It is never just a game

I keep track of several roleplay PWs on various platforms. One of them is Arelith, an NWN1 world. I highly respect the team that built and maintains the world. Their world is a mod of a seven year old game that last had an expansion nearly five years ago, yet the community is still healthy and active.

One thread on its forums however caught my eye, because it is about a design decision that makes me cringe. Like most NWN PWs, Arelith is a diku style world with a serial martyrdom death system. It has the slight variant to this that a dead character can either respawn, or wait until someone brings their corpse somewhere to be raised. If they are revived, they escape the death penalty that would otherwise have been applied. So far, this is typical NWN and there are many variants on this theme; some of which leave the player’s camera at the scene of death and others that move the camera to a death plane, leaving the corpse behind that can be revived. In NWN, the camera is tied to the PC avatar, so the latter is usually accomplished by moving the avatar to another location and leaving a corpse object behind. In Arelith’s case, the corpse object can be picked up and it can also be destroyed. The owner of the dead character currently gets a message indicating whenever their corpse has been picked up, dropped or destroyed, but not who did it. If the corpse has been destroyed, the destroyer gets a skull, a small amount of money and the “victim” loses his/her get out of jail free card and must take the respawn penalty.

This is where the fun begins. There is a mechanism where one player can make gain by inflicting other players with a perceived loss.

I’m not privy to the motivations for that design, though I suspect a pen and paper (PnP) D&D influence. Most likely, the convincing arguments were on a gamist and/or simulationist basis. In the former case, the idea that someone would want to avoid the “legitimate” penalty for failure is anathema. In the latter case, there is a realism factor in carrying corpses. I’m not actually interested in the specifics of the system or in the reasoning behind it so much as in the effect on the players. Some players have little or no advancement motive. If the player does not derive pleasure from counting points, grinding is a job; an entry level, low pay, low status job and loss of avatar capital is akin to not being paid for putting in your hours at McDonalds. Such players will go to great lengths to avoid a death penalty. If there is a way for them to avoid it by waiting for a rez, then they will do so.

If another player comes along and takes away that escape clause, it will cause distress. There are lots of arguments that the whole thing is entirely appropriate because it is in character (IC) and that it is just a game, so nobody should get worked up. Such arguments forget a little aspect of human psychology. When people feel that another player – as a player – has screwed them – as a player – over, then what is IC and what is metagaming simply does not matter. It does not matter if that player feels that it is entirely legitimate that the victim accepts the penalty. The victim feels that they are losing hours of grinding at the hands of another player; especially as it is not on a consensual basis. It is as if someone jumped on their sand castle and losing your sand castle this way is never just a game.

If you wanted to create a gameplay mechanic tailor made for griefing, you could do little better than Arelith’s corpse bashing, except perhaps by adding permadeath to the mix. Well, you could remove the message that a character’s corpse has been destroyed, causing them to wait in vain.

Anyone who has ever read through Nick Yee’s excellent work researching MMO player motivations may have caught the fact that the roleplay motivations are quite independent from other motivations. They might also have noticed that the roleplayer motivations are held more often by females and older players and the competition tends to be a motive for younger, male, players. In short, though you can’t make the blanket statement that roleplayers are carebears, the transect between the darkfall playerbase and the roleplayer motivation is only a fraction of the roleplay crowd. If your world is Darkfall meets roleplay and you are building for those players whole like to be IC while they gank people, then you can just tell the victim that he is being a whiney carebear and be done with it. Otherwise you have to do your best to “goon proof” your world.

There is a general rule that 1% of the population is sociopathic. This does not mean that they are raving serial murderers, but it means that at least one percent of the population – a subset composed almost entirely of males, so 2% of all men – lack empathy. Some are simply indifferent to the needs of others. Others actively enjoy inflicting anguish. The anonymity of online games is an attractant to these people in the same way that sweets attract bears at a campsite; so we have to reckon that our sociopath fraction is higher than 1%. Even if there was no gain to the character that destroyed the corpse, such people will derive pleasure from it for its own sake. That world currently has a problem with a griefer who repeated returns despite being repeatedly being banned. If he were more sophisticated, he’d not be attacking live characters, but instead anonymously bashing corpses. Such an individual could operate indefinitely, cause a great deal of psychological harm and never be caught. How many such individuals are currently operating this way on Arelith, using the IC cover that their character is evil? In addition to the bona fide sociopaths, a sizable portion of the playbase will not be averse to putting the screws to someone outside their monkeysphere if there was something to be gained from it. They’ll give plausible IC justifications of course, but the fact is that they – as players – are defecting when they harm other players in a non consensual way. This may sound strange coming from a advocate of strong IC systems, but “it is IC” can also be used as a pretext for a player to be a wanker. A “be nice” rule is meaningless in such an environment.

With this in mind, consider that your highly networked players are statistically more likely to be older females. 40 year old librarians who live with three cats and can name every character from the Wheel of Time series are a bit more likely to end up victimized by such a scheme than to be perpetrators. It is a hard fact that some of your players – the highly networked ones – are simply more valuable to the health of your community than others. These people won’t say a word, they’ll just leave for greener pastures and their friends will eventually follow them. This has happened before. I’ve spoken to one roleplay NWN2 PW admin who had his playerbase decimated by a something awful faction moving in. (Edit) Some of them (SA people/goons) seem to be excellent roleplayers, but they’ll still leave your world as a bloody, empty corpse if they don’t like it; and still might anyway even if they do. (/Edit) Now imagine if goons invaded your world? Do you have any systems that can abused to empty it?

The question I’d have for the server admin is whether he/she has an overview of the usage patterns of this bashing feature. Any time you have a gameplay feature that can be abused, you should assume that it will be and closely monitor it. The best way would be logging corpse related actions (perp playername & character name, victim playername and character name, the races of the two, was it bashed? Dropped into a container? Rezzed? Etc.) to a database and periodically pull it into Excel to data mine it in a pivot table. Such surveys would tell the server management who is doing what and under what conditions and would let them easily root out abuses of the system.

As for player behavior, he/she should also consider game theory and that if there is a gain and no penalty for defecting, eventually defecting will become the norm. He should resist calls to exacerbate the problem by removing the message that the corpse has been destroyed and instead give a clear indication of who picked the body up and what they did with it (e.g. if it was put down and where or put into another container). Such accountability may be metagaming, but by exposing players to a possible tit-for-tat retaliation if they defect, it enforces the be nice rule.

Your world is ok, but you really need to turn it into a different game entirely…

If you have ever run a world, you’ve probably seen one of these threads.  Heck, if you have only ever played or even just visited a forum or two, you have seen it.  Someone starts a thread pointing out the design “failures” in the world, except that the failures they are pointing out are not failures at all; they are just not suited to the player’s taste.  I’m not writing about the situation where the player appreciates what the designer was trying to do and perhaps actually made a mistake. I’m writing about the situation where a player is asking for classes and tank/rogue/nuker/healer raid instances in a classless world about puppies. What the player is really requesting is that you change your core design and turn it into something else.  These days, that something else is likely to be WoW. 

Heck, you’ve probably committed this sin at some point in the past.  I know I have and I give my apologies to those I’ve annoyed in the past by doing this. 

What made me think of this?  A little Onion video.  A link to that is a great snarky reply to any of those “turn your game into something else” threads.

The HSM (Heroes, not Serial Martyrs) Manifesto

(I’ve been thinking of refinements to the Heroes, not Serial Martyrs (HSM) concept and how to best present the concept in a single article. I promise afterwards to stop beating this dead horse for now and write about other topics)

In the purely action environment of most modern, 3D diku worlds, where players are always on Vent, never in character and focus on game mechanics because “it’s just a game”, then there are no real problems with the currently standard die/respawn cycle (serial martyrdom) of handling character death. For hardcore roleplay worlds, being one part game, one part social world and one part soundstage for collaborative community theatre, death is a core design problem. If standard diku style fuge-plane/respawn systems are used for player characters (and possibly select NPCs) only, then we either we follow the logical rabbit down the hole and face an Orwellian society or conveniently ignore the ramifications as is usually the case. If we extend the serial martyrdom across NPCs as well to avoid the unpleasant Orwellian side, then we face player revolution over the loss of meaning in the actions of their characters (no point in killing the orc king if he just respawns, now is there?) as well and considerations as to the society that may result from everyone having access to the respawn button. It may be an interesting society; it might make a great story or novel, but would have jarring immersion issues preventing players from establishing long term emotional attachment. If we accept that permanent death is an important element – both the fear of it and the risk of it in heroism – of narrative in roleplay environments, then the currently standard die/respawn cycle (serial martyrdom) of diku style worlds is certainly unhelpful. Unfortunately, the obvious solution, permadeath, usually – with the exception of a couple of RPI text MUDS – results in empty worlds.

The problems with Permadeath

1. Players are risk averse; especially in cases where they accumulate considerable avatar capital. Even if there is no on-character-sheet avatar capital accumulation, there is considerable soft avatar capital in a roleplay environment from the storylines that a character is involved in. Weighing risks greater than rewards is part of human nature and designers should accept that. This is the principal reason for the near universal rejection of permadeath by players.
2. If players see something on the character sheet, they feel that they can use it. 100 hit points does not mean 50 to burn and 50 as a reserve; it means 100 to burn and they will happily push it to the very limit. Pulling punches and not being able to use the full range of what is on the character sheet will be unsatisfying to them.
3. Players don’t like being reminded of their aversion to risk. This makes point 2 doubly unsatisfying.

Roleplay persistent worlds are not the only genre with a serious problem regarding how to handle character death. The horror genre relies on its ability to build tension due to fear and the usual death in games problems plays havoc there as well. Unfortunately, the possible solutions laid out by Nels Anderson in part II of his excellent series handling death in the horror genre, especially the “keep it short” bits, are not really applicable to the roleplay PW scenario where the player may spend hundreds of hours with his/her character.

Most persistent current world non PD death systems – the serial martyrdom systems – work by openly allowing the player to use the full range of what is on the character sheet and sparing them any serious risk by removing them from the situation and applying a mild (or non-existent) penalty; which is why they are generally considered acceptable to players, despite their undesirable side effects. Fundamentally, what “death” (or what Richard Bartle once unceremoniously labeled “wuss slap”) in non-PD systems is doing is removing the character from the situation. There may or may not be loss of avatar capital involved, but fundamentally, the character is being moved to a safe place to “start over”.

Most attempts at creating “soft” permadeath systems try to incorporate this “soft landing” to some degree. One approach is to replace the dead character with a relative, thus allowing the player to hold on to the hard avatar capital of gear and/or experience. I’ve seen variations of this idea put forth many times in mailing lists, forums and blogs, most recently by Flatfingers and Syncane. This system is really “respawn with a new avatar image and name” and in the case of roleplay environments still results in the loss of soft avatar capital, so it is not really an acceptable solution. Other approaches are to only kick in after X number of deaths or are used to unlock otherwise unavailable powers. None of these, with the exception perhaps of the last one, are really acceptable in a hardcore roleplay environment, which is why RPI text MUDS tend to have “hard” permadeath with no soft landing.

What is clear is that for – from the perspective of most players – a permadeath system to work, it has to assuage the players’ aversion to risk by providing a landing ramp that can functionally replace the trip to the graveyard/death-plane/respawn-button. The aim of the Heroes, not Serial Martyrs (HSM) approach to permadeath is to overcome the player objections to permadeath by the following rules:

• Offer the option of a safe landing via “safe passage” in the game mechanics for near death characters.
• Give a clear feedback in the game UI to the player that the character is in a bad situation and needs to be immediately extracted.
• Shift the burden of removing the character from the adverse situation from the server to the player.

The actually mechanics of the safe passage in a particular HSM implementation would vary from game design to game design, but the key remains. Whether to press the situation or play it safe is always a player decision and permadeath does not happen without player consent. If the player heeds the warning, and removes his/her character from the situation, HSM is functionally little different from the standard fuge-plane systems in terms of mechanics (though it is a world apart in narrative and lore terms). If the player ignores the warning feedback, it should be in the full knowledge that permanent character death will likely be the result and is in either explicit or implicit (through inaction) agreement with this possible outcome.

One thing you always have to keep in mind with HSM is part two in my problems with permadeath list above. Players will always see everything on the character sheet as resources to be used. Moreover, if they come to understand the underlying principals used in the safe passage mechanism, they internalize that mechanic into their assets, abuse it and be angry when their character dies. The key lies in managing player expectations. It should seem less like “playing it safe” than “whoah, this is hairy, I better scoot”! To prevent players from internalizing the safe passage landing ramp, it should either be completely hidden, or impossible to predict with much certainty.

Lastly, there are the abusers. Every roleplay PW designer/admin that I know can point to at least one “powergamer” who will use anything to maximize his (always a ‘his’) character and will happily abuse the safe passage if it is possible. A good HSM system needs to take these people into account and make attempts at abuse too risky to be a winning strategy, but not punish the average user in the crossfire.

HSM is an early stage idea at this point – a hypothetical general approach to PD. It has not yet been tried and I can think of a zillion ways to do it badly. One critical point however is that the exact safe passage conditions should never be known to the player. It should be enough for them to extract themselves from the situation, but it should never be something that they can rely on and ideally, the actual mechanism is hidden. There are a lot of tricky issues – specifically how to handle the grace buffer based on the mechanics of the world – and the devil is in the details; but I don’t think it is an insurmountable problem. It is instead an interesting design issue.

Anyone who wanted to test HSM in an NWN (or similarly diku) environment could try something like the following:

1 – On the on death event (triggered when a PC’s hit points reach 0), set the player to invulnerable for a random period of time. This time should be the time it takes to traverse the zone (to ensure escape) + something random. If the zone is particularly laggy or if lag tends to occur during peak hours, you can add an appropriate lag factor; for example based on the current number of players on the server and creatures in the zone. We’ll call this time the safe passage countdown time.

2 – Start a thread/process/script that will wait out the safe passage countdown time before continuing in 4.

3- If the PC leaves the zone (and is not going deeper into a dungeon crawl) during safe passage countdown time, then port him/her to a designated safe zone, remove the invulnerability and set him/her at 1 hit point. This represents the PC immediately returning to town. In this case, there is little functional difference between a traditional fuge-plane death system (at least a mild penalty one) and HSM.

4 – If at the end of the wait period, the PC is still invulnerable and still has less than one hit point, kill him/her. This is permadeath.

To prevent abuse, we also need to following elements:

5 – Once safe invulnerability is turned off, it can’t be turned on again for that character for some time (say five minutes). When the invulnerability is removed, record the time and in the onDeath event, ensure that at least five minutes has passed before allowing it to be turned back on. We’ll call this the blackout period. If the OnDeath event is triggered during the blackout period… permadeath.

6 – Extend the blackout period to a longer period (say 15 minutes) in the zone that it was last triggered. This is a good way to prevent the extreme serial martyrdom scenario where a player runs back to town, rests and runs back to finish off an already weakened enemy. (and its people who pull this kind of wankerism who will tell you with a straight face that they disdain “campfire RP” and that their character is a “hero”)

7 – If the player character is healed, remove the invulnerability and start the appropriate blackout timers. This prevents the scenario where safe passage is abused by allowing tanks to go invulnerable and then healing them before the end of the timer and the abuse scenario where the player uses the safe passage invulnerability to finish off an enemy (and probably uses only DPS, leaving healing until after the fight), then heals and continues normally. In both cases, safe passage would likely be unavailable for a long period of time and could potentially put the character in mortal danger. If you stick it out instead of fleeing, make it count.

Death Systems Matter

I started to respond to Opinvu’s offhand comment to my post on HSM that “the focus shouldn’t be on what the death rules are, but what the environment offers in return”, but realized that it deserves its own post. It is one of the very first choices you should make, along with the genre and the levels/skills/no-progression decision and is probably more important that either of the other two. The death paradigm in effect is probably the single most important decision you can make when designing a world. EVERYTHING revolves around it. By death paradigm, I’m not talking about what the death plane looks like, the lore behind it or whether a player needs to answer a riddle or run a maze to get out. I’m talking about whether the choice of gameplay mechanics mesh. If you are planning to use a hardcore world with permadeath around every corner, you will have to take this into account and not require a six month grind to break your character in. If you are planning on a very grindy world, or one where the characters are in play for very long periods of time, you either want to avoid PD completely or make it rare.

RPI worlds deemphasize the need to the collection of avatar capital (Edward Castranova’s term for the “stuff” – levels, gear, etc. – that a character accumulates) and feature permadeath.

A commenter on one of Syncane’s musings about permadeath put very simply the usual criticism of PD:

On the subject of perma-death (or any other harsh penalty for dying) — the problem with that is that it encourages (or even forces) boring “farming” of the mobs that can’t possibly kill your character rather than “having fun” by engaging challenging (risky) content.

The typical NWN environment , like all other Diku style environments (which means almost all of the active non-text worlds) requires a light to non-existant death penalty. Why? Because they are designed around the idea that there will be a mind numbing quantity of repetitive combat. The very first design decision usually made on NWN worlds is where to put the slider on XP, relative to third edition PnP D&D. The slider is normally set to between 5% and 10% to reflect that players in online worlds involve themselves in a LOT more combat than is typical in a pencil and paper scenario. To add to this, designers often deliberately lower the “challenge rating” (CR) in order to further bring down the XP gain to difficulty ratio for specific mob templates. I call this unfortunate and unbalancing practice “CR jacking”. In such an environment, you will die eventually and if you are playing a weak build or lack skills, you will die quickly and often. If a world combines low XP payout, CR jacking and a high death penalty, it will soon be devoid of players. I’ve seen players leave NWN servers over the death penalty to XP gain ratio. If you add PD to the typical NWN/Diku environment, you’ll have a disaster on your hands.

You have to design your world around your death paradigm.

Heroes, not Serial Martyrs

Some time ago, I wrote a post on permadeath. That a lack of permadeath removes death (as well as the fear of death and death’s consequences) as a narrative device is not in dispute. The real problem is how to handle the conflicting needs of roleplayers to have the full spectrum of narrative devices available, with the general desire of players not to lose their characters to a lag spike.

A possible idea comes from Hindu mythology. There was once a demon named Bali. This demon was an ambitious fellow; so he performed many austerities and acts of charity such that he became invincible and threw the gods out of heaven. They could not do anything about him. Eventually, through an act of trickery, lord Vishnu made him give up heaven and retreat to the underworld. What is interesting about this story is the simple fact that Bali had built up so much good karma that he could not be harmed. We could look at this story as a simple case where he achieved very high level, or had monsterous buffs, etc. that he simple owned anyone else who opposed him.

There is an alternative explanation however: the “die rolls were being fudged” to reflect his karma.

If you were to survey literature, one constant is clear; unless the hero is supposed to die as a literary device, he does not die. Normal people die, but heroes (and great villians) are blessed by fate. There are no reoccurring trips to the afterlife. Instead, it is narrow brushes with death that rule the day. The bullet narrowly misses the hero’s head. The bomb is shut off in the last second on the clock. The trolls about to eat the hero argue about how to cook him until daylight turns them to stone.

Which brings us to what I call the Heroes, not Serial Martyrs (HSM) approach to death. Simply don’t kill the player characters off. Give them some sort of buffer of fudged die rolls. The bullets that should be riddling their corpse are instead whizzing past their head. The deathblow from the black knight’s sword instead narrowly misses. While doing this, make it clear to them that they should be dead now and if they push their luck too far, it could end badly. If they heed the warning and remove themselves from the situation, then all will be well. In an HSM environment, a character can only die if the player means to allow it; such as heroically holding the rear guard and ignoring the warning messages. This is in direct contrast to the serial martyrs approach of it being easy to die, but just as easy to get back into the action.

The fudge buffer need not be infinite and there are many variations that can be used depending on how “dangerous” the designer wants the worlds to be. It could be set at character creation and not possible to increase. This is similar to the after X number of deaths, the character is perma’d approach to PD (one of the “soft” permadeath approaches). It could be on a per incident basis, allowing an infinite number of hair raising “near misses”. It could be based on pious acts of the character, such as in the case of Bali, or quests, etc.

The HSM approach may allow us to keep a relaxed approach to play in dangerous situations (and avoid the resultant player loss of hardcore PD), while retaining death as a narrative device. .