Default+Dunefolk era balance

[Celtic_Minstrel]

I think the problem with liminal would be greatly alleviated if it were inverted, giving a bonus during dusk and dawn and no modifier during night and day. This is far easier said than done, though. The most obvious idea I can think of (showing it as a bonus when it's not active and as if it's not active when there's a penalty) will run into trouble as soon as you hit a scenario with a modifier other than 25%. Maybe some slightly more complicated version of this idea could still work, though. The other option would be adding a new key to `[time]` to directly specify the liminal bonus, but that runs into other problems (such as creating the ability to make liminal units not actually be liminal).

[holius]

It's possible to reimplement liminal as a bonus. I have spent some time playing with the code to see which one of my previous suggestions was the easiest to code. It seems Celtic_Minstrel's one is even easier, so I have made a PR to change liminal to a pure bonus (comparable in power to fearless chaotic or fearless lawful).

[Elder2]

Even if you changed liminal bonus values and then tweaked the unit stats it would change balance in some rare cases because it is impossible to change the value of the bonus and units stats and get the same damage values.

[skeptical_troll]

I'm confused, what does the base value refer to now? Herbalist is 6-2, but ad day/night it used to have 5-2, is that changed?

also, can you briefly explain what happens when, for example, lawful bonus is -30%?

[holius]

The PR indeed contains slight balance changes.
Where damage was 5/7/5, I had to chose between 5/6/5 and 6/7/6. Herbalist melee, Rider ranged, Swiftrider melee are impacted.
Where damage was 6/8/6, I had to chose between 7/9/7 and 6/7/6. Explorer melee+range, Windrider melee are impacted.

4/5/4, 5/6/5, 7/9/7, 8/10/8 stay the same, but resistance tweaks may also impact them. For this reason, I tended to chose the higher value for Herbalist and Rider line, to compensate the global slight downgrade when facing a faction with resistances.

Each scenario has a max_liminal_bonus, which will be granted as bonus when lawful_bonus is 0.
When max_liminal_bonus is 25% and lawful_bonus is -30%, the effect on liminal is 0.
When max_liminal_bonus is 25% and lawful_bonus is 10%, the effect on liminal is 15%.
max_liminal_bonus computation algorithm may be discussed. I think my attempt make sense: match the best of effect from either fearless chaotic or fearless lawful, with the restriction that the max_liminal_bonus must be an existing absolute lawful_bonus. For weird schedules, or highly map dependent time of day effect, opens the possibility to set max_liminal_bonus in WML (at scenario level).

[Ravana]

Did you consider fearless liminal?

[holius]

I understood Celtic_Minstrel's suggestion to implement liminal as 0/+25%/0 as an implicit request to tie it with a kind of "always fearless" (no negative effect) hidden trait. But your question makes me realize I could have understood differently.

I don't have much experience of alternative schedules, so I am open to any suggestion on what should be the liminal bonus on those.

[skeptical_troll]

Thanks for the explanation. As Mathbrush pointed out, if you want to recover the -25%, you need a +33.3% from the new base damage, this should in principle avoid the problems above. I haven't checked what happens with the rounding though. Another thing to consider is that units with 'strong' (50% of recruits) will be stronger than before at dawn/dusk, like a strong Herbalist will now deal a 9-2, which may turn useful in some match-up.

I have mixed feelings about how much this helps, a frequent criticism of 'liminal' seems to be that it base damage is not the average over the cycle, and this won't change regardless of where we set the 0%. The other disadvantage, although limited to a few UMC, is that its behaviour sounds less transparent on custom time areas, whatever we decide about it.

Rant: an inelegant but safe solution would be to have the liminal_bonus as a separate key, hence an independent variable with respect to the lawful_bonus. Then if people really feel bad about the 'non-average-base-damage' problem, one can invent a new cycle for the liminal units which, starting at dawn, is as follows: bonus, base , malus, bonus, base , malus. So max at dawn at dusk, base at morning and first_watch, minimum at afternoon and second_watch. This would make sense thinking about the evolution of the temperature in the desert during the day. It would keep the novelty of the different cycle while having base damage=average damage. This however sounds overtly complicated and I still think that the current situation is preferable over all the presented alternatives.

[holius]

Thanks for the explanation. As Mathbrush pointed out, if you want to recover the -25%, you need a +33.3% from the new base damage, this should in principle avoid the problems above. I haven't checked what happens with the rounding though. Another thing to consider is that units with 'strong' (50% of recruits) will be stronger than before at dawn/dusk, like a strong Herbalist will now deal a 9-2, which may turn useful in some match-up.

I have mixed feelings about how much this helps, a frequent criticism of 'liminal' seems to be that it base damage is not the average over the cycle, and this won't change regardless of where we set the 0%. The other disadvantage, although limited to a few UMC, is that its behaviour sounds less transparent on custom time areas, whatever we decide about it.

In fact, I am ready to not recover the -25%, because trying to get this +4/3 would break KISS principle even more than now.

Liminal base damage is not the average over the cycle (in standard schedule).
Always_fearless base damage is not the average over the cycle (in standard schedule).
In non standard schedule, like TWO_SUNS_DEFAULT_SCHEDULE, neither chaotic base damage nor lawful base damage is the average over the cycle.

Yet only Liminal is complained about ?
With my changes, the Dune Rover 4-3 melee displayed has the same average over standard cycle as the Ghoul 4-3 melee.

Rant: an inelegant but safe solution would be to have the liminal_bonus as a separate key, hence an independent variable with respect to the lawful_bonus. Then if people really feel bad about the 'non-average-base-damage' problem, one can invent a new cycle for the liminal units which, starting at dawn, is as follows: bonus, base , malus, bonus, base , malus. So max at dawn at dusk, base at morning and first_watch, minimum at afternoon and second_watch. This would make sense thinking about the evolution of the temperature in the desert during the day. It would keep the novelty of the different cycle while having base damage=average damage. This however sounds overtly complicated and I still think that the current situation is preferable over all the presented alternatives.

That's more or less how the first implementation of liminal failed. It broke Illumination interaction, required modifications of scenarios (mainline maps and UMC) that weren't done.

[skeptical_troll]

Liminal base damage is not the average over the cycle (in standard schedule).
Always_fearless base damage is not the average over the cycle (in standard schedule).
In non standard schedule, like TWO_SUNS_DEFAULT_SCHEDULE, neither chaotic base damage nor lawful base damage is the average over the cycle.

Yet only Liminal is complained about ?

I totally agree, although for me it's also not a problem having always a malus and never a bonus, which I actually find cleaner.
Coming to your implementation, if you want to preserve the 25% difference between day and dusk, then the only way I can see to give a sensible general formula is
liminal_bonus = 0.25 - |lawful_bonus|,
which means basically extending what you are already doing. So deep underground would be -5%, etc.

That's more or less how the first implementation of liminal failed. It broke Illumination interaction, required modifications of scenarios (mainline maps and UMC) that weren't done.

right, didn't think about illumination, forget what I said.

[Elder2]

Bear in mind that khalifate units are not the only units in Wesnoth with liminal. Demons from loti also come to my mind and liminal change weould interfere with them as well, and, given that the damage values in loti are much higher than average and considering the intricate resistances system - any possible change would affect them more than normal units.

[name]

The six turn day was never created with liminal in mind.

The only transparent way to give it a similar average to lawful, neutral and chaotic, is to jump up to +50% during the two separated twilight turns and then straight down to -25% during the remaining four. That is a very dramatic shift with no middle ground. I guess it would make units with this alignment similar to those with the charge ability.

But simply moving the two liminal units over to neutral or chaotic alignment (or even back to lawful) would be the best option.

[Elder2]

I see liminal as just a gimmick and the way it is affected by weird tod can be both interesting and create new possibilities or potentially break units (eg in loti demons can be severely weakened by using items like light devouer which I believe gives +40% chaotic bonus during neutral, +30% during lawful (normal lawful and chaotic +25%) and +50% during chaotic), it is probably one of the reasons why I don't remember seeing any faction outside khalifate with liminal units. I doubt liminal can be "fixed" with regards to custom tods.

Taking this into consideration I guess you could argue that removing it from khalifate and changing the alignment of liminal units could be a good option, but really, what is the point? I believe you can not tweak liminal so it will look "nicer", and khalifate itself is sort of a gimmick, its gameplay is rather special in some ways, I am not sure if you could ever balance khalifate.
Eg mass troll is an undefeatable cheese against khalifate if performed well (though I don't remember seeing many replays with trolls vs khalifate) but another interesting thing is dwarf vs khalifate. I played one dwarf dwarf game vs khalifate and I completely destroyed it when khalifate attacked me because the thing is - guards are cheese vs khalifate. That is right, we have lived to see a time when guards camp is a cheesy strat vs a faction. Khalifate has too little damage and its units are too expensive to have a way of dealing with guards, if you support them with any dwarf unit, thunderers or fighters they can never be killed, khalifate has no poison, no magic, nothing to kill guards. Of course I would like to play more games to confirm it, since CP in our game did not want to believe my prophecy of cheese. xD

[Xalzar]

khalifate has no poison, no magic, nothing to kill guards

Dunefolk*
About this, maybe adding a poison attack to the Herbalist could add some utility to that unit and solve this problem. A healer/poisoner units could be quite unique in mainline, and we can even decide to branch its level-ups so that one specializes more on healing and the other on poisons. Just an idea!

About liminal, I can't comprehend lore-wise how they could possibly be adapted to fight better in crepuscolar light. In nature strictly crepuscolar animals are super rare, and some who are considered as such are in reality nocturnal with the peak of activity in vespertine hours, so...
I know that some scorpions and snakes hunt in twilight because they cannot withstand the temperatures of the desert in the day and in the night, but humans? With clothes and fire? And with eyes adapted to diurnal light? And what about when they fight in non-desert environments? Wouldn't they be better suited to fight in daytime in that case, or in nighttime if they're particularly "rogueish"?

And gameplay-wise, the liminal alignment seems to create more troubles than novel tactical options. If the main advantage of using liminal units is to prolong the turns in which the faction can attack at its best (which could be debatable, since the roles of units are quite specific and diverse) and reduce the vulnerable turns, then let's have neutral units instead. At least the faction would not have the problem of having units unable to support adequately its diurnal attacks and alleviate the malus of the night.

I think changing liminal to neutral could make sense:
-gameplay-wise: a faction with core units split between lawful and neutral isn't present in mainline (the Rebel faction has lawful specialist units), and can press more prolonged assaults like it can partially do with liminal units;
-lore-wise: I can imagine that such a civilization, or at least its less stantiated individuals, are well syntonized to the desert and are able to make the most of their environment all times of the day - even tactically, striking whenever the enemy is at its weakest while they are adapted to fight in all light (and temperature) conditions.

Another option is to convert liminal units to chaotic, maybe with the addition of the fearless trait: if the desert is really too hot in the day, they will surely be active at night and not only in the short twilight hours. They are not lesser animals, they have the means to use all the night if they need.

The last option would be: let's decide if they're lawful or chaotic or neutral, and give them a trait which gives them +25% at dawn and/or dusk (depending on if we want the perfect balance or we go the "fearless trait" way).

[Elder2]

@Cold Steel Also I am not just talking about the 6 turn day, you sound like you think it is only about 6 turn day. Even things like illumination from lava on some maps can create custom tods and mess with dunefolk.

Anyway I think I have made clear what I think about it. I am pretty certain creating anything resembling liminal would run into the same "issues" liminal has and I myself quite like liminal as a gimmick.