User:LadyPavitra

I'm Pavitra of Tonga in-game. Hi.

I happen to have lots of opinions about how the game should develop. Some of them may even be right.

Intriguingness
In my opinion, a good strategy game is founded on nontrivial decisions. It has been suggested to me that I just like making things complicated, but while complication is often a consequence of making the game intriguing, I don't see it as an end in itself. For example, I see most of Baaltar's suggestions as adding complexity, perhaps realism, to the game without actually making the game deeper.

Indeed, this is endemic of most of the suggestions I see on the forum. People call for more "realistic" features, or more frippery things to play with, or more mechanics focused on minigames and RPG-style grinding. This is a bad idea, and will kill the game.

There are too many online browser-based strategy games built on those mechanics already. RoS cannot compete with the existing entries in that genre directly. The giants of the genre have the advantages of age: vast user bases, years of development, pre-established mindshare. They are deep-rooted.

For RoS to survive, it must differentiate itself. If the best one can say of the game is "I like it, it reminds me of / is inspired by / is shaping up to be almost as good as big famous game X", then eventually its players will drift away. If, on the other hand, we as players can say "I like it. It's not like big famous game X, it's something really unique", then people will come to it, and attach strongly to it, and stay. Then our game can survive for a year, three, five, ten.

Every major landmark computer game in the last thirty years has started out that way -- daring and dangerous, unlike anything else.

Political Dynamics
Both for the sake of intriguingness and to focus the game on human-to-human interaction, I have generally been thinking in terms of the kind of intense social dynamics of Werewolf/Mafia -- alliance, trust, distrust, betrayal -- but with the following major differences:


 * Werewolf is best played with about 7-12 people. RoS will eventually be a game for hundreds or thousands of people.


 * Werewolf thrives on short-term psychological tension, and has defined goals. RoS is long-term and has neither defined goals nor end, and therefore needs more different types of interaction.

The formation of international governments and the bureaucratic and political complications that accompany them, emergent economic interactions and structures, personal friendships, and codes of honor all are possible in such a game -- all make such a game possible -- and all serve the ultimate purpose of creating a dynamic and engaging game.

It is not enough merely to enable complex interactions, however. The possibilities must be made transparent. It is my opinion that the actual game rules should be as lean as possible while still making the interactions happen. Every minute I spend playing minigames is a minute I spend on my own, interacting only with the server. It dulls the hot wire of edged-teeth politics, forces me to spend time not playing the real game.

Complexity in the game rules should be introduced only as necessary to introduce complexity in the political landscape, and then in the form of player-to-player interactions. I should spend as little time as possible in the cold void with only the server for company, alone.

If I am speaking only to a computer, then it is easy for me to walk away. Let me, make me, interact with people.

Gratification and Investedness
I realize now that it is not enough for a game to be intriguing. Only mathematicians actually play the [iterated prisoner's dilemma] for fun. The payoff structure of a strategy or RPG game hooks into the deep-rooted sense of reward and effort. It feels like earning assets.

Exponential growth of wealth is good, and necessary. I have come to believe in a "three-day horizon", that major achievements and milestones, signs of accomplishment, should occur about two to four days apart for a dedicated, active nation. This is short enough to give the sense of immediacy, of swift progress, that active players crave; and it is long enough to give the sense of waiting, of distance, that makes covering that distance feel like movement. It's just long enough to feel nontrivial, so that making it feels like accomplishment.

Having given players the ability to gain assets and the psychological sense of having earned them, the game is then in the position to properly screw with players' emotions by giving them cutthroat social dynamics surrounding those assets.

(Caveat: exponential growth makes balancing the game much harder. I suggest that each major milestone produce something that makes building that milestone easier but is otherwise mostly useless, so that the nations that are "ahead" can trade it to the nations "behind".)

Incremental Changes
Large overhauls to the game, sudden major changes, make people feel that they're no longer playing the same game that they've grown familiar with, that the game has been replaced rather than modified. The effect is disengaging.

Some people may be sitting back and taking the time to understand the new system. Others may find that the changes make it easier for them to walk away.

It's better to make small-but-significant changes, to change the game incrementally over a period of weeks, to let every update feel like an update rather than an overhaul.

The ship-of-Theseus paradox is a purely psychological one, but the psychological effect is real.

Specific Recommendations
Some of these aren't originally from me, but are things I merely agree with.

Duck Hunting
Something about this has got to change. It's a ridiculous time sink, and it conflicts with IRC and the forum.

On the other hand, it's brief enough that I remain conscious of the forum and IRC the whole time, so perhaps the decision between ducks and conversation is worth forcing on players. Especially when the clock is ticking on harvesting crops on your land.

But not everyone cares as incorrigibly as I do about forum and IRC. As Andrew Plotkin observes on the topic of the Selenitic maze in Myst, players will generally take the less fun, less challenging route over the more fun, more challenging one -- and then hate the game because it isn't fun.

I care enough to pay attention to IRC at the expense of my duck shooting. Most people won't.

Hourly Timeouts
You should only be able to play N hours worth of duck hunt each day, and only be able to harvest-for-one-hour N times per day. The exact value of N is debatable, but should certainly not exceed 12.

To some extent, hourly timeouts are a good thing. They encourage people to stay on the game's site continuously for longer than it takes to log in, play some perfunctory minigames, click the buttons, and log out. But the game also needs to move with the natural circadian rhythms of waking and sleep. It should shape itself so as to become part of the player's regular daily schedule.

I think N should be about 6. Six one-hour harvests, two three-hour harvests, and one twelve-hour harvest daily, spread out as the player likes.

Sudoku
At this point, it would not be politically feasible to remove Sudoku. However, it should be made not to replace the forum. It may not have replaced the forum on a strict game-mechanical level, or in practice for all players, but I do believe that the forum has suffered from it.

I recommend that Sudoku provide an amplifier, a multiplier, for the forum. A person must still post to earn Knowledge; but with Sudoku, the same number posts can earn more points, or fewer posts can earn the same number of points.

Sudoku is susceptible, in the case of certain skilled players, to grinding. It should either be capped, like duck hunting, or made to yield decreasing returns, like the Finance skill.

Again, I am wary of all minigames, because they take time away from people. Perhaps Sudoku could be replaced with a multiplayer mind game like Psychological Ju-Jitsu, or maybe something from game theory like Colonel Blotto with S&gt;12.

The Harmonic Series
Yeah, I know it was my idea in the first place. I was wrong: it levels off too quickly. There's little practical difference between increasing interest rates by .001% per finance point and increasing them by 0% per finance point, especially since it's not possible to increase finance by more than two points per day. If it were possible to grind finance, like it is with knowledge using sudoku, or even like it is with knowledge using the forum, then I would feel differently. But this is just pointless.

Obviously, though, a linear progression is even worse -- and capping finance (or interest) is an undeserved nerf to a good game mechanic.

I recommend using 1+sqrt(finance) instead: 1% for 0 finance, 2% for 1, 3% for 4, 4% for 9, 5% for 16, and so forth. (Non-square values of finance should yield appropriate noninteger percentages, of course.)

The Forum
The scoring mechanism for posting should somehow distinguish between good posts (which are worth possibly several points) and mediocre posts (which are worth only one or even zero points). (I do not include bad posts because spam is culled and punished by moderators.)

There are a few major options:


 * Users can vote individual posts up or down. This has the disadvantage of sucking up the users' time in countless individual acts of moderation, and also has the possibility for a post to be voted down simply because the voter disagrees with the content, or is at war with the poster.


 * Posts are scored according to how many replies they receive. I personally quite like this one because controversial posts, not just popular ones, are scored highly -- it encourages people to create discussion and debate. Also, it automatically distinguishes many different degrees of quality in a post. It has the disadvantage, however, that it may be difficult to determine who is replying to what. If the forum were more branchily threaded, like Usenet or an email mailing list, then determining this would be very easy, but making the forum do that is likely nontrivial.


 * A post scores one point unless its text matches any previous post. This is simple, elegant, and fair, and the algorithm has an open-source implementation. Its main disadvantage is that it only distinguishes two degrees of quality: yes or no.


 * Currently, a post is not scored if it is shorter than 15 characters. This is... not actually bad, in that it doesn't introduce any new problems that just scoring every post alike didn't have, but it's not much of an improvement. At minimum, the character limit should be increased. 15 characters is fewer than "yes, i definitely agree". 128 feels about right -- it's a short paragraph, or one or two longish sentences. This paragraph is 457 characters long.


 * Perhaps a post could earn ( base2log(length in characters) - 6 ) points. (1 point for 128 characters, 2 for 256, 3 for 512, 4 for 1024, etc.) This has the disadvantages that with diminishing returns the rational strategy is to write many moderate-length posts rather than a few in-depth ones, and that length is a generally mediocre predictor of quality anyway.


 * It was recently suggested that extra points could be awarded for a user's first post in each hour (or other period of time -- I recommend 20 minutes). This has the advantage of encouraging a healthy pace to posting, spread out in the pattern one would expect of an actual dialogue rather than swooping in and replying to every thread once a day. The disadvantage is that it doesn't directly measure the actual quality of posts. This proposal might work well in tandem with one of the others in this list, as its extrinsic approach is sufficiently different that it could be complementary to the other proposals' quality-detection schema.

On Ethics and Game Design
I believe that there are five distinct heirarchical levels of ethical reasoning, each higher than, and subsuming within itself, the last (with the exception that 0. Malice is discarded rather than subsumed).

This discussion may not seem directly relevant to the overall topic of this page, but I believe that it is useful in order to help explain why I behave the way I do, why I care about the subjects here under discussion.

0. Malice
A person acting from this level of reasoning treats suffering as an end in itself, even going out of eir way to harm others, causing suffering even at cost to emself.

1. Depraved Rationality
A person acting from this level of reasoning acts solely in eir own self-interest, with no regard for those around em.

2. Honor
A person acting from this level considers eir actions as part of general patterns and precedents, reasons using superrationality, and generally respects rules and tradition.

3. Legislature
A person acting from this level realizes that the rules themselves are mutable, and seeks to construct customs and precedents so as to bring about the society e believes should exist.

In level 2, the particular customs and ethics are taken as given, and honor consists merely of adhering to them faithfully. Level 3 reasoning involves thinking about the purpose of ethics and deriving for oneself a system of ethics and a code of honor.

Note that this level is partly defined as subsuming the ones before it; a person who attempts to dictate ethical rules without emself adhering to them is a level 1 hypocrite.

4. Game Design
In level 3, the moral axioms of ethics -- what is desirable for human life -- are taken as given, and legislature consists merely of seeking to achieve them. (For example, given the axiom that one should minimize human suffering, a level 3 ethicist might deduce that one should give food to the hungry.) Level 4 reasoning involves thinking about the purpose of morality and deriving for oneself a set of moral rules.

A person acting from this level sees human nature and perhaps even the laws of nature themselves as mutable, and seeks to bring about the best of all possible worlds.

Note that this level is partly defined as subsuming all the others; a person who attempts to influence the game rules, but who does so from a level 1 state of moral reasoning -- manipulating the rules to eir own personal benefit -- without reference to levels 2 and 3, is really only functioning from level 1.

Mere mutability of the world does not make one a game designer in the ethical sense. Without honorable legislature, one is merely a level 1 sorcerer.

Footnote: human nature is mutable because you can decide what to award points for. Do you think people would reproduce nearly as much if it didn't give you about ten thousand points of Mood? :P (If you're under 18 (21 in some jurisdictions), then you didn't see that last sentence. Move along. Go play tennis or something.)

I strive to act from the last and highest of these levels.

Most of the players here will likely vary between levels 0 and 2, attempting first and foremost to win the game, or perhaps to recommend that some particular aspect of the game be changed to be easier for them without considering the larger effect of that change on the strategic shape of the game as a whole.

This is not necessarily wrong; it is normal, and in fact the possibility of game design as such generally depends on such players forming the majority. But I am not such a player. My character and nation are secondary to me; my first priority is the game design.

That is why I, and this page, are here.