Post by Morreion on Oct 4, 2022 12:09:36 GMT -5
To start off, here's a recap of Ultima Online's history and legacy-
The Game Archaeologist: How Ultima Online got made (Massively)
The Game Archaeologist: The legacy of Ultima Online (Massively)
And now to Raph Koster's gaming development memoir, Postmortems.
‘Magic through serendipity’: Raph Koster on the glorious mess that was Ultima Online (Massively)
The Failed Attempt at a Living Ecology and Economy
The Huge Problem of Murder
A Brief History of Murder in Ultima Online (GameDeveloper.com)
The Game Archaeologist: How Ultima Online got made (Massively)
So how did it all begin for this title? Starr Long claimed that the inspiration for an MMO version of Ultima Online came while playing DOOM and realizing how much fun competitive online play could be.
In any case, considering the rapidly connecting online world of the mid-1990s, Garriott leaped at the thought of taking Ultima to the next level. His studio, Origin Systems, was eager to start development on what would be called Ultima Online: Shattered Legacy (and, before that, Multima — as in, “Multiplayer Ultima”). To pull off this feat, he assembled a team of hotshot young developers, including Long, Raph Koster, Kristen Koster, Scott Phillips, Rick Delashmit, and Rich Vogel.
The project hit several snags from the very beginning, the biggest of which was that Origin’s parent company EA didn’t see any profit potential in MMOs. The company was also worried that a risky flop would tarnish the Ultima brand. After several rejections, the small dev team finally convinced the execs and got the go-ahead to make their massively multiplayer game.
But they weren’t treated like rock stars: The tiny team had to make do with office space on a floor of a building that was still under construction and had holes that could plunge a careless person to their death. (Fortunately, no one was actually killed.)
“At that time, we were punk kids doing stuff in the attic, and our parents had no idea what we were up to,” recalled Raph Koster in his book Postmortems, which we’ve quoted extensively in the past. He noted that all of the core programmers and designers came from text MUD backgrounds.
In any case, considering the rapidly connecting online world of the mid-1990s, Garriott leaped at the thought of taking Ultima to the next level. His studio, Origin Systems, was eager to start development on what would be called Ultima Online: Shattered Legacy (and, before that, Multima — as in, “Multiplayer Ultima”). To pull off this feat, he assembled a team of hotshot young developers, including Long, Raph Koster, Kristen Koster, Scott Phillips, Rick Delashmit, and Rich Vogel.
The project hit several snags from the very beginning, the biggest of which was that Origin’s parent company EA didn’t see any profit potential in MMOs. The company was also worried that a risky flop would tarnish the Ultima brand. After several rejections, the small dev team finally convinced the execs and got the go-ahead to make their massively multiplayer game.
But they weren’t treated like rock stars: The tiny team had to make do with office space on a floor of a building that was still under construction and had holes that could plunge a careless person to their death. (Fortunately, no one was actually killed.)
“At that time, we were punk kids doing stuff in the attic, and our parents had no idea what we were up to,” recalled Raph Koster in his book Postmortems, which we’ve quoted extensively in the past. He noted that all of the core programmers and designers came from text MUD backgrounds.
The Game Archaeologist: The legacy of Ultima Online (Massively)
Critics were not lining up to gush about Ultima Online; as Koster noted in his book Postmortems, “Almost all of the reviews were middling to bad. Six out of 10 was a pretty common score.“ He did note that the team received several letters of praise from fellow industry members.
Even so, CNN would later call this launch “a watershed event in the industry… the first that could be termed a breakaway success.” Origin figured that it would see players in the tens of thousands and became overwhelmed when 87K copies were sold by year’s end and over 100,000 players subbed before Ultima Online was six months old.
These were substantial numbers at the time, marking UO as the first MMO to ever cross a six-digit population. And that’s not even counting the alleged half-million or so Chinese players who were flooding onto Ultima Online’s reverse-engineered emulators and rogue servers by 1999.
Ultima Online’s popularity and profitability turned quite a few heads in the video games industry, prompting the greenlight for many other studio MMOs to go into development. By 2003, Ultima Online’s numbers had climbed to a quarter-million paying subscribers who were filling the studio’s coffers every month, a high water mark for the title’s lifespan.
Even so, CNN would later call this launch “a watershed event in the industry… the first that could be termed a breakaway success.” Origin figured that it would see players in the tens of thousands and became overwhelmed when 87K copies were sold by year’s end and over 100,000 players subbed before Ultima Online was six months old.
These were substantial numbers at the time, marking UO as the first MMO to ever cross a six-digit population. And that’s not even counting the alleged half-million or so Chinese players who were flooding onto Ultima Online’s reverse-engineered emulators and rogue servers by 1999.
Ultima Online’s popularity and profitability turned quite a few heads in the video games industry, prompting the greenlight for many other studio MMOs to go into development. By 2003, Ultima Online’s numbers had climbed to a quarter-million paying subscribers who were filling the studio’s coffers every month, a high water mark for the title’s lifespan.
And now to Raph Koster's gaming development memoir, Postmortems.
‘Magic through serendipity’: Raph Koster on the glorious mess that was Ultima Online (Massively)
“To me, the takeaway (which has only been reinforced for me over the years) is that skunkworks really works. UO was a skunkworks project, through and through. I am fairly sure that if we had not been stuck in the ‘closet’ on the top floor, it would not have been made. I should emphasize that this also means we were lucky it worked at all, and it was held together with chewing gum. I don’t mean to paint the team as heroes. We were young, arrogant, and blinkered to consequences of our choices.”
The Failed Attempt at a Living Ecology and Economy
If you’ve followed Koster for any length of time on Twitter, you surely know that he alternates between extolling the virtues of simulation AI and warning against its inevitable misuses by human interlopers. That superficial contradiction is no contradiction at all: He’s been trying to simulate a virtual world for decades and understands possibly better than anyone what can go wrong.
Ultima Online was essentially the Kosters’ grand ecology and economy experiment. Everything from the animals and monsters to the shopkeepers’ inventories was intended to be programmed according to a modification of Maslow’s hierarchy of needs: Each mobile had a hierarchical list of wants and activities based on those wants, like a dragon wanting to protect its hoard of gold when sated but risking that hoard by chasing down random adventurers to eat when hungry.
But the technology of the mid ’90s was insufficient to actually make most of that happen. Even had Origin been willing to fund the team’s wildest ambitions, the servers simply couldn’t handle the scripting underpinning – for example – NPC beggars who searched for and then followed around only rich players. The sheer expense of massive radial searching and pathfinding was just too much to process. And if you’re currently wondering whether players would even be able to tell the difference between a dragon whose AI had cycled it through 500 need queries and a dragon that simply spawns and tries to kill randomly, then you’re asking the right question – Koster asks it too.
“There were a fair amount of team members who saw the whole system as a boondoggle, and not worth pursuing,” Koster writes, and so most of the plans were discarded before the game ever got to beta.
“At some point, reality will catch up to our designs from 1995.” -Raph KosterThat wasn’t the only problem. As originally designed, Koster admits, the closed economic system of UO wouldn’t have worked perfectly either; he calls it a “mistake,” with mechanics that ended up gutted because they just weren’t fun. And that’s chiefly because of the players. Players are incurable hoarders, which means a large amount of resources in the game will just sit around in someone’s bank (or a shopkeep’s bloated inventory) rather than keep moving through the world, at least if there are insufficient drains or uses for the item. Another fascinating bit: He reminds the reader that while in the real world, we find it acceptable that making wooden foozles may not make money, but in a virtual world, players will literally file bug reports when there’s no market for their craftables. After all, why would a dev put it in the game to be made if it was worthless?
That doesn’t mean Koster has given up on the dream; he’s just waiting for the tech to catch up. “I’d much rather be burning CPU on this sort of thing, frankly, than on 3-D collision,” he says, essentially arguing that scripted AI ecology and economy matter more than whether you can jump or whether every tile has been hand-placed. “The kinds of immersive power that a simulation like this can bring opens a lot more doors. In the long run, I believe that all the pressures are towards simulationist environments, rather than handcrafted ones. CPU power continues to outpace the cost of human capability to design static scenarios. At some point, reality will catch up to our designs from 1995.”
Ultima Online was essentially the Kosters’ grand ecology and economy experiment. Everything from the animals and monsters to the shopkeepers’ inventories was intended to be programmed according to a modification of Maslow’s hierarchy of needs: Each mobile had a hierarchical list of wants and activities based on those wants, like a dragon wanting to protect its hoard of gold when sated but risking that hoard by chasing down random adventurers to eat when hungry.
But the technology of the mid ’90s was insufficient to actually make most of that happen. Even had Origin been willing to fund the team’s wildest ambitions, the servers simply couldn’t handle the scripting underpinning – for example – NPC beggars who searched for and then followed around only rich players. The sheer expense of massive radial searching and pathfinding was just too much to process. And if you’re currently wondering whether players would even be able to tell the difference between a dragon whose AI had cycled it through 500 need queries and a dragon that simply spawns and tries to kill randomly, then you’re asking the right question – Koster asks it too.
“There were a fair amount of team members who saw the whole system as a boondoggle, and not worth pursuing,” Koster writes, and so most of the plans were discarded before the game ever got to beta.
“At some point, reality will catch up to our designs from 1995.” -Raph KosterThat wasn’t the only problem. As originally designed, Koster admits, the closed economic system of UO wouldn’t have worked perfectly either; he calls it a “mistake,” with mechanics that ended up gutted because they just weren’t fun. And that’s chiefly because of the players. Players are incurable hoarders, which means a large amount of resources in the game will just sit around in someone’s bank (or a shopkeep’s bloated inventory) rather than keep moving through the world, at least if there are insufficient drains or uses for the item. Another fascinating bit: He reminds the reader that while in the real world, we find it acceptable that making wooden foozles may not make money, but in a virtual world, players will literally file bug reports when there’s no market for their craftables. After all, why would a dev put it in the game to be made if it was worthless?
That doesn’t mean Koster has given up on the dream; he’s just waiting for the tech to catch up. “I’d much rather be burning CPU on this sort of thing, frankly, than on 3-D collision,” he says, essentially arguing that scripted AI ecology and economy matter more than whether you can jump or whether every tile has been hand-placed. “The kinds of immersive power that a simulation like this can bring opens a lot more doors. In the long run, I believe that all the pressures are towards simulationist environments, rather than handcrafted ones. CPU power continues to outpace the cost of human capability to design static scenarios. At some point, reality will catch up to our designs from 1995.”
The Huge Problem of Murder
Easily the most amusing – and horrifyingly frustrating – chapters of Koster’s Ultima Online reminiscing revolve around the game’s infamous early PvP mechanics. It’s clear the Origin team struggled with a naive sort of libertarian mindset when developing – or more accurately, not developing – the PvP at launch. As Koster’s now-ancient blog posts relate, the developers truly believed that their laissez faire “let the community police itself” philosophy was the best approach. You don’t even need to read Koster’s account; you can hear it repeated in the plaintive appeals of FFA PvP players even here in 2018, as the wolves desperately try to convince the sheep to come play victim in the service of the simulation.
“We must have playerkillers in UO because the world would suffer if we did not have them,” he once wrote. “But they also must be channeled so that their effect is beneficial and not detrimental. […] We don’t want to exterminate them completely anymore than we want to make rattlesnakes, black widow spiders, and sharks extinct because they fill a valuable role in the virtual ecology.”
Of course, we all know the story of how the rattlesnakes, black widow spiders, and sharks played out in practice: Rampant ganking and griefing and PKing drove players out of the 1997 sandbox by the thousands – “a truly distressing number of our new player acquisitions,” Koster laments – until the developers engineered countermeasures, patch by patch. They tried flagging. They tried a notoriety system. They tried reputation. They tried bounties. They tried a faction system. They tried guild wars. They tried newbie protection mechanics. But for every attempt to curb the PKs the devs put in, the reds invented some fresh hell for their victims, exploiting every ruleset change. It’s hard not to chuckle as Koster rattles off each new idea and how the players thwarted it (especially when I remember it happening!).
And that’s a theme Koster riffs on over and over – that “no matter what you do, players will decode every formula, statistic, and algorithm in your world via experimentation.” In fact, he argues, you don’t even need combat in your game for griefplay to be present (as anyone who’s read any comments anywhere can attest).
Eventually, the team, by then absent Koster, implemented Trammel, effectively creating mirror worlds safe from player-killers. “I wouldn’t have done it, personally, but there is no question that the userbase doubled once this went in,” he says.
“The result [of UO’s PK environment] was an exodus driven not only by the more modern 3-D graphics of [EverQuest] but by the safety. Everything I had thought about the impossible admin load of having a PK switch with a large-scale game was disproven in short order, and players wasted no time in telling me bluntly that I had been drastically and painfully wrong. In the name of player freedoms, I had put them through the slow-drip torture of two years of experiments with slowly tightening behavior rules, trying to save the emergence while tamping down the bad behavior. The cost was the loss of many hundreds of thousands of players. Ultima Online had churned through more than twice as many players who quit than EverQuest even got as subscribers that year.”
Even so, he frets over the loss to the verisimilitude of the virtual world. The griefer environment may not have been realistic, but it did create “endless stories and excitement, the stories that people tell and retell to this day,” as players were forced to work together to overcome the true evil in the game: the PK players themselves. That, he argues, was “empowering” in a way that “casual” post-Trammel player towns never were. Remember Kazola’s Tavern and the multiple PK guilds that ravaged Great Lakes, and the anti-PK guilds that rose up to fight them? I do; I was there, I defended Kazola’s, and I was once guilded with folks from one of the groups he mentions (SIN – I’ve written about that before). How many games have that sort of meaningful history? I can count them on one hand, and it’s no accident they’re all sandboxes. And what if participating in a genuine struggle against evil players – or figuring out ways to deal with the worst elements of society – “means [we] are more likely to dare to do it in real life instead of living in passivity”? In eschewing free-for-all games, are we just “giving up on the hard problem of freedom co-existing with civility”?
It’s intoxicating. But still maybe too idealistic. And Koster admits as much in essays written years after he’d moved on to new “experiments,” clearly having re-examined his ideas under the light of implementation and disaster.
“I like safe and wild zoning now. I really, really didn’t,” he pens. “I used to think that you could reform bad apples, and argue with hard cases. I’m more cynical these days. […] I used to think that people were willing to act communally for the good of the community. Now I know more about the Tragedy of the Commons and the Prisoner’s Dilemma and think that people are mostly selfish. This isn’t Ivory Tower theory gone looking for empirical evidence. It’s experience gone looking for explanation.”
When I spoke to Koster about this part of his book to get his most current take on MMO PvP, he explained that the massive shifts in the modern gaming landscape have changed his outlook even further.
“Now there are so many worlds and so many gamers that you can actually have a pure PvP game and it can survive. That wasn’t the case back then. I think we’ve even conditioned players to it over the years, and now we have stuff like Fortnite and PUBG. There are a lot of things in the survival genre which are reminiscent of UO, to me. Crowfall [the in-production PvP MMORPG on which Koster has consulted] is, of course, Todd [Coleman]’s game, not mine; I just advise on it. But if I were making my own game now, it’d still have wild PvP areas in it. Just, I’d find a way for Kazola to actually win, so to speak, by using the sorts of town mechanics we later developed in [Star Wars Galaxies].”
“We must have playerkillers in UO because the world would suffer if we did not have them,” he once wrote. “But they also must be channeled so that their effect is beneficial and not detrimental. […] We don’t want to exterminate them completely anymore than we want to make rattlesnakes, black widow spiders, and sharks extinct because they fill a valuable role in the virtual ecology.”
Of course, we all know the story of how the rattlesnakes, black widow spiders, and sharks played out in practice: Rampant ganking and griefing and PKing drove players out of the 1997 sandbox by the thousands – “a truly distressing number of our new player acquisitions,” Koster laments – until the developers engineered countermeasures, patch by patch. They tried flagging. They tried a notoriety system. They tried reputation. They tried bounties. They tried a faction system. They tried guild wars. They tried newbie protection mechanics. But for every attempt to curb the PKs the devs put in, the reds invented some fresh hell for their victims, exploiting every ruleset change. It’s hard not to chuckle as Koster rattles off each new idea and how the players thwarted it (especially when I remember it happening!).
And that’s a theme Koster riffs on over and over – that “no matter what you do, players will decode every formula, statistic, and algorithm in your world via experimentation.” In fact, he argues, you don’t even need combat in your game for griefplay to be present (as anyone who’s read any comments anywhere can attest).
Eventually, the team, by then absent Koster, implemented Trammel, effectively creating mirror worlds safe from player-killers. “I wouldn’t have done it, personally, but there is no question that the userbase doubled once this went in,” he says.
“The result [of UO’s PK environment] was an exodus driven not only by the more modern 3-D graphics of [EverQuest] but by the safety. Everything I had thought about the impossible admin load of having a PK switch with a large-scale game was disproven in short order, and players wasted no time in telling me bluntly that I had been drastically and painfully wrong. In the name of player freedoms, I had put them through the slow-drip torture of two years of experiments with slowly tightening behavior rules, trying to save the emergence while tamping down the bad behavior. The cost was the loss of many hundreds of thousands of players. Ultima Online had churned through more than twice as many players who quit than EverQuest even got as subscribers that year.”
Even so, he frets over the loss to the verisimilitude of the virtual world. The griefer environment may not have been realistic, but it did create “endless stories and excitement, the stories that people tell and retell to this day,” as players were forced to work together to overcome the true evil in the game: the PK players themselves. That, he argues, was “empowering” in a way that “casual” post-Trammel player towns never were. Remember Kazola’s Tavern and the multiple PK guilds that ravaged Great Lakes, and the anti-PK guilds that rose up to fight them? I do; I was there, I defended Kazola’s, and I was once guilded with folks from one of the groups he mentions (SIN – I’ve written about that before). How many games have that sort of meaningful history? I can count them on one hand, and it’s no accident they’re all sandboxes. And what if participating in a genuine struggle against evil players – or figuring out ways to deal with the worst elements of society – “means [we] are more likely to dare to do it in real life instead of living in passivity”? In eschewing free-for-all games, are we just “giving up on the hard problem of freedom co-existing with civility”?
It’s intoxicating. But still maybe too idealistic. And Koster admits as much in essays written years after he’d moved on to new “experiments,” clearly having re-examined his ideas under the light of implementation and disaster.
“I like safe and wild zoning now. I really, really didn’t,” he pens. “I used to think that you could reform bad apples, and argue with hard cases. I’m more cynical these days. […] I used to think that people were willing to act communally for the good of the community. Now I know more about the Tragedy of the Commons and the Prisoner’s Dilemma and think that people are mostly selfish. This isn’t Ivory Tower theory gone looking for empirical evidence. It’s experience gone looking for explanation.”
When I spoke to Koster about this part of his book to get his most current take on MMO PvP, he explained that the massive shifts in the modern gaming landscape have changed his outlook even further.
“Now there are so many worlds and so many gamers that you can actually have a pure PvP game and it can survive. That wasn’t the case back then. I think we’ve even conditioned players to it over the years, and now we have stuff like Fortnite and PUBG. There are a lot of things in the survival genre which are reminiscent of UO, to me. Crowfall [the in-production PvP MMORPG on which Koster has consulted] is, of course, Todd [Coleman]’s game, not mine; I just advise on it. But if I were making my own game now, it’d still have wild PvP areas in it. Just, I’d find a way for Kazola to actually win, so to speak, by using the sorts of town mechanics we later developed in [Star Wars Galaxies].”
A Brief History of Murder in Ultima Online (GameDeveloper.com)
Fighting the losing battle
Even whilst putting in features like this, designed to reduce the incidence of playerkilling, the team was busily adding new simulation features that increased it. I mean, just one week later we tried to curb thievery by allowing players to add traps to locked containers. A tinker could use metal and crossbow bolts or potions to make explosive, dart, or poison traps. The intent was to let players defend their possessions from theft in their homes or in their bags by letting them put them inside locked containers.
But what happened instead? Locked chests blew up inside backpacks, killing you, when thieves opened them. We left in the ability for thieves to disarm the traps, so they weren’t always effective. People made chain reactions of explosives, so that they could light a fuse outside town and cause a death inside town. People sold trapped locked containers to shopkeepers, who then resold the booby trap to unsuspecting victims. Leaving trapped chests at crossroads was a common ambush tactic. Yes, of course there were skills for detecting traps and disarming them… but you had to be a canny player to know of, and use, these tools. When we ran some metrics that year, the number one killer in the game was named TinkerBoy and had personally been responsible for more than 3,000 deaths.
The victims were disproportionately new players who didn’t know the ropes. And we were losing a truly distressing number of our new player acquisitions — Ultima Online was the fastest selling Electronic Arts game in history, well on its way to being a massive massive hit. But our subscriber numbers, while stratospheric for the day, weren’t keeping up because the losses were so high.
We had a world where a bard could entice an NPC shopkeeper out of town safety, kill them, and steal everything. Or provoke them to anger, get them to attack a random passerby, then call the guards on the shopkeeper for illegal behavior. Where you could die while polymorphed into a deer, resurrect still in that body that resembled a deer in every way, and therefore be able to wander into a player’s house without them suspecting a thing — and rob it blind. Where people would find already locked chests in a house, and leave a trap for the unsuspecting actual owner! Even the “good guys” took part, luring guards out of town and leaving them near Blackthorn’s shrine, where murderers resurrected.
That Christmas, we spawned Santa Clauses in every town, and put a gift in every player’s backpack. Players stole the clothes off of Santa, leaving naked men chanting “Ho ho ho!” right where new players logged in. Then they formed roving bands of Santa Clauses and roamed around slaughtering everyone with chilling war cries wishing people Happy Holidays.
The worst of all these exploits were around player housing. UO allowed players to build houses anywhere in the world that they fit. The patch notes for the six months are a litany of exceptions: no houses on tilled fields. No houses on roads. No houses inside dungeons.
Houses had tilted thatch or tile roofs, which were accomplished with an optical illusion, rather than being a solid floor (like all the flat roofs in the game). This meant that if you could get up to the roof, you could simply fall in, steal whatever you liked, and walk out the front door. Ultima Online didn’t simulate gravity, so players would place a chair next to the house, stack a second chair on top of it, stand on top of the upper chair, remove the lower chair, and repeat until they stood in midair floating atop suspended chairs, and simply walk onto the roof. They found ways of sneaking past doors, of teleporting in by exploiting minor collision bugs when dropping items, and worst of all, of obtaining player keys.
You see, there wasn’t much of a concept of “ownership” in early UO. It existed for actions, as we have seen in the case of notoriety, and it existed for pets, but it did not exist for objects. Locked items, including houses, were tied to keys. Keys could be duplicated, and critically, stolen. Lose your key, and you effectively lost the house and all its contents: potentially months and months worth of character investment.
Even whilst putting in features like this, designed to reduce the incidence of playerkilling, the team was busily adding new simulation features that increased it. I mean, just one week later we tried to curb thievery by allowing players to add traps to locked containers. A tinker could use metal and crossbow bolts or potions to make explosive, dart, or poison traps. The intent was to let players defend their possessions from theft in their homes or in their bags by letting them put them inside locked containers.
But what happened instead? Locked chests blew up inside backpacks, killing you, when thieves opened them. We left in the ability for thieves to disarm the traps, so they weren’t always effective. People made chain reactions of explosives, so that they could light a fuse outside town and cause a death inside town. People sold trapped locked containers to shopkeepers, who then resold the booby trap to unsuspecting victims. Leaving trapped chests at crossroads was a common ambush tactic. Yes, of course there were skills for detecting traps and disarming them… but you had to be a canny player to know of, and use, these tools. When we ran some metrics that year, the number one killer in the game was named TinkerBoy and had personally been responsible for more than 3,000 deaths.
The victims were disproportionately new players who didn’t know the ropes. And we were losing a truly distressing number of our new player acquisitions — Ultima Online was the fastest selling Electronic Arts game in history, well on its way to being a massive massive hit. But our subscriber numbers, while stratospheric for the day, weren’t keeping up because the losses were so high.
We had a world where a bard could entice an NPC shopkeeper out of town safety, kill them, and steal everything. Or provoke them to anger, get them to attack a random passerby, then call the guards on the shopkeeper for illegal behavior. Where you could die while polymorphed into a deer, resurrect still in that body that resembled a deer in every way, and therefore be able to wander into a player’s house without them suspecting a thing — and rob it blind. Where people would find already locked chests in a house, and leave a trap for the unsuspecting actual owner! Even the “good guys” took part, luring guards out of town and leaving them near Blackthorn’s shrine, where murderers resurrected.
That Christmas, we spawned Santa Clauses in every town, and put a gift in every player’s backpack. Players stole the clothes off of Santa, leaving naked men chanting “Ho ho ho!” right where new players logged in. Then they formed roving bands of Santa Clauses and roamed around slaughtering everyone with chilling war cries wishing people Happy Holidays.
The worst of all these exploits were around player housing. UO allowed players to build houses anywhere in the world that they fit. The patch notes for the six months are a litany of exceptions: no houses on tilled fields. No houses on roads. No houses inside dungeons.
Houses had tilted thatch or tile roofs, which were accomplished with an optical illusion, rather than being a solid floor (like all the flat roofs in the game). This meant that if you could get up to the roof, you could simply fall in, steal whatever you liked, and walk out the front door. Ultima Online didn’t simulate gravity, so players would place a chair next to the house, stack a second chair on top of it, stand on top of the upper chair, remove the lower chair, and repeat until they stood in midair floating atop suspended chairs, and simply walk onto the roof. They found ways of sneaking past doors, of teleporting in by exploiting minor collision bugs when dropping items, and worst of all, of obtaining player keys.
You see, there wasn’t much of a concept of “ownership” in early UO. It existed for actions, as we have seen in the case of notoriety, and it existed for pets, but it did not exist for objects. Locked items, including houses, were tied to keys. Keys could be duplicated, and critically, stolen. Lose your key, and you effectively lost the house and all its contents: potentially months and months worth of character investment.
The reputation system
The redesign was called “the reputation system,” which is of course the general term for all systems that track an overall rating for an individual based on the feedback of other users; eBay’s stars are a reputation system, your upvotes on Reddit are a reputation system, and so on. Technically notoriety doesn’t count as a reputation system, because it’s simply adjusting a value in code, without a user getting to decide how they feel about another user. It is more like what we might term an alignment system, drawn from Dungeons & Dragons. Murder counts, however, are a form of a negative reputation system (a system with only upvotes would be a positive reputation system).
Rep systems in general were a relatively new idea at the time, with UO’s murder report system as one of the early mainstream examples alongside Slashdot’s karma and eBay’s star ratings. Slashdot can probably be credited as bringing the concept to broader awareness. But there were many antecedents: at Xerox PARC in 1992, the Tapestry email system used annotations as a way to filter email; annotations were effectively upvotes. BBS systems often had “leech scores” to track people who downloaded without uploading. And the pioneering American Information Exchange system was developing early forms of smart contracts that basically tracked reputation, clear back between 1988 and 1991. Many folks were wrestling with the same issues that we were: Sybil attacks, whitewashing attacks, and distributed reputation...
All of the complex feedback and color coding was boiled down to Blue, Gray, and Red, with innocents Blue and Murderers red. Gray indicated a recent flagging for doing something that harmed or might harm others. These two things were different internally: one indicated aggressive behavior and the other criminality, but they looked the same.
The key difference was that aggressive behavior only turned you gray for people who got hurt. Criminal behavior generally turned you gray to everyone, except for the wrinkle that failed thievery could flag you as a Criminal just to the victim until you next died. This allowed the victim of thievery to catch the thief later and take revenge even though time had elapsed. Looting corpses that were not your own also made you a criminal.
Anyone you saw as gray could be attacked without repercussions, but it wore off after two minutes. This meant that if you hurt someone with an ill-considered earthquake spell, anyone hurt in the earthquake could attack you without penalty, but third parties who weren’t hurt would still see you as blue.
Killing a blue person (an Innocent, in the system’s terms) meant that they had the ability to report you, just as in the earlier murder system. In fact, just hurting an Innocent who later died in a separate incident before they were fully healed up meant they could report you. A murder count incremented on you; it faded away at the rate of one murder every eight hours of real-time gameplay. If you hit five murder count, you were turned red, and became a Murderer.
Murderers suffered severe penalties upon death (scaled by how many people you had killed), were killed on sight by guards (effectively blocking you from towns while red), and also generated bounties like the older system.
Alongside this system were a number of extra features for guilds — they were able to declare themselves Order or Chaos, which replaced the older system, and had their own green and orange color-coding.
The system also put an emphasis on people taking proactive action. Guards no longer just appeared when you attacked someone; instead, someone had to call for them. Thieves had to be “noticed” by someone, and then you could call guards to kill them. It meant that towns actually got slightly less safe.
This system did have a noticeable effect on the amount of playerkilling, but it remained still too high. The sight of Murderers locked in their houses running macros overnight to reduce murder count became a common sight. The stat loss could be significant, so Murderers simply spaced out their kills so they could stay above the threshold of five murder count...
EverQuest had a simple solution to all the above: they had a PK switch. Meaning, unless you flagged yourself as PK-enabled (which few did), you were safe. The end.
This worked acceptably in EverQuest because it was simply a far more constrained game than Ultima Online. You couldn’t drop things on the ground. No laying of traps. No stealing. No houses. No stacking chairs. No chairs. Grief players happily led monsters into newbie zones and killed them, but you couldn’t loot others, so there was no reward cycle there except the gnashing of teeth. There wasn’t crafting. It was a game driven by player-vs-environment combat, and to a veteran MUD player felt very strongly like a DikuMUD with first-person 3d.
The result was an exodus driven not only by the more modern 3d graphics of the newer game, but by the safety. Everything I had thought about the impossible admin load of having a PK switch with a large-scale game was disproven in short order, and players wasted no time in telling me bluntly that I had been drastically and painfully wrong.
The result? In the name of player freedoms, I had put them through a slow-drip torture of two years of experiments with slowly tightening behavior rules, trying to save the emergence while tamping down the bad behavior. The cost was the loss of many hundreds of thousands of players. Ultima Online had churned through more than twice as many players who quit than EverQuest even got as subscribers that year.
The redesign was called “the reputation system,” which is of course the general term for all systems that track an overall rating for an individual based on the feedback of other users; eBay’s stars are a reputation system, your upvotes on Reddit are a reputation system, and so on. Technically notoriety doesn’t count as a reputation system, because it’s simply adjusting a value in code, without a user getting to decide how they feel about another user. It is more like what we might term an alignment system, drawn from Dungeons & Dragons. Murder counts, however, are a form of a negative reputation system (a system with only upvotes would be a positive reputation system).
Rep systems in general were a relatively new idea at the time, with UO’s murder report system as one of the early mainstream examples alongside Slashdot’s karma and eBay’s star ratings. Slashdot can probably be credited as bringing the concept to broader awareness. But there were many antecedents: at Xerox PARC in 1992, the Tapestry email system used annotations as a way to filter email; annotations were effectively upvotes. BBS systems often had “leech scores” to track people who downloaded without uploading. And the pioneering American Information Exchange system was developing early forms of smart contracts that basically tracked reputation, clear back between 1988 and 1991. Many folks were wrestling with the same issues that we were: Sybil attacks, whitewashing attacks, and distributed reputation...
All of the complex feedback and color coding was boiled down to Blue, Gray, and Red, with innocents Blue and Murderers red. Gray indicated a recent flagging for doing something that harmed or might harm others. These two things were different internally: one indicated aggressive behavior and the other criminality, but they looked the same.
The key difference was that aggressive behavior only turned you gray for people who got hurt. Criminal behavior generally turned you gray to everyone, except for the wrinkle that failed thievery could flag you as a Criminal just to the victim until you next died. This allowed the victim of thievery to catch the thief later and take revenge even though time had elapsed. Looting corpses that were not your own also made you a criminal.
Anyone you saw as gray could be attacked without repercussions, but it wore off after two minutes. This meant that if you hurt someone with an ill-considered earthquake spell, anyone hurt in the earthquake could attack you without penalty, but third parties who weren’t hurt would still see you as blue.
Killing a blue person (an Innocent, in the system’s terms) meant that they had the ability to report you, just as in the earlier murder system. In fact, just hurting an Innocent who later died in a separate incident before they were fully healed up meant they could report you. A murder count incremented on you; it faded away at the rate of one murder every eight hours of real-time gameplay. If you hit five murder count, you were turned red, and became a Murderer.
Murderers suffered severe penalties upon death (scaled by how many people you had killed), were killed on sight by guards (effectively blocking you from towns while red), and also generated bounties like the older system.
Alongside this system were a number of extra features for guilds — they were able to declare themselves Order or Chaos, which replaced the older system, and had their own green and orange color-coding.
The system also put an emphasis on people taking proactive action. Guards no longer just appeared when you attacked someone; instead, someone had to call for them. Thieves had to be “noticed” by someone, and then you could call guards to kill them. It meant that towns actually got slightly less safe.
This system did have a noticeable effect on the amount of playerkilling, but it remained still too high. The sight of Murderers locked in their houses running macros overnight to reduce murder count became a common sight. The stat loss could be significant, so Murderers simply spaced out their kills so they could stay above the threshold of five murder count...
EverQuest had a simple solution to all the above: they had a PK switch. Meaning, unless you flagged yourself as PK-enabled (which few did), you were safe. The end.
This worked acceptably in EverQuest because it was simply a far more constrained game than Ultima Online. You couldn’t drop things on the ground. No laying of traps. No stealing. No houses. No stacking chairs. No chairs. Grief players happily led monsters into newbie zones and killed them, but you couldn’t loot others, so there was no reward cycle there except the gnashing of teeth. There wasn’t crafting. It was a game driven by player-vs-environment combat, and to a veteran MUD player felt very strongly like a DikuMUD with first-person 3d.
The result was an exodus driven not only by the more modern 3d graphics of the newer game, but by the safety. Everything I had thought about the impossible admin load of having a PK switch with a large-scale game was disproven in short order, and players wasted no time in telling me bluntly that I had been drastically and painfully wrong.
The result? In the name of player freedoms, I had put them through a slow-drip torture of two years of experiments with slowly tightening behavior rules, trying to save the emergence while tamping down the bad behavior. The cost was the loss of many hundreds of thousands of players. Ultima Online had churned through more than twice as many players who quit than EverQuest even got as subscribers that year.
Trammeling players
By 2000, I was off the project, working instead on a series of pitches for new MMOs that Origin might make. And while I was doing that, the team’s new design leadership arrived at a new solution that put paid to the early wild and crazy era of UO forever: the Trammel/Felucca split.
Simply put, the map was cloned. One side was termed Felucca, and had the same rules that already existed.
The other was called Trammel, and in Trammel, there simply wasn’t any ability to attack other players. It was a peaceful place.
To this day, this is controversial. I wouldn’t have done it, personally, but there is no question that the userbase doubled once this went in.
Had the game been like this from the beginning, would it have reached even greater heights? I don’t know. We lost an enormous amount of players to bad behavior. But we also gained endless stories and excitement, the stories that people tell and retell to this day. Lord British would not have been killed by a player. The sense of excitement would not have been there. Even the player economy would have collapsed — as indeed, it did almost immediately once goods were largely safe in Trammel.
The names came from the two moons over Britannia. A felucca is a kind of sailing ship. The dictionary says this about the word “trammel.”
noun
a restriction or impediment to someone's freedom of action.
verb
deprive of freedom of action.
That’s basically exactly what happened. We trammeled players, and tamed the crazed flow of exploits and inventions. At that moment, a crucial piece of the virtual world was lost, in favor of the old player switches and safe zones of the past. EverQuest launched with a PK switch, and set the template. Later games would do realm vs realm combat, maybe even some simple criminal flags. But never again would online worlds let you drop stuff on the ground, try to pick a pocket, entice creatures hither and yon, set a bomb, or even just track their good or bad behavior. The dream of letting players police themselves was over.
Instead, it was now the admin’s job.
Within five years, they wouldn’t bother.
Where once LegendMUD would have kicked you out for cursing, we got Xbox Live chat with rampant sexism and homophobia.
Where once Ultima Online would have tracked your behavior to try to warn other players of bad apples, we eventually got Twitter, where Nazis can post freely and spam others off the Net.
By giving up on solving the hard problem of freedom co-existing with civility, I fear that the result is that on today’s Internet, we have neither.
By 2000, I was off the project, working instead on a series of pitches for new MMOs that Origin might make. And while I was doing that, the team’s new design leadership arrived at a new solution that put paid to the early wild and crazy era of UO forever: the Trammel/Felucca split.
Simply put, the map was cloned. One side was termed Felucca, and had the same rules that already existed.
The other was called Trammel, and in Trammel, there simply wasn’t any ability to attack other players. It was a peaceful place.
To this day, this is controversial. I wouldn’t have done it, personally, but there is no question that the userbase doubled once this went in.
Had the game been like this from the beginning, would it have reached even greater heights? I don’t know. We lost an enormous amount of players to bad behavior. But we also gained endless stories and excitement, the stories that people tell and retell to this day. Lord British would not have been killed by a player. The sense of excitement would not have been there. Even the player economy would have collapsed — as indeed, it did almost immediately once goods were largely safe in Trammel.
The names came from the two moons over Britannia. A felucca is a kind of sailing ship. The dictionary says this about the word “trammel.”
noun
a restriction or impediment to someone's freedom of action.
verb
deprive of freedom of action.
That’s basically exactly what happened. We trammeled players, and tamed the crazed flow of exploits and inventions. At that moment, a crucial piece of the virtual world was lost, in favor of the old player switches and safe zones of the past. EverQuest launched with a PK switch, and set the template. Later games would do realm vs realm combat, maybe even some simple criminal flags. But never again would online worlds let you drop stuff on the ground, try to pick a pocket, entice creatures hither and yon, set a bomb, or even just track their good or bad behavior. The dream of letting players police themselves was over.
Instead, it was now the admin’s job.
Within five years, they wouldn’t bother.
Where once LegendMUD would have kicked you out for cursing, we got Xbox Live chat with rampant sexism and homophobia.
Where once Ultima Online would have tracked your behavior to try to warn other players of bad apples, we eventually got Twitter, where Nazis can post freely and spam others off the Net.
By giving up on solving the hard problem of freedom co-existing with civility, I fear that the result is that on today’s Internet, we have neither.
