Longturn input for freeciv-3.0

Planning and discussing Freeciv Longturn gaming
cazfi
Elite
Posts: 1345
Joined: Tue Jan 29, 2013 6:54 pm

Longturn input for freeciv-3.0

Postby cazfi » Sun Sep 03, 2017 5:52 pm

In the past we've had problems that longturn(.org) people have not participated in testing of upcoming major freeciv releases at all to find shortcomings in the longturn type of gaming before datafile formats have been frozen and it's too late to do anything about it until next major release. Only after there has been a release they have started asking why there's no this or that feature, or how the new features should behave differently. I'm worried that it might have happened again with freeciv-2.6, but as that d3f went 1.5 years ago already, let's concentrate on freeciv-3.0 case. Its datafile formats are already stabilizing so that only well justified changes (completing existing features, ones agreed previously already, etc) may go in any more. I hope that someone from longturn community makes some testing with current freeciv-3.0 engine to find out what needs fixing from the point of view of longturners.

User avatar
Corbeau
Elite
Posts: 622
Joined: Mon Jan 13, 2014 11:13 pm

Re: Longturn input for freeciv-3.0

Postby Corbeau » Sun Sep 03, 2017 6:10 pm

Well, to have more people testing it, it would be good to have the thing available. I would gladly test it, but I have no clue how to have it working on my computer.

wieder
Hardened
Posts: 165
Joined: Thu Jul 27, 2017 10:43 am

Re: Longturn input for freeciv-3.0

Postby wieder » Sun Sep 03, 2017 9:42 pm

Has there been some recent issues with longturn support? I remember having discussion about setting the city vision smaller than the actual city work area, but it's really not that big issue. We are using smaller vision for cities to encurage people to actually put some units inside the city. Now it was found out that when this happens, there may be some conflicts with enemy units blocking access to tiles that are not really seen but which the city still is able to use.

It's nice if the cities can keep working on those tiles, but it's certainly nothing too serious is that ability is removed. I don't see any real problems emerging from removing something like that.

The biggest problems we have on longturn.org games are issues with real time actions happening arount the turn changes. Stuff like moving workers to make a road just before the tc and then using that road to attack right after the tc. In a situation like that the enemy should be online when the tc happens and also try to do some fast paced defensive moves in order to prevent the road from happening. The problem is that no one has really figured out a good way for preventing the various exploits one can take during the tc.

For roads there has been an idea for some kind of setting for workers not doing work if it's started too close to tc. This is of course just one suggestion for specific tc exploit and doesn't really solve everything. The idea was that if a worker can see enemy unit n seconds before the tc, the work would never be completed in just one turn. I have no idea how complicated it would be to actually code that kind of feature and no one never really tried to design the idea as a proper suggestion to fix that kind of tc stuff.

One solution might be changing the unitwaittime to affect all the units so that the attacking units couldn't be moved in right after the tc but even that has some serious downsides and I might not want to include a setting like that on a multiplayer game.

Other tc exploits include the last minute tech target changes, switching production to some wonder (with tons of caravans:) and maybe also the classical attack where the troops are moved seconds before tc. There is unitwaittime, but also the problem that the enemy has no time to move defensive units before the tc or maybe switch the production to something that would help defending the cities/units vulnerale to those units. I have no idea how to solve issues like that.

One great problem we could have suggested fixing are the trade routes. We have disabled the trade routes for longturn.org games because the huge map sizes easily create an issue with too big one-time bonus. With a map size of 200*200 tiles you can't really set the route values to give a reasobably good bonus and routes for close enough cities while also not giving too big bonus and routes for situations where the caravans are moved on the other side of the map. Instead of even trying to suggest something we have simply disabled trade routes. This has been working reasonably well and as a compensation we have been using wonders, terrain bonus alterations and changing the costs and output values for city improvements.

Now there is also tech trading :D

With tech trading the "problem" happens with players getting free techs without the need for using resources to actually produce that stuff. There is of course tech upkeep, but it's something that can be exploited if you have allies who are willing to give you all the techs once they are lost at the tc. There have been suggestions how to fix those but at the moment it semms no fool proof method is yet to be suggested. At least not one without too big drawbacks. Instead of tech trading we have been using tech leakage and experimenting with some tech stealing options.

I wouldn't say we are complaining about the lack of some features or settings. It's more like we are asking what can be done and what options there are to cope with the multiplayer related issues. Freeciv is a great game and I'm personally happy with the possibilities there are to cope with some exploits.

One of the latest changes was to remove partisans inspired from cities lost to the enemy. The replacement partisan unit is yet to be tested in a real game but it looks like there is no problem in making that kind of changes with normal ruleset features.

I would be happy to talk more about the subject. For longturn.org games most of the issues, except the tc ones, have been dealed with editing the ruleset. The rulesets are really flexible and allow these adjustments we have seen necessary. No complain there, really. Instead, thanks for making all this possible!

Wow... This was a long one :)

cazfi
Elite
Posts: 1345
Joined: Tue Jan 29, 2013 6:54 pm

Re: Longturn input for freeciv-3.0

Postby cazfi » Sun Sep 03, 2017 10:50 pm

wieder wrote:One great problem we could have suggested fixing are the trade routes. We have disabled the trade routes for longturn.org games because the huge map sizes easily create an issue with too big one-time bonus. With a map size of 200*200 tiles you can't really set the route values to give a reasobably good bonus and routes for close enough cities while also not giving too big bonus and routes for situations where the caravans are moved on the other side of the map. Instead of even trying to suggest something we have simply disabled trade routes. This has been working reasonably well and as a compensation we have been using wonders, terrain bonus alterations and changing the costs and output values for city improvements.
Trade routes have gained small number of changes in 2.6, and a bit bigger changes in 3.0 with the introduction of Goods. For the specific issues of map size, there's now server setting tradeworldrelpct

Code: Select all

> explain tradeworldrelpct
Option: tradeworldrelpct  -  How largely trade distance is relative to world size
Description:
  When determining trade between cities, the distance factor can be
  partly or fully relative to world size. This setting determines how
  big percentage of the bonus calculation is relative to world size, and
  how much only absolute distance matters.
Status: changeable
Value: 50, Minimum: 0, Default: 50, Maximum: 100


It would be good if you tested this and reported if it needs some tweaking.

wieder wrote:I wouldn't say we are complaining about the lack of some features or settings. It's more like we are asking what can be done and what options there are to cope with the multiplayer related issues. Freeciv is a great game and I'm personally happy with the possibilities there are to cope with some exploits.
I think the case in the past was that change in freeciv engine was considered an regression in some longturn use-case. We would want to avoid the situation where the game turns worse for you.

As for 3.0 I'm a bit worried about the generalized actions related popups (unit usually does not automatically attack when you order it to move toward enemy unit, but it popups a menu of actions unit may do) They may not be that well suited for competitive multiplayer game (they have been ok in my own relatively friendly games)

sveinung
Elite
Posts: 421
Joined: Wed Feb 20, 2013 4:50 pm

Re: Longturn input for freeciv-3.0

Postby sveinung » Mon Sep 04, 2017 10:12 am

cazfi wrote:As for 3.0 I'm a bit worried about the generalized actions related popups (unit usually does not automatically attack when you order it to move toward enemy unit, but it popups a menu of actions unit may do) They may not be that well suited for competitive multiplayer game (they have been ok in my own relatively friendly games)


Do you think we should set popup_attack_actions to false by default?

cazfi
Elite
Posts: 1345
Joined: Tue Jan 29, 2013 6:54 pm

Re: Longturn input for freeciv-3.0

Postby cazfi » Mon Sep 04, 2017 3:29 pm

sveinung wrote:Do you think we should set popup_attack_actions to false by default?

Don't know. Maybe the kind of players who want it to be false are the people who will more likely find the setting, while those who benefit from it being true are more likely to play with the default value?

User avatar
AndreasR
Elite
Posts: 686
Joined: Thu May 02, 2013 10:26 pm
Location: Norway

Re: Longturn input for freeciv-3.0

Postby AndreasR » Mon Sep 04, 2017 4:18 pm

sveinung wrote:Do you think we should set popup_attack_actions to false by default?


Suggestion: The attack dialog could have three buttons: "Attack", "Attack immediately from now on", and "Cancel".
(Perhaps use different words)

wieder
Hardened
Posts: 165
Joined: Thu Jul 27, 2017 10:43 am

Re: Longturn input for freeciv-3.0

Postby wieder » Tue Sep 05, 2017 11:43 am

That stuff I listed on that lengthy message, how are they dealt with in Freeciv-web multiplayer games?

I'm guessing that at least the tc (turn change) actions also affect those games. Would be really interesting to hear about that from the people playing the games.

User avatar
AndreasR
Elite
Posts: 686
Joined: Thu May 02, 2013 10:26 pm
Location: Norway

Re: Longturn input for freeciv-3.0

Postby AndreasR » Mon Sep 18, 2017 5:01 pm

sveinung wrote:Do you think we should set popup_attack_actions to false by default?


Freeciv-web now has a "Auto attack from now on" button. Added here.

wieder wrote:That stuff I listed on that lengthy message, how are they dealt with in Freeciv-web multiplayer games?

(The biggest problems we have on longturn.org games are issues with real time actions happening arount the turn changes.)


Freeciv-web LongTurn games set set unitwaittime=36000. Full settings here.

In particular, this are the rules for allowing players to join the game after it has started: 10 gold per turn, up to max 700, and every tech known by at least other 40 players.

wieder
Hardened
Posts: 165
Joined: Thu Jul 27, 2017 10:43 am

Re: Longturn input for freeciv-3.0

Postby wieder » Tue Sep 19, 2017 11:23 am

Longturn.org games have also been using the 10h unitwaittime for ages now. The problem with double turning is however only partially solved by that. The best example is building a road 10sec before TC and then using that road 10sec after the tc. It's possible if the attacking units didn't move and usually they can wait for the moment of action if all is planned well.

700 gold for compensating late joining is not really that much while it's much. It equals 350 shields and the top 10-20% of the players are usually able to have a production like that by T70 on longturn.org games. Then again the settings are very different and the production is not comparable really. Would be interesting to see the productions / economics for a web game. There are less start units but the multiplayer ruleset allows really fast growth with rapture.

We have been using something like 4 settlers, 5 workers and one explorer as the start units. No military units. The web multiplayer uses 2 settlers, 3 warriors and an explorer. No military units. There are more start units because we try to keep the players more interested during the early game when there is less stuff to be moved and less actions to be taken. Too little to do may result with people getting bored and leaving the game. 6 settlers on the other hand resulted with some new type of strategies we didn't want to really embrace. However. Never say never :D