).

Slightly More Bookkeeping Method

The above method does make more expensive traits take longer to learn, but doesn’t capture the geometric feel of the multiplicative exp. That is, under this system, the fourth dot takes just as long to learn as the second. If you’d like to retain some of that feel, you can make lower point values take less time to pick up than higher.

The simplest way to do that is to let the players put two points of exp into a skill per week if they’re trying to buy rank 1 or 2 and only half a point in per week (or one every other week) if they’re trying to buy rank 5.

In nWoD, that means a character completely untrained in a skill (and with ample exp to spend every week but no bonus exp) gains rank 1 in four weeks, rank 2 after 8 weeks total, rank 3 after 16 weeks total, rank 4 after 24 weeks total, and rank 5 after 56 weeks total. Meanwhile, a x6 power gains rank 1 in 8 weeks, rank 2 in 15 weeks total, rank 3 in 30 weeks total, rank 4 in 45 weeks total, and rank 5 after 105 weeks.

If you have enough downtime that a year still seems too fast to go from untrained to mastery in a skill, you can slow the progression down even more. Do keep in mind that doing so will make learning more expensive traits than skills (i.e., attributes and powers) take even more time. Make sure that your time progression leaves enough room for players to bother getting the last rank in a power if it takes over a year to go from 4 to 5.

And if, as a GM, you want to keep track of all the bookkeeping yourself, you can simply ask the players what they’re trying to learn and then tell them when they can level it up. Just one day you’ll be like, “all that time you spent learning X has paid off, you may increase it by one dot.” Your players might even be amazed at the way their characters grow along with their intentions but without their direct involvement.

Flashbacks vs. Exp Progression

As last week’s post might indicate, I’ve been playing a lot of Assassin’s Creed lately. The game itself is, of course, entirely framed as a flashback as Desmond explores ancestral memories. However, even within this framework, the game has introduced the concept that it’s glossing over certain memories until they become relevant: it will sometimes feature sequences much earlier than the current date of the main narrative (e.g., the story of what happened to Ezio’s girlfriend who was in his introductory scene then never mentioned again). This is, of course, not that unusual for fiction: TV shows, in particular, have time to periodically do long-form flashbacks to give context to the current situation (and if your show’s about immortals, they might do that every week…).

Despite their utility in other media, long-form flashbacks are really hard to do in tabletop games. A lot of games have experimented with short-form flashbacks, in the sense of player-directed 30-seconds-or-less statements to get a bonus on a roll, but asking players to spend the bulk of a session (or more!) in a flashback has a couple of hurdles:

• If the game features exp progression, it’s often really hard to figure out how to use stats from earlier in the characters’ careers without accurate record keeping.
• Many players aren’t going to want to play their characters earlier in their careers (i.e., with worse stats) for extended periods.

Obviously games that eschew progression are going to have an easier time of this, but my group’s never been happy without numbers that go up, and I’m sure a lot of groups are in a similar boat. And, besides, an interesting vignette about something that was previously glossed over has just as much place in D&D as it does in SotC. So here are a few ideas on how to do that in a way your players might enjoy:

• Get in the habit of giving out and taking away cool things (powers, gadgets, etc.). Take away some of them in downtime with the explanation that you’ll eventually run the flashback explaining it. Periodically introduce cool things that have been mostly used up (such that they were much cooler at full power) with the explanation that you’ll explain how they got it in an eventual flashback. The goal here is for players to look forward to flashbacks where they’re potentially less powerful in stats for the joy of playing with toys that they don’t otherwise have access to.
• Since you’re primarily running long-form flashbacks when they’ll provide some kind of expanded context for the current situation, include a mechanic in the flashback to provide benefit in the present based on success. For example, “Let’s flash back to the time you were trying to gain treaties from the barbarian tribes which would now be a big help against the Dark Lord.” This creates resources that the players have theoretically had access to for a while, but which weren’t relevant (and, thus, weren’t quantified) until the flashback.
• Keep flashbacks extremely rules light and just let the players succeed without rolling on just about everything within reason that they try. After all, they clearly survived the events none the worse for wear such that it wasn’t even relevant until now, so might as well let them have unusual success. Don’t even reference stats unless you have to. The session should mostly resolve around revealing more information about something, possibly even something that the PCs didn’t realize until now that they hadn’t shared with one another (possibly because it happened before they even met; see Leverage’s The Rashomon Job).
• Just let them use their up-to-date stats and get exp normally. Gloss over how a player is using a power he just got last session in a flashback set years ago as imperfect recollection.
• Everyone had AMNESIA! And didn’t know it! Until NOW! (Don’t actually do that, it’s overused.)

Anyone else have tricks for running long-form flashbacks that players are happy to participate in?

Achievement-Based Leveling

Harbinger returned my attention to achievement-based advancement as mentioned in one of my previous posts. Obviously, my original concept was designed for a video game that could track all the achievements for you, but with the correct phrasing of achievements you could allow your players to take on the role of the computer: there’s a ton of bookkeeping potentially involved, but it’s the fun kind of bookkeeping where players do it for you and it’s in their own self interest to keep on top of it.

Effectively, you’d need all the possible achievements typed up and printed out (potentially on several sheets of paper), kept with the character sheet, and checked off by the player as they’re fulfilled. You’d probably also want to have a blank line next to each achievement for the player to summarize when it was achieved in case the GM has doubts.

A lot of inspiration for this is taken from the Dungeon World variant of Apocalypse World.

System

The core intention of the system is to provide players lots of directed goals, accomplishing any of which will support the concept and intended playstyle of the game. Additionally, the intent is to reduce grind/repetition: each achievement can only be gained once, so repeating the same action has little further system benefit (unless there is a more difficult version of the achievement that incorporates the same actions as the easier version). To that end:

• Each player character has a list of possible achievements. These are short phrases with a method of accomplishment that requires minimal interpretation by the GM.
• Achievements are divided into groups based on similar theme. These groups might vary based on the kinds of actions incentivized by the current game and campaign.
• All achievements completed in a category are totaled, as well as a global total of achievements completed. Achievements vary in difficulty to accomplish, but each contributes the same amount to these totals: there will be low hanging fruit that players can get early on, and harder achievements that will serve as higher level goals.
• The total achievements are used to allow characters to level and to determine what classes they have access to when leveling.

Note: Any numbers below are purely arbitrary and used for illustration. Actual numbers used will depend on how many achievements you come up with and how fast you want the players to level. A game with a ton of achievements that expects several sessions between levels will need higher numbers than a game with fewer achievements and a faster leveling pace.

Level Up

A character’s total level is determined by total achievements across all categories. When a character accumulates the requisite number of total achievements for each level, he or she is eligible to level up. This level up can happen instantly, at the end of a session, or after training (however the GM would normally award a level up).

For example:

1. 0 total achievements
2. 10 total achievements
3. 20 total achievements
4. 30 total achievements
5. etc.

Note that the achievement total is linear: in theory, as you level up you’ll run out of low hanging fruit and be left with the harder achievements, keeping the expected time between levels that normally is accomplished by a curved exp-to-level requirement.

Class Requirements

This system expects D&D 3-style multiclassing; some modifications should be required for versions where you’re less free to multiclass. When you level up, you can choose to take a level in a class that you meet the sub-requirements of. Level 1 in any class has no requirements: you can always multiclass into the first level of a new class. The subsequent levels have increasing requirements in specific achievement totals within different groups appropriate to the class.

• Barbarian: Combat and (Neutral + Chaotic)
• Bard: Combat, Spellcasting, and Lore
• Cleric: Spellcasting and (Deity’s Alignment)
• Druid: Spellcasting and Neutral
• Fighter: Combat and Adventuring
• Monk: Combat, Lore, and Lawful
• Paladin: Combat and (Lawful + Good)
• Ranger: Combat and Defeat
• Rogue: Adventuring and Neutral
• Sorcerer: Spellcasting and Adventuring
• Wizard: Spellcasting and Lore

For example, a Paladin’s requirements might be something like:

1. None
2. Combat 2 and (Lawful + Good) 1
3. Combat 4 and (Lawful + Good) 2
4. Combat 6 and (Lawful + Good) 3
5. Combat 8 and (Lawful + Good) 4
6. Combat 10 and (Lawful + Good) 5
7. etc.

Note how this interacts with leveling up: a single-classed Paladin with 30 total achievements is qualified to level up into level 4… but might not have gained the requisite Combat or alignment achievements since the last level to take the next level of Paladin. In that case, the player would need to choose whether to hold off on leveling until the necessary achievements are completed, or to multiclass.

Example Achievements

All achievements are both things that characters might describe as learning experiences in character and also actions that the GM believes supports the style and goals of the campaign.

Quest

Each completed quest in the game counts as an achievement. These are not required by any particular class, but contribute toward leveling up. It’s the GM’s choice (based on overall number and difficulty of achievements) whether only large quests count as an achievement or even smaller goals might count. In the former case, players will often have the same number of Quest achievements, but in the latter there might be divergence due to personal quests.

• Quenched the Black Flame of the Boneyard
• Recovered the Princess of the Platinum Lands
• Slew the Dragon Gygyragax

Combat

Combat achievements are devoted to doing interesting and dangerous things in physical combat. They are typically required by non-casting classes.

• One-shotted a 1 HD creature (full HP to unconscious/dead) with a melee attack
• Was attacked and missed by 3 different enemies in a single round
• Hit a target with a ranged weapon at its maximum range increment

Alignment

Alignment achievements are subdivided into Good, Evil, Lawful, Chaotic, and Neutral. They are required for several classes with alignment restrictions.

Note that a GM might choose to determine a character’s alignment based on achievements: Subtract the lower of Good and Evil from the higher, and then subtract Neutral; if the number is positive, the higher alignment is the character’s; if it’s 0 or less, the character is Neutral on that axis. Do the same for Lawful and Chaotic. For example, a character has achievement totals of Good 5, Evil 1, Lawful 1, Chaotic 0, Neutral 2; this character is Neutral Good.

Good

• Took damage while defending an innocent
• Accepted the surrender of a repentant foe

Evil

• Sacrificed an innocent to gain another achievement
• Killed a surrendered and bound foe

Lawful

• Delivered a defeated criminal to the rightful authorities
• Accomplished a city-based quest achievement without breaking any laws

Chaotic

• Set a prisoner free
• Accomplished a city-based quest while breaking laws without being caught by the authorities

Neutral

• Killed a dangerous and untrustworthy foe when taking it prisoner was an option
• Allowed an innocent to die to accomplish a greater good

Spellcasting

Spellcasting achievements are based around doing interesting things with magic, and are required for most casting classes.

• Used a 1st level attack spell to defeat a target
• Used a 1st level attack spell to accomplish a non-combat goal
• Used a 1st level defense, heal, or utility spell to defeat a target

Lore

Lore achievements are based around learning more about the setting, and may often be similar to more general Quest achievements. They are required for certain educated classes.

• Found a rare tome in a monster’s horde
• Spent 100 GP at one time to buy books
• Translated a warning in a dungeon

Adventuring

Adventuring achievements are more general actions appropriate to dungeon-delvers, and are required for classes of a more mercenary bent.

• Possessed 1000 GP worth of cash and gems at one time
• Disabled a level 1 trap (either intentionally or by blundering into it)
• Defeated a 1 HD enemy during a surprise round

Defeat

Defeat achievements are triggered based on getting in the blow that drops a creature unconscious/dead or by being the primary negotiator that convinces an enemy to surrender. As such, each defeated creature generally only contributes to one player’s total. They are used for hunting-related classes.

• Defeat 5 goblinoids
• Defeat 15 goblinoids (10 more)
• Defeat 30 goblinoids (15 more)

To End an Era of Exp, Conclusion

Over the past three installments, I’ve argued that experience points are outdated and no longer solve the problems they did originally. Meanwhile, I believe they’ve been frequently misused by becoming a staple of RPGs where they often lead to an unnecessary psychological reinforcement. In theory, games are fun without having to reduce them to a single axis of progress, and doing so can cause players to fixate on exp accumulation rather than what’s actually cool in the game.

So, assuming you’re with me this far, here are a few of the potential solutions to remove exp but still include advancement. They are likely to be more useful for video RPGs, particularly MMORPGs, but may still be useful in some way for tabletop games.

Time-Based

Eve Online is one of the few games of which I’m aware that have eliminated the ability to grind for progress by the simple mechanism of divorcing skill progression entirely from game actions. Instead, players designate which areas they’d like to improve, and timers begin counting down until that skill level is learned. I believe there may be some degree of requirement to do a few in game tasks to open up a new tier of skills, but that may just have been my limited understanding of the very complicated systems for the little time I played.

Reducing improvement to a pure time-based system is pretty much anathema to grind, at least for advancement. When you’ll improve at the same rate no matter what you do (and even if you don’t even log in), in theory there’s no reason to do anything besides what’s fun. There’s at least nothing boring to do that might have a better exp rate.

Conversely, time-based progression can create a game where late-comers can have a hard time competing with earlier players, as there’s no way to bridge a skill gap other than an earlier player forgetting to set the next skill to train. In the real world, this is solved by older generations making way for the younger, but such a thing is unattainable in video games until and unless permadeath suddenly comes back into vogue. Effectively, it may make sense to supplement a time-based system with something else that doesn’t discourage new players and alts.

Use-Based

A game with use-based skills gives a chance to increase a skill whenever it is used (or at least whenever it is used in appropriately difficult situations). Many video games include some kind of use-based system for skills. WoW has use-based weapon skills (or did the last time I played), and EverQuest had athletics-related scores that went up as you used those modes of movement.

Strangely, most use-based games also have an experience system that leads to actually leveling up. WoW weapons skills were mostly superfluous, and just a way to force more grinding: you’d only really be much under the level cap on weapons that you never used, and would have to spend a couple of hours training if you got a good magic weapon of an unused type. A lot of this is ease of balance: modern games gain a lot of mileage by being able to assume that a player of level X can do Y. And some games that go entirely use-based, like Oblivion, can generate some odd behaviors (in Oblivion, it’s often better to create a custom class that has skills that are easy to raise but you don’t plan to use; that way, you can game the level-up system by only leveling when you want to).

Despite making balancing harder, there’s still likely some mileage in use-based skills, particularly in a game that’s not rigidly class-and-levels-based. Specifically, if you can come up with a good way to make skills improve faster when the player is actually being challenged, and tie raising particularly useful skills (like defenses and mez resistance) to more interesting foes, you can avoid the grind while having a system that isn’t tremendously different from exp. The challenge-detection algorithms would have to be pretty good, and constantly subject to exploit-detection, of course.

You could even hybridize such a system with a time-based one, granting reduced time to improve a skill rather than a direct skill up.

Achievement-Based

Nearly every modern RPG, especially every modern MMO, has some form of achievement system. Do a certain countable thing enough, or do an unusual or difficult thing once, and the game can track it and award an achievement. For most games, these are largely for bragging rights, though some do grant decent bonuses for achievements (or collections of them).

What if constellations of achievements replaced experience for leveling? For level 2, you need three easy quest achievements, one medium difficulty quest achievement, two exploration achievements, and one first tier monster-killing achievement. You’d see them on your level-up screen: “oh, I need one more exploration achievement to level, let’s go explore!” Maybe you’d have to move on after killing the easiest group of monsters sufficiently for an achievement, or maybe subsequent achievements for that group would just get harder (“Achievement 1: 20 Total Goblins, Achievement 2: 60 Total Goblins, Achievement 3: 120 Total Goblins. Man, let’s just go kill zombies instead, I only need one kill achievement to level, and I don’t have any zombie ones yet.”).

The essential idea is that achievements allow you to granularly toggle rewards to what’s theoretically fun about your game. Exp is a lowest-common-denominator; you can’t make players do other things if they’ve found one supremely efficient source of exp. But if leveling required a variety of tasks at any time, it might be easier to convince players to try them.

This isn’t a mechanism for the faint of heart: it requires actually having a good idea of what’s fun about your game, and not trying to use the system to force players to go after filler. It also might require some multiple choices to avoid alienating different classes of player (“One exploration achievement or one PvP achievement” might be a good way to catch two different player agendas, for example). But, at the end of the day, you’re going to design an achievement system anyway, and a large number of your players will want to pursue it, so there’s no reason not to get some practical use out of it as well.

Final Thoughts

Experience is an easy fallback for designers. It’s so pervasive that most designers probably don’t even consider doing something else. Even if you do, that something else is going to require a ton of balance to prevent exploitation, and it’s probably an easier sell to just deliver a small variation on the traditional experience point.

But experience already demands a ton of balance, it’s just not getting it. It lies at the root of a host of player behaviors at odds with what should, theoretically, be foremost on a designer’s mind: delivering the fun. It, by its nature, steals the focus of whatever you’re doing and pulls it down to an accountant’s zeal for leveling up.

None of the ideas above may ultimately be ideas that survive vetting at the hands of a cunning playerbase, but I’m firmly of the opinion that trying something is better than using the same old mechanic without question. Nearly every other remnant of OD&D has been gradually replaced over the last 30 years of game evolution. Some of what was lost was of much more debatable value than experience points.

Why are they the last sacred cow to fall?

To End an Era of Exp, Part 3

Experience points started off the hobby as a necessary (or at least convenient) tool for allowing characters to grow. But the evolution of systems has trended towards the more granular and the evolution of game play has trended towards more coherent stories. It’s questionable why many games still rely on exp for advancement, but first we need to understand the power of advancement itself.

Conditioning

Back in 2005, Dan Bayn wrote a very persuasive article about Advancement Addiction. To sum up: psychological conditioning theories potentially indicate that linking character rewards to game play makes players increasingly crave these rewards and associate them with fun. Experience (and advancement in general) isn’t just a convenient counter to justify a character gradually growing more competent, but is a powerful psychological driver in itself. Players of games with any form of classic advancement begin to confuse the advancement with the fun. Try an experiment at your game table:

1. Run an all-out, dramatically exciting game session that hits stated character in-story goals, but claim that kind of session doesn’t count for advancement rewards.
2. Run a pretty slow and mundane session (maybe a combat slog with boring foes), but give out a bunch of advancement afterwards.
3. Get your players to give their honest assessment of perceived fun and which type of session they’d like to see more of.
4. Now do the same thing again but flip which session gives rewards and see how that changes opinions.

Clearly, you’re hoping your players are cool enough to realize the trap you’re setting and admit that the better game was better. But even excellent players having an excellent time still miss the thrill of building up their characters, as Harbinger of Doom mentions in his Spirit of the Century actual play writeup. There’s anecdotal evidence that players that primarily focus on indie games without much in the way of advancement can get over the urge, but I think it’s pretty reasonable to suggest that most players of classic RPGs are conditioned to expect rewards in the form of our characters kicking more ass the more we play. Somehow it doesn’t feel fun to have a character that doesn’t get more powerful, even if the character started out with tremendous power to start with (start a Nobilis game with double the suggested character build points and then count the sessions until the players start asking when they can make their literal demigods more powerful).

So advancement addiction is a hard thing to overcome, and you may not even want to. But why does exp make everything worse?

Reinforcement Schedules

Experience, by its nature, is designed to make advancement more granular. You don’t get a level or even just a skill point, but a collection of experience points that can be traded for such once a certain accumulation is reached. Experience is generally, by virtue of being so granular, designed to be awarded at variable rates. In classic use, you defeat a goblin for X exp and an orc for Y. In more story-based games, you get X for a slow session and Y for a session where a lot was accomplished. Experience awards explicitly reinforce doing the stuff the game/GM wants you to do vs. doing other stuff (potentially more fun stuff, if experience is not correctly synched).

The classic example (exp for monster defeats) is the one that frames all understanding of exp awards. Awarding experience for more difficult defeats implicitly creates a world in which defeating challenging opponents is the most valuable teaching tool. This is not necessarily wrong, but leads to all player characters essentially being violent thrill seekers each looking for the next challenging fight. Sure, they prioritize threats for all the classic reasons—stealing treasure, defending innocents, pursuing a villain—but the classic style means that it’s a completely valid choice to rough up targets that could be just as easily left alone (i.e., who runs from random encounters?). Classic experience awards incentivize making every PC a bully.

Variable rewards certainly have their uses, of course. Sometimes a GM has to rely on the exp card to get players to tackle a challenge in a way that’s probably more fun for everyone (e.g., face the tunnel of challenges or hire a pack of sherpas to guide you safely to the other side of the mountain?). It provides a carrot for players when there’s no reasonable in-story justification for doing things (always remember. though, that the stick version of the same behavior is often called “Railroading”).

Ultimately, however, most GMs cheat, especially when there is a big change in power with each level. If the next big part of your campaign involves a lot of challenges appropriate to 5th level characters, you probably don’t want to let the characters get much past or short of 5th before they get there. So monsters get added if the players are falling behind and taken away if they’re getting far enough ahead, all based on some kind of dogmatic interpretation of how many encounters the game designers thought was appropriate between each level.

Module series are forced to do this shamelessly: if Module 1 starts at 1st level and Module 2 expects the characters to be around 4th level, you can count up the exp awards from every fight in the book and see that the module writer made sure you’d get pretty much exactly to the right spot for Module 2. If your players skip fights, you’ll need to make up the exp somewhere. If you add any improvisation on your own part, you may want to scale back something from the official text. In many ways, the GM is trapped using a mathematical system to attempt to justify “this is a 4th level adventure so I want you to be 4th level now.”

Maybe you buy into the math, and you carefully arrange it all to add up. I know I frequently just start going holistic: “it feels like it makes sense for you to level up now.” The GM may or may not continue the fiction of experience points and their relationship to character advancement, but, in a tabletop game, it’s actually not really that big of a deal. A competent GM will make sure to lay out challenges that feel reasonable and give out experience or just pure advancement when it seems appropriate. In a GM-run game, I feel like experience points are generally an unnecessary complication for an already worrisome addiction to advancement, but it ultimately doesn’t matter because the GM can compensate to make the fun and the advancement continue to line up.

The real problem is that video games don’t have a GM.

Skinner Box

To paraphrase a quote of forgotten attribution, “Gary Gygax doesn’t have to let you keep fighting goblins if it’s not fun anymore, even if they’re still worth exp.” When a GM is running a game, it’s very unlikely that the players will be able to experiment fighting with a bunch of different monsters and then choose to only fight the monsters with the best risk to reward ratio until they level up and then start going after bigger targets with an even better risk to reward ratio. You can’t decide goblins are an easier fight than kobolds and then only fight goblins until you feel capable of moving on to the easier of ogres and worgs.

In video games, players do this all the time.

As mentioned in the first post in this series, “adding RPG elements” to a video game almost always includes exp. Computers are great at counting things. If a goblin is worth 1 exp, a computer can give you an accurate exp count after going through a whole warren of goblins without missing a beat. Theoretically, computers are also way more impartial than a GM, so the exp numbers that would likely get fudged in a tabletop game can be rock solid. Hell, there’s often a benefit for letting a player keep fighting stuff at the current level until he wants more challenge and moves on.

But the problem is that computer games automate the problem too well. A variable reinforcement conditioning that is debatable behind a tabletop game is so obvious as to be a running joke about video games, especially MMOs like World of Warcraft. Click the attack button a few times and you get a reward! Click it enough and you get a big prize! It’s very clearly a Skinner Box, and a lot of that is due not just to rigid interpretation of the awarding of exp, but due to the capacity to choose targets.

That’s right, we’re talking about grind (and we have been this entire series!).

Just to pick an example that fits my MMO of choice at the moment, City of Heroes features a wide variety of high-level enemy groups. One group uses a lot of crowd control and stays at range. Another has lieutenants that are completely impervious half the time and all of them drain endurance when they die. Still another turns invisible and always drops piles of slowing effects at your feet until you can barely move. Then there’s a group that does basically normal damage, automatically clusters up for area effect attacks if you’ll let them, and their major trick is that, when you defeat them, they sometimes resurrect themselves with less powers so you can defeat them for exp again. I’ve never really been in a pickup group where the members wanted to fight anything else but Freakshow if they could help it.

As a designer of video RPGs, particularly MMORPGs, the use of experience points has the subtle and invasive effect of causing your player behavior to flow like water seeking the lowest point: an unfortunate mass of your players will eventually hit on the behavior with the best time or risk to reward ratio in the game, and perform that behavior far more often than the fun inherent in the behavior supports. They can’t help it, everything in the game told them that leveling up is fun, and, in fact, implied that it’s the whole point. They’ll murder Freakshow, Earth Revenants, or your game’s grind-mob du jour for hours and hate you for how boring it is, while hundreds of more interesting fights await them. With exp in play, you can’t make players participate in things because it’s more fun unless you also make it worth a better exp ratio. Sure, you can write quests, but that just shifts the goalpost. As soon as you have one category of quest that gives better rewards no matter what it’s about, watch the players do that category more.

At the end of the day, if your game has exp your player has a number on his screen (perhaps represented as a colored bar). You have taught that player not only that making that number go higher is fun, but that it is implicitly the point of your game. It doesn’t matter how well written and scripted your quest or how interesting the AI is on your monster, its success and failure will really come down to whether there’s anything else comparable that makes the exp number go up faster. The only solution is to discard exp entirely.

And if you’ve followed me along this far and are now screaming, “oh yeah, even if I buy that it’s a good idea to kick out one of the core features of every RPG video game, what would I do instead!?” then I ask for only one more week of your patience. Next week, in the final installment of this series, I’ll talk about things you might do instead to replace exp with something less conducive to grind.

To End an Era of Exp, Part 2

Last week, I gave my (potentially ill-informed) explanation of why the concept of experience points entered the hobby at its earliest moments. Exp was simply the best solution for the design goals at the time. But times quickly began to change…

The Rise of Story

Anecdotes tend to point to Dragonlance as one of the (if not the) first official module series that included a story. Players had likely begun playing that way long before, but modules had always been much more focused on simply providing a venue for play: here is a dungeon, here is how your players can get into it, here are things they will find when they explore it. It didn’t need much of a story to be useful, as it provided an interesting place to fight bad guys who had treasure.

Dragonlance was a departure. It wasn’t just a setting that had novel tie-ins featuring stories set in the world. It was a module series specifically intended to make playing through the novels possible. The Dragonlance novels are a well-regarded fantasy epic, and you could produce a reasonable approximation of the events within simply by playing through the modules. Within the modules there were villains, and countdown timers, and quests, and other techniques that were fairly uncommon at the time to mold the play experience into the feel of a fantasy epic.

The modules did very well and became the model for most subsequent printed adventures. Quickly, the official examples of what constituted an RPG session became less Howard and more Tolkien: treasure and glory more and more became side effects of the pursuit of much grander goals. The PCs weren’t just trying to get rich and level up, they were trying to save the world, or at least accomplish much more character-specific long term goals than wealth and power.

By the 1990s, this had become so much the default method of play that a series of games that flat out replaced the moniker of “dungeon master” with “storyteller” became the new hotness in RPG circles. Most games abandoned the concept of levels entirely, instead relying on more granular improvement that was easier to map to non-game experience: in fantasy stories, characters rarely improve all their skills at once (save perhaps between novels). In fact, the core assumptions of the hobby had changed pretty drastically:

• Advancement was largely granular, increasing stats individually
• There was a renewed effort to balance characters mechanically against one another (even, perhaps especially, in games that still retained levels)
• PCs in-game tracked progress by their accomplishment of goals set before them, be they personal or for the good of civilization; characters might even come out of chargen with all the wealth and glory they’d ever want
• Death became less and less common, as allowing a character to die unexpectedly would ruin all the plot threads the GM had invented to tie that character and his or her goals to the world and story

But, despite these shifted assumptions, almost everyone still used exp…

To End an Era of Exp, Part 1

There is no rules concept more thoroughly tied to RPGs than Experience Points. They’ve been with the hobby since the earliest editions of D&D and are the first thing added to give “RPG-elements” to a video game that didn’t have them before. If your RPG features character advancement at all, chances are that advancement is tied to some variation of exp.

Why?

In the beginning…

Anyone running a game blog focused on the return to old school D&D can tell you that the earliest forms of gaming don’t really match the assumptions of most modern games. Whether or not that’s a failing, it’s pretty obvious that some things have become significantly different as the hobby has evolved.

The first editions of D&D (and experiments that led to it such as Braunstein) were iterations of the concept of simply zooming in on the individual units in a wargame and playing them rather than acting as their commander. That is, the first RPG characters were operating on game engines modified from wargames. This is another thing very obvious to those that have been in the hobby a long time: anecdotes indicate that the earliest versions of D&D relied heavily on players also owning a copy of TSR’s wargame, Chainmail, as a reference.

I don’t have sufficient knowledge of Chainmail to speculate on whether it used an exp mechanic, and would welcome any input on that in the comments. Unit advancement in early computer war sims leads me to believe it must have had some impartial way to advance, or at least differentiate, units from novice, to trained, to veteran. This is an abstraction that works very well for wargames: it’s pretty much impossible to track the individual capabilities of units when you’re managing dozens, so it makes sense to divide them into fixed tiers of competency.

It was only natural that the earliest RPGs did things the same way, based as they were on wargame rules. Most early systems were heavily level-based, and D&D even used titles for levels clearly defining the gradual improvement of competency tiers. The question was how to discretely allow characters to progress from one level to the next.

An interesting facet of level-based games that didn’t really even go away until D&D’s third edition was that character classes weren’t explicitly balanced against one another. In a straight up fight, a fighter was simply better than a thief, and a mage would eventually become powerful enough to exceed them both. There needed to be some mechanism by which balance could be achieved, and it made sense to do that by speed of gaining levels. A class that was half as capable as another would level up twice as fast, thereby using higher level to compensate for less effectiveness at the same level.

Finally, old school play was dominated by the concept of the adventurer in the style of Conan or Fafhrd and the Grey Mouser: a charming rogue that used wit and combat savvy to improve his place in the world. These were not superheroes, gamely laying down their lives for the good of a community, but whirlwinds of disaster separated from villains mostly by a preference for preying on other predators rather than upon the innocent.

This ethos expressed itself in gaming revolving around the dungeon crawl: your character’s primary goal was the discovery and robbery of hidden places guarded by creatures whose deaths no one in civilized society would mourn. Certainly, these warrens of monsters were often a significant threat to the surrounding humans and demi-humans, but the player character’s primary motivation was treasure and glory. Life as an adventurer was difficult and deadly: the monsters were going to do everything in their power to beat you, and you were going to do everything in yours to ensure an unfair fight in your favor, or to bypass the monsters entirely unless combat was absolutely required to take their treasure.

All these assumptions cascaded into the need for experience points:

• Advancement pinned to across-the-board upgrades with levels
• Classes not balanced in capability against other classes at the same levels
• Roguish and often competitive PCs seeking to minimize effort and maximize wealth and glory
• A high chance of dying and having to start a new character fresh

With these four inputs, exp simply made sense as a concept. It provided a somewhat objective method to determine when character could advance in level. It allowed weaker classes to advance faster than stronger ones. It gave players a carrot for clever play by allowing them to improve faster. It gave players a stick for incautious play by providing something that death would take away (either via resurrection penalty or simply starting a new character at nothing). It was an excellent fit for the needs of the era.

But those four assumptions have less and less of a place in RPGs every year…

Levels vs. Verisimilitude

The concept of increasing in level has long been a boon to progress-based RPGs, as it creates a potentially unlimited path of advancement for player characters. And players love to advance their characters. RPGs without levels typically have to pin their advancement scheme to a fixed scale, often striving for realism. Barring hideously expensive powers generally not expected to be possessed by the populace at large, it may be hard to keep a determined player from becoming amongst the best in the world at a specialty very quickly, possibly even at character generation. For some RPGs, this is a boon, but others are built around a continually ramping series of challenges. For these games, basing advancement on levels makes balance and maintaining player interest far easier.

But it makes world building far harder.

The hallmark of levels is how much of a difference there is between one level and the next: a progressive ramping up of power. It’s difficult to make a level-based game where a several-level power difference doesn’t make the lower-level character completely ineffective in most challenges vs. the higher level character. It’s hard to make these settings “gritty” or even marginally realistic; just in the realm of combat, a kid with a gun or a punk with a knife very quickly becomes completely negligible in a way he wouldn’t in the real world. And the real world (or at least Hollywood’s interpretation of it) is how players expect the game world to work intuitively until it’s proven otherwise.

This intuition is less of a problem for actual play; players can very quickly adjust to being superheroes. But it makes world building problematic: a designer’s default assumption is to create a world space that just doesn’t make any logical sense if there is a multiple-order-of-magnitude difference in capability across the spread of levels in the world. High level characters and creatures are so much more powerful than low level ones that it very quickly makes little sense if they don’t completely dominate the socio-political landscape of the setting, creating a massive base of the population that will literally never do anything of historical note if they don’t level up. Even attempting to mandate that high level characters are rare is a very hard assertion to make when having to design adventures that continue to challenge leveling heroes: a single high level mook could live a life of power and prestige dominating a lower level region rather than waiting to serve as a speedbump for high level adventurers going after his boss.

Massively multiplayer online RPGs have this problem worst. Tabletop GMs and the designers of single-player computer RPGs can more easily contrive situations that less arbitrarily segregate power levels in the game world. Sealed underground warrens, extraplanar threats, and a constant pressure to deal with the next big challenge can keep the goalposts moving long enough to keep suspension of disbelief afloat. And, at great need, a high-level PC can be allowed to journey back and realize the massive power now wielded over most of the cultural landscape. But MMOs can’t do this as easily. Gameplay needs to funnel players from zone to zone where threats increasingly ramp up, resulting in huge fields of high level creatures and NPCs waiting for the slaughter. Balance and protecting against jerks requires that safe areas include lots of friendly, high-level NPCs. The world can very easily become an inexplicably ramping heatmap of levels along the optimal player path.

So, how do you solve the problem?

In a tabletop or single player computer game, it’s about consistency. If leveling is a common result of the physics of the world, logic can be applied to the problem: powerful political figures are very high level, cities are built to account for the abilities of powerful creatures that might attack them, and the only reason to field armies of lower level individuals is to try to level them up into higher level individuals before a single champion of the other side crushes them. If leveling is an unusual fact, making the PCs more and more exceptional as they grow, the world needs to reflect this as well: player characters very quickly exceed threats on a mortal scale (no 20th level city guards), more and more adventures take them to places of great danger that are somehow blocked off from casual interaction with the world, and there is an expectation that they will be able to enact great change on the political landscape should they so desire.

In an MMO, it’s much harder. Unless you’re doing very complicated things with instancing, a player character can’t have a significant effect on the world. And, as more and more players race to the level cap, you wouldn’t really want them to anyway. It’s almost inevitable that player characters can return to areas where previously challenging enemies now fall like wheat, and the question starts: why don’t some of the max level characters take some time off of showing off their outfits in the central zone and deal with the problems that are now easy to them, so the people slightly below them can deal with easy problems, and so on down the line? Sure, the answer is, “because it’s a game,” but it’s also supposed to be using as many tricks as possible to preserve immersion. Auto-scaling things to the right level is not much more of a solution, as it can lead to feeling like leveling is mostly meaningless. Some possible, modular concepts to mitigate the problem off the top of my head:

• Periodically spawn higher level NPCs in low level zones that are somehow set up as the bosses of the area but not required in the main quest line for the low level story. Encourage high level PCs to return to deal with them in a way that doesn’t also encourage them to wipe out quest targets for the lower level characters in the zone.
• Frequently spawn much lower level NPCs in high level zones as the toadies for the appropriately-leveled threats. Players will mow through them and potentially enjoy getting to feel awesome.
• Avoid putting in drastic attack scaling between high and low level combatants: armies of low level characters vs. high level threats makes a little more sense if they can actually hit the target, even if for only a tiny fraction of a percent of its hit points. If PCs take on low level swaths of foes, they might get chewed up a little (and might be able to similarly chew up higher level targets).
• Set up the major threat of the world with some kind of story reason for clustering power in a central location and waning as it spreads away. Maybe it’s societal, or maybe there’s an actual power source of some kind that improves effectiveness based on proximity. Essentially, there should be some story based reason why the weakest, rather than the strongest, of an enemy group are able to penetrate furthest towards the good guy seats of power.
• Spawn non-boss NPCs at low level. Give them exp and let them level when they kill players on roughly the same scale as players gain exp (and maybe make them not target PCs for whom they’d get no exp unless attacked, just like PCs will rarely attack gray NPCs). Figure out how to get levels and location to persist through server restarts. When they die, have them rez like players at the nearest stronghold of their enemy group (or in the next nearest of higher level if they die in their stronghold) and then wander out to staff it. See what happens. (This could obviously go horribly awry.)
• Use instancing/phasing and gating to prevent players from returning to low level areas and preserve the fiction that time is passing. Everything in the world effectively levels up slightly more slowly than the PC so the gains in power feel more gradual. If players somehow return to low level zones, preserve the illusion that this is some kind of flashback, possibly even temporarily reducing their level to match (i.e., let them help their low level friends but not wander into low level zones as modern gods).

Some of those have been tried to some extent in previous or upcoming MMOs, but most of the time, AAA games tend to punt on this particular problem because it’s a game and games have levels and players, as a whole, are willing to ignore the inconsistencies. But I hold that, at some level, suspension of disbelief is being strained and it makes the world feel like a theme park, even if most players don’t consciously worry about it. With a few tricks to incorporate this weird inheritance from wargames, though, I believe immersion in games could be greatly increased, and everyone enjoys a little more immersion in their escapism.