Discuss the massively-multiplayer home defense game.
You are not logged in.
Announcing the preview period for the Dutch auction of one-of-a-kind non-fungible crypto tokens of the pixel paintings from The Castle Doctrine. Only one person can own each painting. View the collection here: https://opensea.io/collection/thecastledoctrine More details here: http://thecastledoctrine.net/crypto.php
I'm curious about how much interest there will be in these, because they're not cheap to mint.
On the peak day, there were 3,612 unique people playing the game in one day. On that day, houses were broken into 81,051 times.
Over the life of the game there were:
$214,622,597 worth of in-game money stolen
1,858,757 robberies committed
951,296 security tapes watched
720,989 player deaths
The most active player lived 7,769 separate lives in the game
1454 players lived more than 100 lives each
78 players lived more than 1000 lives each
3 players lived more than 5000 lives each
It's happening tomorrow on Steam, after six long years.
Well, a mobile port is beyond the scope of what I have time for, sadly.
I could set up "emails when you get robbed" pretty easily, but that won't work for Steam players (I don't have their email).
Okay, thanks folks!
Both issues have been fixed now.
First, the value of the tool is no longer a surprise at first (it was a weak surprise anyway, because the value estimate was recomputed after the first visit). I don't want to muck with the code that checks houses back in, so instead I'm just getting rid of the surprise. The value estimate is correctly computed for abandoned houses from the very beginning now.
Second, for every 24 hours that an abandoned house lives on the list, extra seed money is added. So, they will eventually become "worth it."
Maybe 24 hours is too slow, but in the first batch update, 33 abandoned houses got extra money.
I don't want it to climb too fast, because I want players to be conducting a careful value calculus.
So, see how it feels now.
Who wrote that really long review?
One of the best things written about TCD ever... :-)
Okay, I joined.
Needs a logo and such.
There were 72 active players in the past 24 hours.
The last time we had this many active players was in Februrary of 2015, Almost two years ago.
Furthermore, there are currently 68 real, non-abandoned houses on the list.
Yes, the house list is padded out to 100, but only 32 abandoned houses are needed for that padding. The rest are REAL.
I've been thinking more about the way this game degenerates in low-population situations. Part of the problem is that The Castle Doctrine is a content-eating machine. Players design and make the content in the game---the houses---and players consume that content by robbing those houses. The problem is that making a house involves a significant time investment, but that same house can get "used up" and removed from the game very quickly. All it takes is one clever player robbing it, and it's gone. Or the owner can make a careless mistake and die, taking the house with them to the grave.
But the bottom line is that each piece of content only serves one other player in the end---the one who successfully robs it. And that's in the best case scenario, if the owner doesn't die first. And of course, most players would want to rob and beat more than one house. Each player creates only one house, but wants to rob multiple houses. That equation simply doesn't add up. The content demand is way greater than the supply.
In high population situations, this issue gets smoothed over somewhat in the shuffle. But in low population situations, the problem is glaringly obvious, most importantly because the house list itself gets so short, and the few remaining active players effectively run out of "game" to play.
Many solutions to this problem have been floated over the years. The most obvious one was NPC houses of some kind. Either state-designed dummy houses, or recycling dead player's houses, or something like that. And if you think about all the player-generated content that has been wiped over the years, it is kind of a tragedy.
20,000 people have played the game, many of them generating dozens of houses each. So, potentially 50- to 100,000 houses have been created over the years. And only 14 are left!
But it was important to me that everything remain real in this game. You're robbing a real person, and that person may someday watch the security tapes to see what you did. So, every house has to have an owner, or else the "real" aspect goes out the window. You can't be robbing an NPC or a ghost or a robot.
And the lost houses have been purged from the database anyway, so even if I wanted to use those 100K lost houses as content, I can't really.
However, there IS a collection of "real" houses in the game that have not been deleted. Houses that will potentially be sealed in stasis for an eternity. Houses that still have living owners, but have been robbed down to $0, so taken off the list. Abandoned houses.
The owners could come back someday and watch their final tapes, and maybe even fix up their houses and bring them back online.
And get this.... there are over 4,000 of these abandoned houses in the database! Content just sitting there, unplayed and unseen.
So, what if some squatters come by in the night and leave a few valuables in some of these houses?
Yes, many of those 4,000 houses are abandoned default houses with absolutely no challenge, but I have a bunch of stats in the database for each house. What about houses that have killed at least one robber in the past? That's a good filter, and there are over 1,500 of those houses. Some of them are in pretty rough shape, with walls cut all the way to the vault. But others are still mostly in tact, with puzzles still present. All of them are somewhat unnerving to go poking around in. Dogs are still around, and they're extra hungry after all these years...
So, the house list is very full now, padded out at the bottom with abandoned houses that have been laced with a bit of loot by squatters.
And in terms of realness, the houses are still owned by someone, and they may come back someday and see the mess that has been made over the years.
Think of it as Las Vegas, post-housing-crash. A lot of empty houses in the neighborhood.
For those of you who are still actively playing the game, but taking occasional breaks, the squatters won't consider your house unless you've been gone more than a week. And you, the owner, can always come back after the squatters have been there and claim whatever they left behind. Abandoned houses can also still build up bounties if robbers die in them.
Over time, the push-over abandoned houses will get robbed clean and taken back off the list, while the harder houses will tend to stick around longer.
This change will also inject quite a bit of extra money into the economy for the remaining active players. Homeowners will build stronger houses---which I'm not too thrilled about---but robbers will also be more well-equipped. Thus, the arms race among existing players will no longer stagnate. Whoever robs more of these abandoned houses will become a powerful threat.
And by the way, Zed, if you're reading this, your SDF mailbox is bouncing... out of disk space?
Okay, this is done.
A few little wrinkles:
1. Only houses that haven't been touched by the owner in more than a week are considered abandoned. So if you're actively playing, you won't have your house seeded by the state.
2. The amount seeded in a house is a random value between $50 and $200 (so there's not a list of 100 houses all with the same value).
3. A random tool item is included in each vault. The value of this item is NOT included in the displayed value estimate, so it's a surprise. The $50 house might contain a $2400 crowbar... This is a bit of a slot machine element, and goes against my usual sensibilities. However, robbing a real house always had this element, because you never knew what was in the vault. In terms of hiding the true value, which is NOT done for real houses, it was too weird to see an abandoned house way high on the list when it got seeded with a $2400 item. Better to leave the item as a surprise. It is also fitting thematically. Some squatter was in the house and left a bit of money around... and oh look what else they left behind! I've explored abandoned buildings in real life, and it does have that surprise feeling to it.
The game currently tries to maintain a list of 100 houses. In the first run, there were 14 houses around, so 86 abandoned houses were seeded.
And, messing with the game myself, some of these houses are still great houses! Some have no walls cut, and are still the same puzzle they always were. Some are scary and risky.
Can you folks check it out and see what you think?
If it's good, I'll post a Steam update and also email the player list.
Immhotep, I think what Zed is saying is correct. I'm pretty sure that a house with exactly $0 left in it drops off the list, even if it has never been robbed. So, to protect your family, build exactly $2000 worth of walls and then exit the game.
Zed: I'm not sure that my proposal would solve this problem. Currently, I don't know that any such houses exist:
select count(*) from castleServer_houses where value_estimate = 0 and rob_attempts = 0;
+----------+
| count(*) |
+----------+
| 0 |
+----------+
Whoops... both Zed and I are wrong here. A zero-value house that is freshly edited WILL go onto the list. From the WHERE clause:
AND ( houses.value_estimate != 0 OR houses.edit_count > 0 )
So it tracks how many times the house has been edited, and flips negative after a robbery, but flips back to positive after it is visited by the owner and edited/tested again.
And edit_count is defined thus"
// times edited since last successful robbery
// = 0 if never edited (and not robbable at all)
// > 0 if successfully edited and robbable
// < 0 if successfully robbed at least once and still robbable
Hmm... I think I could only use abandoned houses that had at least one robber death.
There are 1574 of those. Which is enough!
Your method of picking them is sound. While there are less than N houses with > 0 value, pick a 0-value robber-killing house at random and add $M to it.
Yeah, probably not going to reset the puzzles at all.
Not sure where the money is coming from, exactly, story-wise.
One idea is to put extra money into houses that have orphaned children. Either they are out scrounging, or maybe receiving state aide?
There's an edge case where someone commits suicide in a starter house, logging a robber death that will put a too-easy house on the list. But players can (and do!) do that anyway with their real houses by running out to rob before building anything.
Yeah, you're probably right.
Can you think about this (and anyone else who wants to) and report back with any lightningbolt ideas?
I'm tempted to just put all 4000 of them online with $100 each in them and see what happens.
I think the harder ones will naturally collect bounties, just like real houses, right?
So that's how to get it started.
But how do we keep it going? When do the houses get more money after they're cleaned out once?
What about dog houses?
Over time, stray dogs move into these abandoned houses, making them more and more dangerous.
Instead of placing a random egg, we challenge players to get to the vault, but place random extra dogs around the house.
If the real owner ever comes back, they get to keep/move/sell these dogs.
Oh, but randomly-placed loot in abandoned houses...
...which makes some thematic sense--squatters have lived there over time and left something behind, separate from the vault...
...has the side-effect of making the abandoned house game a kind of randomized egg hunt. Which feels very different from the main game. In the main game, you're going into a real person's design and trying to analyze it, looking for patterns, weakspots, or clues. In a random egg hunt, you're just cutting every wall randomly, hoping to find the right spot.
Anyway, I still don't see how putting money back in the vault is going to work, because the vault is usually easily reachable...
Also, the vault is reachable again in exactly the same way later, once the house re-appears again. So players who memorize all the abandoned houses can farm them repeatedly with no trouble.
Given that there are 4000 of them... I guess they could be cycled through in an order that makes memorization unlikely. Only the oldest houses get money added to them, and a house that is successfully farmed goes to the bottom of the list. But if we do it with vault-reaches, then the damage will accumulate in these houses over time... they'll get more and more run-down and easier to get through.
So, I'm looking for some smart ideas here.
Remember, these 4000+ houses still have living owners! They just haven't been back in a while. So there's still someone who can see the tapes, eventually. So that's better than ghost houses or contest houses.
And the conservation of money will be an issue, but it'll just push the game up to a higher level. Both robbers and home owners who are farming will have better stuff. I'm not thrilled about that prospect... it could be reduced by making money accumulate in these abandoned houses at a very slow rate.
Well, by "not the real game" in the case of TCD, I also was referring to ways to make money and climb the home defense ladder through income sources that aren't real.
When I look at the existing house pool, I already see that most of the houses look way too well-funded and well-built. Those kinds of houses aren't that interesting to poke around in.
And that's WITH the conservation of money that's currently in place.
If there's some "campaign" source of money (even if it is real houses that have been abandoned, but we're adding state money to them over time), then players can grind that campaign indefinitely and build up however much money they want over time.
The game will then become, "Who can play the campaign the most/fastest/best," because that will be the real way to get a lot of money and build up your house. And there's no end to it. Even if you're the only player playing, you can keep running that treadmill until your fingers fall off. Come morning, your house is going to be an impenetrable fortress. AND you're going to have so many great robbery tools that you'll be able to wipe out every house on the map.
And yeah, money could trickle into these houses at a very slow rate or whatever.
But it still violates a conservation of money principle that keeps things from getting out of control in the game.
And the "real" players would be getting so much money from this that all the real houses would be impenetrable fortresses, making robbing the abandoned houses the only game in town...
Anyway... the interesting thing here is that all those "abandoned" houses aren't gone! They're all still there, all the way since the very beginning. They're just unlisted, waiting for their owners to return.
Guess what? There are 4,045 of the things.... hundreds of pages worth of houses.
And listing them all would involve a few lines of code.
My intuition would be to just list them all at 0 and then trickle money into them.
But here's the thing... most of them would be cake-walks, having their main walls cut.
I don't know how to turn them into "good" puzzles.
Their last self-test map is also saved (the last time they actually got through their own house without tools).
So this could be used instead of the final state. But that definitely doesn't feel as real. Abandoned houses should feel abandoned, not squeaky-clean again.
The other idea is to hide something in each house. The vault is empty forever, but there's a state-sponsored treasure in each house, in a random location, that grows in value over time.
Over time, people will cut through all the walls looking for it... but houses still reset as usual when a robber leaves without reaching the vault.
The other idea is to just start by showing all the 0 houses and see what happens. People may poke around in some of them and die, getting bounties started there...
That sounds like a great design there, Zed.
But the question is this: what are we trying to do here? Give bored players something else to do when the "real game" isn't happening currently?
I guess that helps somewhat with player retention, which may fight the vicious cycle and lead to eventual player growth over time.
But, any of these solutions (including abandoned houses, or challenges) also takes players out of the real game for a while. They aren't part of the real robber pool while they are robbing an abandoned house or competing on a contest house.
This is the equivalent of adding AI or dummy practice opponents in Cordial Minuet. It would give the lone player something to do, but have the unfortunate side-effect of occupying them, potentially indefinitely, so that even when they are "needed" by other players, they aren't available. Imagine a 3-player situation in CM. Two players match up, and the third is waiting around. Now engage that third in a heated AI practice match. Now the fourth player shows up... now what?
I've been working for months improving the single-player experience in One Hour One Life. But it's fully integrated into the multiplayer experience. It's just a facet, not a separate mode. So if you're playing the "single player game," and someone else joins, well, they're born as your baby. You're always there for them when they need you for their multiplayer experience.
In other words, there is no "campaign" mode.
And it sounds like these fixes involve essentially adding a campaign to TCD.
(Also, you can imagine in TCD or CM some sort of "real player is present" alarm that pulls you out of your campaign mode when a real opponent is ready or a real house is added to the list.... but that's a whole additional user experience design problem.)
So can we imagine a fix to this problem that doesn't feel so campaign-like?
We have talked about ways to make one piece of content last way longer.
Multiple vaults is one way. Say you get 5 vaults to place, each of which holds 1/5 of your money. Obviously, this lacks the design clarity of a single vault...
I have thought about this.
There are two problems:
1. It obviously makes everything "way less real," in that some of the houses you're robbing no longer hurt a real person. No one is watching the tapes, etc. Of course, if there simply aren't enough players, then the "real" thing goes out the window anyway, so perhaps I'm clinging to a wasted vision of realness here (how real is it if no one is playing?).
2. This would have the effect of injecting a bunch of extra, "free" resources into the game. People could rob one of these abandoned houses, get some money, spend it on tools to rob another, and repeat, forever, until they had as much money as they wanted. They would obviously use all this extra money to reinforce their own house. This means that all the "real" houses would get harder and harder very quickly. And if the "real" houses served as the pool of abandoned houses later, then even the abandoned houses would get harder and harder. I'm pretty sure things would spiral out of control pretty quickly.
TCD has a very tight and fragile balance. You want people to be heavily constrained when it comes to building their houses. Something as simple as salaries throws the balance way off.
Yes, in the old days, there were 100+ houses to rob---wasn't that a limitless pool of free resources? It seems like it might have been, and the very best players certainly built monster houses. However, that pool was always connected to the number of players. Not only more houses, and more resources, but also more robbers! And more wealthy robbers. If there were suddenly 30 extra houses, there's also be 30 extra robbers trying to get into your house.
The suggestion of adding extra "content" to the game in isolation means more resources for building defenses without an increased threat to go along with those resources.
And yes, the few extra players who also have access to the extra resources would be more formidable as robbers, so maybe it would balance out somewhat. But more likely, the whole thing would just stagnate into fortress-ville. Why would you go after one of these top houses when you could farm abandoned houses forever?
Right now, money enters the game through two sources:
1. Fresh life players get $2000 (but to prevent farming these repeatedly, a player's next house is invisible to you for a while if you robbed their last house).
2. You get bounty payments for tricking other players into dying in your house. More money based on how successful they were elsewhere before they were tricked by you.
Notice that both sources are accompanied by a proportional threat from other players.
Perhaps this discussion is academic at this point, given that very few people are still playing the game. But I worry that a quick fix won't work. If it throws the balance off, we'll be right back here in a month or two trying to fix it again.
If we really want fake houses, I'd suggest that we need fake robbers to go along with them. Anyone interested in taking a crack at a TCD bot?
You're right about all that lost content.
Some of it is cached in the database (from the last 30 days or something), but eventually, it's purged. So I'm pretty sure I don't have 4+ years of houses in the archive somewhere. That would be pretty cool.
And you're right about the great curve that was present in the old days. You could go low down on the list and find easier houses to try, which is what they did in that Rooster Teeth vid.
Keeping abandoned houses around is also problematic because once you know how to beat it, you can keep beating it over and over later, after more money builds up. Or are you allowed to beat it only once?
There are certainly some interesting design problems there that would be fun to solve.
HOWEVER, it still cuts to the core of the game, which is that everything is real. When you rob a house, someone real is hurt by your actions, and they may potentially watch the security tapes and take revenge on you.
Keeping abandoned houses around will reduce that feeling.
Of course, it seems clear that keeping that "real" feeling around is at odds with a game that actually functions over the long term. Seems like there'd at least have to be some kind of separation.... like the abandoned houses would have to be on a separate list.
There's also the problem of ever-growing wealth that will naturally make houses harder and harder. Abandoned houses keep injecting more money into the game.
In working on my new game One Hour One Life, I've been doing a lot of thinking about how to make the game function well for smaller player populations. I've realized over the years that for multiplayer games, low population playability is crucial for building and maintaining critical mass. If the game isn't interesting for one player, it will never grow to two players, because that one player will leave before the second player has a chance to join. Even if you start with a larger population (during the Castle Doctrine launch, there were 3600 players on the peak day), the game will generally get slightly less interesting for each player that leaves, resulting in a mild vicious cycle that affects all multiplayer games. But if the game really falters for very low populations, once the game hits the bottom during a lull, it can never recover.
The Castle Doctrine and Cordial Minuet both suffer from this problem in a big way. Neither game is playable by a solitary player. The Castle Doctrine is a bit better, because it can be played asynchronously by a small population of players that are not online simultaneously, but Cordial Minuet absolutely requires two simultaneous players to function at all, and even worse, it requires pairs of players. A third player is stuck without a game partner and is thus motivated to leave before a fourth player arrives.
All that said, I've been thinking about how to improve The Castle Doctrine in this regard.
One of the biggest problems with low populations on the server involves chill timers. After you rob a house and die trying, the house has a chill on it that prevents you from re-entering it in your next life for a while. The standard setting on the server has been a two hour chill. The point of this timeout is to prevent you from spam-robbing a house carelessly across multiple lives. You can't just randomly try every corridor while dying and respawning over and over with additional throw-away lives. The chill timeout is long enough to make each attempt matter.
However, there's a problem with this long timer when the house list is very short: you very quickly run out of houses to try robbing. Aside from building your starter house, robbing is the main activity, and when you run out of houses due to chill timeouts, you essentially are forced to stop playing for a while. When the list is very short, this can happen in just a few minutes. Bam. Content of the game exhausted. Better quit and come back later. This only serves to strengthen the vicious cycle described above.
So, I've implemented a simple solution to this problem. The chill timers are now adjusted based on the number of available houses, with five minutes added per additional house that's available to you, up to a max of 2 hours for very long lists of houses.
It works like this, for the number of houses not including your own:
house_count chill minutes
1 5
2 5
3 10
4 15
5 20
6 25
7 30
...
The idea is that, if you are conducting legitimate, non-spam robberies, you will spend about this many minutes working your way through the full house list, and by the time you've tried every house, the chill timer on the first house that you tried will be about ready to expire.
Thus, the game's content pool won't be forcibly exhausted for you as a robber, pretty much ever.
That is, of course, unless you successfully rob every house in the neighborhood.
Also, I realize that many people mitigate that pain by using second accounts.... there's nothing that I can do to stop that. BUT, I still don't want to build that directly into the game. It's not the way the game is meant to be played.
Hi there, Storm! Thanks for trying to revive the game.
I've been busy with One Hour One Life, so I haven't been checking these forums...
But I just happened to check them today BECAUSE I just implemented a server-side fix to TCD that should help with the small player pool. It changes the way chill timers (purple houses after you die in them) work. I'll be posting more about that in a minute.
ANYWAY....
Regarding your suggestions. Way back in the day, we DID have "salaries," and there's a server-side switch in there for turning them back on. However, we found back then that they didn't work, because they encouraged people to build "scary" facade-type houses and just "turtle" until they built up a lot of money through salary.
An example facade house is a solid wall with just a single, simple commit gate. No one is going to try walking across that gate to have a look around.
Those scary facade houses aren't interesting content for the game. So, the game was changed so that the only way to earn money was by tricking people into dying in your house and collecting their bounties. Thus, you had to make a house that looked inviting and was filled with psychological tricks if you wanted to grow your wealth as a home owner.
Of course, you could still rob to get money, but there WAS a viable "homeowner" path to wealth through a very well-designed house.
I don't know that small-vs-large player pools changes the effect that salaries would have in terms of encouraging people to turtle inside scary, uninteresting houses.
As for your second point about second chances....
Well, one of the main points of this game is that "everything is real" and that permadeath is waiting for you around every corner. That amps up the tension and drama. It was a core aesthetic of the game from the very beginning.
I don't doubt that it makes the game "worse" in terms of retaining players or keeping the world populated with player-generated content. It causes people to rage-quit, etc.
However, I feel like I need to stick to my vision-guns on this one.
The main problem that I see with it is that it destroys game content along the way. If you build a great house and then die due to a robbery mistake, that piece of content that you made is removed from the game along with you before other people get to experience it.
I've thought about keeping "abandoned" houses on the list until they are beaten. I'm not sure how to keep the owner away, though, because obviously, they'd have a huge advantage in the race to get to the vault in the house that they designed. Yes, I could hide it from the original owner's account, but this would just further encourage multiple accounts. Imagine dying and losing a week's worth of loot in your house.... you'd be pretty motivated to cough up another $16 for a second account to get that loot back.
There's a general problem in the game that's even bigger than this content loss issue, and that's a general content shortage. Everyone makes only one house per life, but most people would like to successfully rob many houses per life. That means that there aren't enough houses to go around. So, keeping abandoned houses on the list would help a bit, but it wouldn't completely fix the problem.
Aside from the content loss to the rest of the game world, the incredible pain of losing everything is part of the game.
Well, I feel like we've kinda hit a wall here.
We want you to only succeed in the game by outsmarting other players and getting your reward for this directly from them. This makes everything real.
But we have no way to know for sure if the player you're besting is really another player or is actually just a puppet under your control.
The alternatives involve somehow NOT getting money directly from other players. There's an NPC fountain of some kind that doles out money (lottery, contest, captcha).
It's hard to envision how this source of money, given that it's the only direct source in the game (players can't create money themselves) won't overshadow the indirect sources of money (robbing) that make up the "main game" or the "game we want."
Thus, there needs to be a way for players to create money, but it can't be free money.
The only solution, which partially how I solved this problem, is to have creating money cost REAL money. In the original game, two accounts cost twice as much. The problem was that after you bought two, you could essentially create an infinite amount of money over time and pump in into one account.
It seems that, if each new life cost real money, and you could pick how much you start with based on how much you spend on that life, the problem of multiple accounts would go away.
But this is STILL subject to non main-game grinding. If you wanted to get richer in the game, you'd just go grind more in real life by working extra hours so you'd have more real money to pump into the game.