Undermountain Update
Since my last column we have played 4 sessions of Undermountain and I am happy to give folks an update. If you’re unfamiliar with the Undermountain concept you can read more about it in my last column, but in summary it’s a stripped down version of Pathfinder optimized for simple dungeon crawling and light rules.
There Is No Ruler
The first major change I’ve made since my initial draft of the rules is to eliminate measurements. Undermountain no longer has specific distances for anything. Instead, objects are either within “Melee”, “Near” or “Far” distance from other objects.
There are only 3 movement options:
• “Step”, which is a generic movement of short distance (aka 5 feet) to deal with line of sight, doors, etc. that doesn’t require a substantial change in range.
• “Close”, which allows a character to go from Near to Melee distance (or vice-versa) and still make a Standard Action
• “Run”, which allows a character to take a Full Round action to move from “Far” to “Near” distance.
When an area of effect is used, the effect is “Small”, “Medium” or “Large”. “Small” effects include any objects that are at Melee distance to each other. “Medium” includes any objects that are at “Near” distance, and “Large” can group objects at “Far” distance.
One big consequence of this system is that there are no Attacks of Opportunity and no Threat zones. Likewise, there’s no stacking or facing rules.
We are still using Terrain types, specifically Difficult Terrain, which stops any object entering it and requires a Full Round action to traverse even when making a Close action.
This has added time for a full encounter to our game sessions. Before, in a 4 hour session we were averaging 4 encounters and we are now up to 5. I’d like to get that to 6 and am considering how to make that possible without reducing the simulation of the game level to something so abstract that it is hard to provide immersion.
Know Who Is At Your Table
In an earlier column I talked about ways to profile your players and assign them into psychographic segments based on the way they approach the play of the game. So far I’m still gathering data but I thought it would be interesting to give people a glimpse at my table.
My game consists of 6 regular players: 3 very experienced RPG players, 1 with intermediate experience, and 2 rookies.
My players are:
Delaina Dancey: Attorney at the Saedi Law Group, and my wife. Delaina has played on and off for years and is our intermediate player. Her character is a Half-Orc Warrior who usually serves as the tank. She’s a Power Gamer in this group, usually being the first to rush into the fight and playing her character as impatient to start cracking skulls.
Steve Chambers: Cook at CCP North America. Another of our very experienced players, Steve often helps out as assistant GM remembering various rules faster than I can look them up. His Sorcerer has been instrumental at dealing with some of the toughest foes the party has faced, and since he’s playing his Sorcerer like a Power Gamer, that’s been a good thing.
Paul Jacquays: Game Designer at CCP North America. To say Paul is “Experienced” is like saying Jimmi Hendrix was “Experienced”. Paul has his fingerprints all over some of the earliest work in the tabletop RPG hobby. He’s involved now in the MMO field and hasn’t played a “D&D” game in years but he’s having a great time as the party’s Cleric (aka: Saving Their Asses from the TPK Every Session). He’s playing in the center spot on the segmentation graph so far, not surprising given his long history with the format.
Mike Tinney: President of CCP North America and former CEO of White Wolf Game Studio. Mike is a very experienced gamer who knows the rules well. He’s clearly got a Tactical focus. His character is a Rogue and he works hard to get in position to do Sneak Attack damage.
Cara Waiswilos: Our other rookie. Cara is a sweet southern girl right out of the Bible Belt (“War Eagle!”). She’s jumped in with both feet, playing a Japanese-themed warrior princess as a tank and Power Gamer. It’s been so fun to see the game through her eyes: Her summary of her (first ever) TRPG game session was “This is all about problem solving, teamwork and fighting evil! It was awesome!”
Bill Winter: General Counsel at CCP North America. Bill is a relative newcomer to TRPGs. He’s unfortunately borne the brunt of the party’s character deaths, having walked face-first into a Gelatinous Cube, been torn apart by a Roper, drowned by a Giant Sea Scorpion, and kicked in the head by a number of Ant-Man Monks. Bill's Ranger has done more dying and less living than the rest of the party but he’s been cursed with really unfortunate die rolls. He’s being advised on kharmic die preparation by members of the group. His playstyle is surprisingly Character Acting, which I found unusual for someone who is having to digest the rules as they play.
Lessons Learned So Far:
Game Balance:
Whenever you make as many changes to a game system as I have done with Pathfinder, you’re certain to upset elements of game balance. So far, we’ve struggled with a good mix of damage inflicted vs. armor class. The characters are all the equivalent of 5th level so the Warriors can take & dish out a lot more pain than the spellcasters, and as a result I have to be careful not to let the melee combat accidentally overwhelm the casters while still challenging the Warriors. So far I’ve done that by a lot of die roll fudging, but the last session I did a lot less, so I’m getting an intuitive feel for the right balance (now I just have to get that codified somewhere in writing).
Bonus Complexity:
D20 in Undermountain mode still has too many modifiers. I literally had to make a flowchart to help my new players figure out how to calculate the attack & damage bonuses for their characters, how to deal with critical hits, and how to tell me how much damage they’re inflicting. That’s too bad but so far I haven’t found a good way around it. Of course, once they have the numbers figured out they don’t need to use the flow chart every time they make an attack roll but it still reveals the underlying complexity. And since they’re leveling up next session we’ll have to do it all over again.
Injuries Are Meaningless
In Undermountain, characters don’t die. When they are reduced to 0 hit points or less, they go unconscious. When they are restored to positive hit points they take a Wound, which gives them a persistent injury of a random type. However, the rules are structured so that these injuries are removed between sessions if the characters end the session in a place of safety, or if the characters use a Healing Kit to remove the injury. So far, the injuries have been more about bookkeeping and less about “fun” (in the form of having to cope temporarily with a setback). Either the rules need to be tweaked or the system needs to be junked because it’s not adding More Fun to our 4 Hours.
Some Comments On The Future of Gaming
I just returned from the GDC Online event in Austin TX. There was a lot of talk at the show about AI and improving the way players interact with NPCs. And on my return home I got my new iPhone 4s, including the key new feature Siri, the virtual assistant in iOS 5.
A few months ago I did an interview on the Fear the Boot podcast (Fear the Boot Blog Archive Interview Episode 17 – Ryan Dancey) and one of the things we discussed was how fast AI would improve the experience for gamers interacting with NPCs in on line games.
I thought it was interesting to reconsider those remarks in light of Siri. Remember that because of Moore’s Law, the processing power of a given system will double roughly every 18 months, or the price of that system will halve. You can draw some very interesting 10 year timelines starting with Siri and the iPhone 4s and imagine that it won’t be that far away that NPCs in on line games will be the front ends of virtual agents powered by something like Siri, able to interact with a wide range of player inputs.
Siri is not really an AI of course, it’s just an expert system married to a natural language interpreter. What many people don’t know is that Siri is actually a front end to a large backend server system. Siri does both local processing on your phone, and back end processing in the Cloud to deliver the unique experience of a virtual agent. But think about the fact that Siri has to be limited to some extent by what can be crammed into an iPhone form factor – whereas the agents in an on-line game will be running on the fastest, most robust hardware available without any client-side limits.
It is very reasonable to expect that the next generation of games from today (i.e. the games that will come to market in about 5 years) will have server-side virtual agents designed into their core. It will become essentially unremarkable in the way that 3D and networking have become unremarkable features of current generation games – just another thing that players will take for granted.
This is of course just the tip of the iceberg. The agent’s ability to interact with humans will expand rapidly. I suspect that you’ll have “sidekicks” in games within 10 years that would come close to passing the Turing Test. When you play a game like Dragon Age, for example, your party will seem like distinct individuals with whom you can interact across a wide range of topics and experiences, who learn and change based on the adventures you have together and how you treat them, and who surprise you as they follow their own agendas and character arcs even when they don’t nicely complement your own.
And putting this level of agency into monsters will have a huge effect on gameplay. It’s a bit of an open secret right now that AI in most games is dumbed-down substantially because the tools already available would overwhelm most player’s ability to cope if the opponents were really turned loose. But that’s mostly just brute force applications. Coming soon will be subtlety and deception, a much harder thing to simulate and a much more immersive experience when you confront it. You won’t just be attacked by the monsters, you’ll be betrayed by them.
For Next Time
In my next column I expect to be writing about bringing more “social interaction” gameplay into Undermountain. We’re getting the combat systems debugged and its time to start looking forward to challenging the characters with problems that can’t be solved by just hacking away at anything that moves.
As always, I am interested in your feedback and will do my best to respond in the forum comments following this column.
--RSD / Atlanta, October 2011
Since my last column we have played 4 sessions of Undermountain and I am happy to give folks an update. If you’re unfamiliar with the Undermountain concept you can read more about it in my last column, but in summary it’s a stripped down version of Pathfinder optimized for simple dungeon crawling and light rules.
There Is No Ruler
The first major change I’ve made since my initial draft of the rules is to eliminate measurements. Undermountain no longer has specific distances for anything. Instead, objects are either within “Melee”, “Near” or “Far” distance from other objects.
There are only 3 movement options:
• “Step”, which is a generic movement of short distance (aka 5 feet) to deal with line of sight, doors, etc. that doesn’t require a substantial change in range.
• “Close”, which allows a character to go from Near to Melee distance (or vice-versa) and still make a Standard Action
• “Run”, which allows a character to take a Full Round action to move from “Far” to “Near” distance.
When an area of effect is used, the effect is “Small”, “Medium” or “Large”. “Small” effects include any objects that are at Melee distance to each other. “Medium” includes any objects that are at “Near” distance, and “Large” can group objects at “Far” distance.
One big consequence of this system is that there are no Attacks of Opportunity and no Threat zones. Likewise, there’s no stacking or facing rules.
We are still using Terrain types, specifically Difficult Terrain, which stops any object entering it and requires a Full Round action to traverse even when making a Close action.
This has added time for a full encounter to our game sessions. Before, in a 4 hour session we were averaging 4 encounters and we are now up to 5. I’d like to get that to 6 and am considering how to make that possible without reducing the simulation of the game level to something so abstract that it is hard to provide immersion.
Know Who Is At Your Table
In an earlier column I talked about ways to profile your players and assign them into psychographic segments based on the way they approach the play of the game. So far I’m still gathering data but I thought it would be interesting to give people a glimpse at my table.
My game consists of 6 regular players: 3 very experienced RPG players, 1 with intermediate experience, and 2 rookies.
My players are:
Delaina Dancey: Attorney at the Saedi Law Group, and my wife. Delaina has played on and off for years and is our intermediate player. Her character is a Half-Orc Warrior who usually serves as the tank. She’s a Power Gamer in this group, usually being the first to rush into the fight and playing her character as impatient to start cracking skulls.
Steve Chambers: Cook at CCP North America. Another of our very experienced players, Steve often helps out as assistant GM remembering various rules faster than I can look them up. His Sorcerer has been instrumental at dealing with some of the toughest foes the party has faced, and since he’s playing his Sorcerer like a Power Gamer, that’s been a good thing.
Paul Jacquays: Game Designer at CCP North America. To say Paul is “Experienced” is like saying Jimmi Hendrix was “Experienced”. Paul has his fingerprints all over some of the earliest work in the tabletop RPG hobby. He’s involved now in the MMO field and hasn’t played a “D&D” game in years but he’s having a great time as the party’s Cleric (aka: Saving Their Asses from the TPK Every Session). He’s playing in the center spot on the segmentation graph so far, not surprising given his long history with the format.
Mike Tinney: President of CCP North America and former CEO of White Wolf Game Studio. Mike is a very experienced gamer who knows the rules well. He’s clearly got a Tactical focus. His character is a Rogue and he works hard to get in position to do Sneak Attack damage.
Cara Waiswilos: Our other rookie. Cara is a sweet southern girl right out of the Bible Belt (“War Eagle!”). She’s jumped in with both feet, playing a Japanese-themed warrior princess as a tank and Power Gamer. It’s been so fun to see the game through her eyes: Her summary of her (first ever) TRPG game session was “This is all about problem solving, teamwork and fighting evil! It was awesome!”
Bill Winter: General Counsel at CCP North America. Bill is a relative newcomer to TRPGs. He’s unfortunately borne the brunt of the party’s character deaths, having walked face-first into a Gelatinous Cube, been torn apart by a Roper, drowned by a Giant Sea Scorpion, and kicked in the head by a number of Ant-Man Monks. Bill's Ranger has done more dying and less living than the rest of the party but he’s been cursed with really unfortunate die rolls. He’s being advised on kharmic die preparation by members of the group. His playstyle is surprisingly Character Acting, which I found unusual for someone who is having to digest the rules as they play.
Lessons Learned So Far:
Game Balance:
Whenever you make as many changes to a game system as I have done with Pathfinder, you’re certain to upset elements of game balance. So far, we’ve struggled with a good mix of damage inflicted vs. armor class. The characters are all the equivalent of 5th level so the Warriors can take & dish out a lot more pain than the spellcasters, and as a result I have to be careful not to let the melee combat accidentally overwhelm the casters while still challenging the Warriors. So far I’ve done that by a lot of die roll fudging, but the last session I did a lot less, so I’m getting an intuitive feel for the right balance (now I just have to get that codified somewhere in writing).
Bonus Complexity:
D20 in Undermountain mode still has too many modifiers. I literally had to make a flowchart to help my new players figure out how to calculate the attack & damage bonuses for their characters, how to deal with critical hits, and how to tell me how much damage they’re inflicting. That’s too bad but so far I haven’t found a good way around it. Of course, once they have the numbers figured out they don’t need to use the flow chart every time they make an attack roll but it still reveals the underlying complexity. And since they’re leveling up next session we’ll have to do it all over again.
Injuries Are Meaningless
In Undermountain, characters don’t die. When they are reduced to 0 hit points or less, they go unconscious. When they are restored to positive hit points they take a Wound, which gives them a persistent injury of a random type. However, the rules are structured so that these injuries are removed between sessions if the characters end the session in a place of safety, or if the characters use a Healing Kit to remove the injury. So far, the injuries have been more about bookkeeping and less about “fun” (in the form of having to cope temporarily with a setback). Either the rules need to be tweaked or the system needs to be junked because it’s not adding More Fun to our 4 Hours.
Some Comments On The Future of Gaming
I just returned from the GDC Online event in Austin TX. There was a lot of talk at the show about AI and improving the way players interact with NPCs. And on my return home I got my new iPhone 4s, including the key new feature Siri, the virtual assistant in iOS 5.
A few months ago I did an interview on the Fear the Boot podcast (Fear the Boot Blog Archive Interview Episode 17 – Ryan Dancey) and one of the things we discussed was how fast AI would improve the experience for gamers interacting with NPCs in on line games.
I thought it was interesting to reconsider those remarks in light of Siri. Remember that because of Moore’s Law, the processing power of a given system will double roughly every 18 months, or the price of that system will halve. You can draw some very interesting 10 year timelines starting with Siri and the iPhone 4s and imagine that it won’t be that far away that NPCs in on line games will be the front ends of virtual agents powered by something like Siri, able to interact with a wide range of player inputs.
Siri is not really an AI of course, it’s just an expert system married to a natural language interpreter. What many people don’t know is that Siri is actually a front end to a large backend server system. Siri does both local processing on your phone, and back end processing in the Cloud to deliver the unique experience of a virtual agent. But think about the fact that Siri has to be limited to some extent by what can be crammed into an iPhone form factor – whereas the agents in an on-line game will be running on the fastest, most robust hardware available without any client-side limits.
It is very reasonable to expect that the next generation of games from today (i.e. the games that will come to market in about 5 years) will have server-side virtual agents designed into their core. It will become essentially unremarkable in the way that 3D and networking have become unremarkable features of current generation games – just another thing that players will take for granted.
This is of course just the tip of the iceberg. The agent’s ability to interact with humans will expand rapidly. I suspect that you’ll have “sidekicks” in games within 10 years that would come close to passing the Turing Test. When you play a game like Dragon Age, for example, your party will seem like distinct individuals with whom you can interact across a wide range of topics and experiences, who learn and change based on the adventures you have together and how you treat them, and who surprise you as they follow their own agendas and character arcs even when they don’t nicely complement your own.
And putting this level of agency into monsters will have a huge effect on gameplay. It’s a bit of an open secret right now that AI in most games is dumbed-down substantially because the tools already available would overwhelm most player’s ability to cope if the opponents were really turned loose. But that’s mostly just brute force applications. Coming soon will be subtlety and deception, a much harder thing to simulate and a much more immersive experience when you confront it. You won’t just be attacked by the monsters, you’ll be betrayed by them.
For Next Time
In my next column I expect to be writing about bringing more “social interaction” gameplay into Undermountain. We’re getting the combat systems debugged and its time to start looking forward to challenging the characters with problems that can’t be solved by just hacking away at anything that moves.
As always, I am interested in your feedback and will do my best to respond in the forum comments following this column.
--RSD / Atlanta, October 2011
