Diplomats on ships, mid-game changes and ruleset editing

[ilkkachu]

Earlier this year, during LT57, there was discussion about the fact that the ruleset in LT57 allowed Diplomats to
do the Establish Embassy action from board of a ship. This was seen as surprising by some, and calls were
made to change the behaviour.

I meant to write about this before, but didn't get it done earlier. Even though the game is over, I think the
issues here also apply more generally, so I think it's still worth to post this. Those who weren't in the game
and didn't follow the discussions may concentrate on the points applying more generally.


- -

I must say, I was a bit surprised and disappointed about the discussion during the game.

First, that feature is plainly documented in the NEWS file for Freeciv 2.6 [1], within the list of changes
between Freeciv 2.5 and 2.6. The same feature was also present in LT54. Surely LT players have read
the documentation of the game they are playing?

Second, the meta-rules given for LT57 in the forum [2] explicitly state that minor issues that
do not break the game are not subject for mid-game changes. Changes in how a single Diplomat action works,
do not affect the general progression of the game: expanding, colonizing, building an economy, research,
and the technicalities of warfare are all unchanged. Hence, I found the suggestion of a mid-game change
for such a feature somewhat odd.

Third, there was a call from the game admin that if all players agreed, the change could still be made.
While I understand the idea of letting players choose, and changes with unanimous support should be
possible to make, a process that asks players to object is not what the meta-rules describe. Quite the
contrary. Players who want to play the game without spurious changes should not be forced to insist on
that point, it should be the default. With LT53 it was already seen that players objecting to a change of
a minor rule were branded as obstructionists, regardless of the same meta-rule nominally being in effect
in that game.

- -

Now, at this point, I expect someone to point out that the NEWS file is inaccurate. That's partly true
in the LT context. For example, the very sentence just before the one mentioning establishing embassies from
shipboard mentions that Explorers can also do the same action. This is false in Longturn games, since it
has been explicitly undone on Longturn. However, the document should serve as a very strong hint on
changes that might have happened when upstream changes were included. Any further documentation on LT
rulesets must be made by LT itself.

- -

The above said, I appreciate that noticing something is off mid-game is jarring; I've felt the same
myself. However, I don't think it's useful to call such ruleset features bugs, or errors, let alone exploits,
at least not in the cases we've seen lately. The very simple reason for this is that for something to be a
bug or an error, there must be some expected behaviour to compare against. And what is that expected
behaviour? The way LT57 was described in the forum was that the rules post contained a link to github, with
the note "Here is the ruleset".

Nowhere did I see listed what features were to be expected from the game. Nothing about Diplomat actions,
nothing about equality with earlier rulesets, or changes from them. Changes have of course been discussed
in Discord at the time they were made, but as far as I can see, they are not listed in full anywhere.
Perhaps players and editors have had some goals in mind, but they do not seem to be written down either
and it's likely that each have had their own opinions on what should be. As far as the publicly available
information tells, the game's ruleset was supposed to be what the game's ruleset happened to be.

In other words, if it is wanted for a game (be it LT57 or any) to be identical to some earlier game (LT52/LT51/
LT50 or whatever) or some "golden" ruleset, it should be brought up very clearly before the game. Partly so that
other players could know what to expect, but also so that they could work towards that end, to try to find differences
and bugs to be squashed. And, importantly, also so that it could be agreed on what will be done when (not if)
differences to the desired standard are discovered mid-game.

- -

With the benefit of hindsight, similar comments could be made for the whole LT 2.5 to 2.6 transition.

The 2.6 LT ruleset was based on the Freeciv 2.6 civ2civ3 ruleset, which had received some changes between
Freeciv 2.5 and Freeciv 2.6. In addition, new changes were being introduced to the LT ruleset. Hence, there
were no less than three sets of changes being juggled at the same time:

1) Changes from (2.5) civ2civ3 to (2.5) LT that had been there before,

2) Changes from 2.5 civ2civ3 to 2.6 civ2civ3 and

3) New LT specific changes that were being made at the same time.

The second group is a problematic one, they were the ones that someone, somewhere, outside LT had deemed useful
to make, in at least some context. But at the same time, the ones that would be most surprising for LT players
moving from 2.5 LT to 2.6 LT if and when included in the LT ruleset. However, it was not clear at least to me
during the time of transition if these changes were supposed to be included, or ignored, or some combination of
the both. In major transitions like this, clearer goals would again help.

- -

Also, a point I feel I should emphasize, about the labelling of unexpected features.

In the computing sense, the word 'exploit' is usually used to refer ways to take advantage of security vulnerabilities,
for privilege escalation, and in general, breaking in. It has a rather negative connotation. Ruleset details are somewhat
up to personal preference, meaning that there really is no absolute right or wrong. With that context in mind, throwing
around weighted words like 'exploit' should be avoided, and some effort taken to aim for a more productive way of
conversation.



[1] Freeciv 2.6 NEWS file: https://github.com/freeciv/freeciv/blob/S2_6/NEWS

[2] Forum post "The winning conditions and the ruleset for LT57": http://forum.longturn.net/viewtopic.php?id=1390

[ilkkachu]

Some suggestions which I think would make LT games more accessible to new players, and would be useful
for older players also.


- More care should be taken to document the LT ruleset.

Right now, there doesn't seem to be any comprehensive English description on the changes from civ2civ3 to LT,
The ruleset files are in github, but I don't count them as exactly human-readable documentation.

I mention changes from civ2civ3, since that's the stock ruleset that most closely matches LT in style, and
it already has a somewhat comprehensive documentation that tells how it differs from the classic rules and
Civ II. For someone familiar with Freeciv, but not LT, this would probably be a good starting point. Of
course for those not familiar with Freeciv or Civ II at all, even it wouldn't be enough.

- The current LT rulesets should be compared to the earlier 2.5 ones and to stock civ2civ3, the differences
identified and decisions made on which of the changes are desired, and which are not. Preferably by the
LT players as a group.

- Changes between LT games should be discussed and documented more clearly. Right now, there are many
changes that have not come with accompanying help text changes. This should not be so: the help texts should
be verified to match the behaviour at the same time the rules changes themselves are tested. Changes should
also be posted using e.g. the forum or some such tool, so that they don't get hidden in piles for unrelated
Discord discussions. Same goes for known bugs.

- The meta-rules of new games should be verified to match what players expect, regarding both the ruleset, and
mid-game changes. If the intent is that mid-game changes may be made by vote, the meta-rules should say so.
If the intent is that only sufficiently major problems are subject for mid-game change, the meta-rules should
say so, and it should be clear who the final arbiter is. The arbiter should be someone who is not playing in
the game or part of the teams involved.

[louis94]

I agree that rulesets should be better documented. Unfortunately, this is hard and there will always be incomplete. I think that the documentation should come in two parts: a general intro to Freeciv and game-specific details. One has to be careful that documentation tends to make people use the ruleset as intended by the person writing the docs, while creative use of the rules is often possible.

- The meta-rules of new games should be verified to match what players expect, regarding both the ruleset, and
mid-game changes. If the intent is that mid-game changes may be made by vote, the meta-rules should say so.
If the intent is that only sufficiently major problems are subject for mid-game change, the meta-rules should
say so, and it should be clear who the final arbiter is. The arbiter should be someone who is not playing in
the game or part of the teams involved.

What do you think of the rules for LT59? The expectation for this game is that the rules are fixed (and closely match Warclient, but it's not part of the rules since the Warclient ruleset is undocumented). Changing the ruleset is made possible but discouraged by a somewhat heavy procedure.

[wieder]

Here are some thoughts and this is not a comprehensive list of issues, ideas or anything.

In the end there were no changes to the ruleset mid game. The players were asked about the possibility of changes and any players could have objected in private, without letting the other players to know who that objecting players was. It may or may not be that it was already known that at least one player will object to the change. I was counted as one of the players.

The current documentation relies on the in-game help and everything else is additional. There may be bugs and issues on the in-game help and that stuff will be fixed when the bugs are found. Additional documentation has been planned and even made, so to say, but the in-game help remains as the primary documentation.

For some part it's impossible to know what would be the proper behavior. In some past games we had the behavior done differently and for some games we have changed that for some reasons. The base rulesets may give some ideas about how things should be but the documentation from Civ2Civ3 or from some other game doesn't define how the games should be implemented. If the rulesets have some weird issues, please report on those. The LTT ruleset is now frozen and this means that we no longer add features like we used to. However like it is with software, we sometimes need to fix things. Fixing may involve changing stuff and this is a process we are trying to improve. The federation science bonus is a good example of this. It was made too high for some games and this resulted with a distorted gameplay. For this case there is no original behavior and the best possible solution is to simply change the bonus to a reasonable level.

For LTT the basic idea is this.

1) the ruleset and the server code defines how the game should work
2) this behavior is documented on the in-game help
3) we fix this if there are issues
4) there are some additional pages for summaries but this may or may nit be accurate since our current process doesn't support additional documentation that well
5) we have the forum and discord for talking about it

Now for LT57 we had some players who didn't play the game since 2.5 and we had most of the pending changes done with the 2.5 transition. In addition to that we had plenty of bugs. If the players see something weird or the behavior has significantly changed, you can help by reporting it. Some players may not know how this stuff should work so it comes to using common sense and feedback from the veteran players. Not the best possible solution but that's the best we can do. All the players can of course suggest changes mid game, say that there are exploits or unfair stuff but in the end it usually comes to this -> the games are not changed mid game and we will try to fix things for the future games.

We have some examples of what constitutes as a major issue and if it becomes uncertain if this is the case, we need some way to deal with that. Since this is a game and it's for fun, it was suggested that the change would be made if no one objected. And objecting could have been made in private. It would be too rigid to my taste if all the players together wouldn't be allowed to make a change. Now this probably should be added to the rules.

We have no golden ruleset.

LTT is based on previous rulesets from 2.5 and versions before. On top of that we have experience from the game and all this was added to LTT when it was created for 2.6. Even if we wanted to, it would be impossible to have exactly same behavior from the past games. Some new things are now possible and some things are even impossible to implement. Like the single hut in the game. At least impossible with only ruleset changes.

Some players have wished for less or no changes while also reducing the tc exploits. At least no changes with less tc exploits would be impossible to have. I doubt the players would want to have a game that would never change even if there would be a single golden path and lots of features allowing the players to gain not fun advantage over other players. We of course have LTX for extra stuff that can be put there. However if we would completely end all the fixes on LTT we might end having LTT as a relic and LTX as the alternative too many people would prefer over LTT since we can do more changes there. This of course doesn't mean that LTT should get all the possible features we can figure out.

The problem is that for a ruleset a bug may not produce error or something that would be instantly noticeable. A bug may make some strategies too weak or too powerful and often only the real games and time will tell if this is the case.