Defending by running away megathread

[Suriname]

Problem statement
MoM/CoM are harsh. When you make a choice you face the consequences. There is an in-game trick that allows to make a choice and not face the consequence. It cheapens the game, and it reduces its value of a hard and difficult one. It also forces challenge-minded players to use it, because it's just that much of a boon. It's abuse without any advantage to the game and it should go.
Do you disagree? Then please tell me: what is the advantage that having this trick brings to the game? How is it different from summoning wraiths on turn 1?

but I'm strongly against excluding flying creatures from the "conquest" trigger. It's unfair.  If I'm attacking a city and my creatures are flying, why should I be penalized for that and forced to engage the enemy when I can just move into the city tiles and conquer it. What, the enemy can't attack me? That's their problem, they should have researched web. If my cavalry can do it, my death knights or doom drakes should be able to as well. Again, this is subjective, for me this would be a greater deal breaker than the city not becoming mine even though I enter it. At least the current rules are consistent : You don't get the city until you kill all enemies. New rule would be : You don't get the city until you kill all enemies if your units are flying, otherwise you do. Well, that's just bullshit. But if we don't exclude flying, then I can abuse that to steal cities. Which is even worse.

This is a perfect example of not only "do not fix if it isn't broken" but also "do not fix if fix is not an improvement" at the same time.

Except, you've already received many reports of the fact that flying is absolutely broken. Reducing its advantage is therefore a desirable feature of any proposed solution. Solution implies a problem, and I can show that there is a problem, as the fact that you've avoided to answer any of these points makes me suspect that you don't have an answer and this makes you uncomfortable:
- people can see vids on the web of shitty run away to keep a city tactics, making your work look bad
- tactics become tricks making the game simpler, see the answer to C.Mike. I hope we all agree that the richness of the game, its complexity, is what attracts us to it. Clearly a tactic that is always used presents therefore a problem.
- freedom doesn't mean lack of consequences. One can choose to abandon a city before the battle if the enemy stack is seen coming, or during the battle with "flee". But, in tactical one can run around and not engage. That's so useful that I feel obliged to plan for it - not as the main strategy, but rather as an additional parachute that gives me more leeway when facing an attack. At the beginning this just needs 40 prod - the cavalry - or some cheap summon, nagas and bears are quite ok for this against slower stuff as they are tanky too. But this is lack of consequences from choices. I choose not to engage in my city and I keep it despite that.
- if it were a multiplayer game nobody would accept this, let's be honest. To take any cities you'd have to bring cavalries only to avoid this trick, and enough to withstand the initial barrage of spells/ammo, this would self-reinforce in a loop until some realms wouldn't ever be used, it'd soon become Caster of Life buffs on cavalries and all games would be decided by turn 50. You'd receive soon enough enough reports that you'd be forced to change it.

will be forced to flee

Never been a fan of games where I don't get to pick my action and the game forces me into one, especially if that one is not favorable for me.

Sure thing. You are free: pick your action. Have you decided that you want the city? Then don't flee, remain and fight.

What you have right now is have your cake and eat it too. But you'll agree that that's not MoM (or, hopefully, CoM) - the game is effing harsh, if you take a decision you have to live by it and confront its consequences. Or would you like to go back to the days of summoning wraiths at the beginning too?

One thing to note here : we fixed the production interruption bug. That doesn't sound very related but it actually is. I bet about 40-60% of the time, when a player manages to get to keep a city through running in circles, this bug helped them do so - so the 75% damage destroyed the armorer's guild? No worries, the paladin you bought last turn is already completed sir, we are ready to defend now! This will be gone now. You lost a military building? No new unit to defend you next turn. You either can instantly summon a strong creature there, or you can forget about keeping the city.
So my claim is the "problem" might be already fixed, after this change, the tactic will be even less effective.

In what version? Because I've just had it happen. Anyway, it's not enough. If I  choose not to defend a city, I want to lose it - either it gets razed, or it changes hands. And if I refuse to engage, well, then I've made my choice...

The advantage
Clearly the above is true only if there's an advantage obtained from this trick. Advantages:
1. defensive
In the game one can attack or defend. By defending you get the first turn casting, and the advantage is obvious. So, by keeping the city you force the enemy stack to attack again and can further reduce the AI doomstack: next round those 2 confuse'd units won't be there, you'll manage another confusion or 2, and even if you eventually lose the city the doomstack's been reduced to nothing. When you manage to suppress 2 AI summons you have suppressed not only the units themselves, but all the time/effort taken in moving them from the summoning circle to your city. The advantage, at turn 60, is great. Currently it's the only thing that makes lunatic possible, because you can do this with a cheap unit and overcome the difference in resources.
2. reinforcements
If there's reinforces on their way - and with all the cities at distance 4, this is perfectly possible - a single turn makes or brakes keeping a city.
3. lost buildings
Sure, you lose some buildings... But do you? If you get lucky you won't, especially if the enemy has run after you till turn 10, then gone back and entered the city around turn 13, and not many enemy units remain. In any case, even when you lose some buildings and population the destruction will be much less than the total in the city. Confront this with losing the entire city, including the time and resources of the settlers, and you see how the current loss does not compare.

This works so well that eventually I started to use it on attack. With cheap tanky summons - bears, nagas - or with cavalries if the enemy is faster it becomes cheaper to go into an enemy stack one by one, cast save or die spells - confusion, sleep etc - and run around the slower enemy until turn 25. But this is for another thread....

Just don't do it
That's dumb, why wouldn't I do something that the game allows me to? But I don't like it, I think it makes the game look ridiculous when it's seen on videos, and that it cheapens the whole experience for everybody involved.

The solution
1. the algorithm
After all the back and forth in the other thread this algorithm seems to be easy enough to implement and the closest to a satisfying solution:
Counter for both sides, starting on turn K to see which side keeps the city. K can be subject to balancing considerations, offering a way to fine tune this approach. I'd use 10 as a start, the same turn in which the AI units run for the city, because it feels the right number, but it's just a first idea. It could be from turn 1 to keep the advantage to the defender, for example.
At turn 25 the side with the highest counter "wins".

2. Definition of win
For me the win means forcing anyone who's refused to engage to face the consequences of the choice. If it is the defender then the city should change hands, or be completely razed - I don't really mind either way. Any remaining troops should then either flee - because that's what running around is - or "withdraw exhausted", but the latter might be too complicated as it's currently possible only on the attacker side. It would however be better in that it compensates the player the effort of running around and keeping the unit alive... This brings me to,

3. Extending this to flee
If you look at the flee function, it's a simplistic 50% survive/die for each unit.
We can extend this approach to the flee function and have speed and movement mode (flight, etc) be considered for the odds of fleeing, replacing the simplistic dice roll.
At this point, we could replace all "retreating exhausted" with simply the flee retreat - because it happens mostly in cases of faster or equal movement speeds. Therefore, the cavalry wouldn't be lost by fleeing ever, or most of the time, resulting in a similar outcome as the exhausted withdrawal. This however is more complicated and probably deserves its own thread.

[Suriname]

I've gathered some comments on this. I think they represent most opinions, if I've forgotten anything then just post.

I don't get this "cavalry defends by running". Could someone point to a video (and time in said video if/when hours long) that would showcase it? Sure, I get you can avoid getting the city captured, but you should lose most buildings, so what's the point of it as this "game breaking trick"?

That said, perhaps dropping the move of cavalry to 4 would be a simple help for the worst of this? That would make it unrealiable to keep avoiding more than 1 move 3 creature (like naga, hell hounds, war bears)

It's not only cavalry that can do this, against anything that moves 2 you can do it quite simply with movement 3. Cavalry is just the best example.

You can find it in zitro's movies I think, it's used quite regularly at the beginning of the game, turns 40 to 70, but it can happen later on as well.

[Suriname]

Realism

Thinking about it, it shouldn't be that unrealistic (for what stands for realism in a magic game): If the defender can avoid the attacker, they can do guerilla warfare of all sorts, preventing the enemy from gaining control. But the attacker could make a ruined heap of the area, if so inclined. And that's precisely what does happen.

Furthermore, the tactic seems very precarious as a lot of things can kill your cavalry: direct damage spells, ranged attacks, web.

On the realism: is it realistic to have your cake and eat it? That's what this trick is. More seriously: in game, battles are represented with exchange of hits, cavalry's guerrilla warfare is already represented with first strike. Fast skirmishers are represented by ranged combat ability - you should explain to me why horse archers can do this exactly as well as cavalry, by your definition of realism they should be much better. This trick takes it further and is therefore not for realism, but merely a game design limitation.

On the precariousness: nah, the AIs target ranged attackers first. Keep some bowmen to absorb the first few spells and you're good to go running, and if you're far enough by turn 10 even web won't matter. This happens and matters less after the 50 skill point is reached, but if you're still around at 100 skill then nothing actually matters anymore: you have already won and you just want to see the score.

[Suriname]

It's fun

Once you can beat the game like this, where's the challenge? What's the difference with sim farm or some my little pony game? How does repeating a learned trick even make you feel smart, as you say? In short... Keeping these tricks in the game, when they could be very simply fixed, limits the game, reduces its re-playability, and makes it less alluring to newcomers when people see videos on the web where they see these tactics and understand that it's "just all tricks".

Imho it is all BS. I play games to have fun. To discover these tricks and execute them, to deal with RNG gods and dodging stupid AI. To waste time building imaginary empire. And surprise -- I have fun... About replayability: these games are with us since 90s -- if this is not replayability then what is?

Executing the same trick alll the time is monkey job, there's already work for that, what I like is having rich tactics and strategies and engaging with always different ways the battle. I do play online some stuff, but nothing fantasy related despite having passed through many titles. Nothing beats MoM in terms of potential, but having a few options that are overpowering the others is a problem, and I'd like to fix it.

Besides, you don't believe what you say either. Because if you did, then you'd be playing MoM with the fixes, and summoning a wraith to beat the game. Why don't you do it? The cavalry trick is a bit less easy but it's still a trick, and if not right now eventually you're going to tire of it and abandon the mod as well.

[Suriname]

Feasibility

On higher difficulty you can not use running with cavalry, maybe just for neutrals. On Master/Lunatic AI has de-facto infinite casting skill and mana and extra pick further increases the chance of AI having at least one damaging or immobilising spell.

I'm using it right now in my lunatic mono sorcery game, it's possible in a lot of cases. Keep in mind that the AI shots ranged units first, and infinite skill happens later than this trick is an issue - but even then the same can happen with a more tanky unit, you can do this with say an archangel, a nightmare, a hero... But at the beginning it changes the outcome: you can plan for it by having 1 cavalry and archers rather than swordsmen to give to yourself the security that you'll hold the city that 1 turn necessary to bring the real army in.

[Suriname]

And a post dedicated to Seravy cos he's the boss

I don't think proving "the problem exists" is possible on something that's entirely subjective and up to the players' personal preferences.

If I have a belief, and I can prove my assumption with reasoning starting from bases we both agree on, then I have proven my position to you. If the reasoning shows a conflict between some of your positions then you are either choosing one or the other, or you are assuming an irrational position, but at least I've made you aware of it. Let's not go too philosophical.... I'll try to structure the discussion in as streamlined a way as possible so that we can actually engage constructively despite the difference in positions. You'll see that everywhere my starting point is the richness of MoM, what I consider to be the attraction that still keeps people around.

As for your "simple" fixes, they weren't simple. At least not on the side of game code, but I believe also as rules. We haven't really got there to discuss that in detail -as I'm against the whole idea anyway

Well, thread closing tends to prevent discussion. But that's not exactly productive, it's the receipt for having an echo chamber. Is that what you want? If that's the case I'll just leave, no need to waste further time for either of us. Is losing a contrarian such as me going to make you feel better? Yep. Sure thing. Is it going to enrich your mod? You tell me, but if you want to form an informed opinion I suggest reading this: https://www.idc.ac.il/en/schools/governm...f-1973.pdf and when you reach the line "leader’s influential position" think about your power on the mod and this community.

[Suriname]

Flying, invisibility, and other ways to not engage different than retreating with speed

Simple enough to be implemented without trouble:
Any unit flying, invisible?, <?> won't be counted despite being inside the walls 4x4 area.

Reasoning:
if it's flying, or in any way not engaging, then it's not actively occupying the city in the in-game representation. But apart from the interpretation... It goes contrary to the committal thing, and it is again having the cake and eating it too. So, it shouldn't count for holding a city, nor should it for conquering it.

If both sides have flying units then it's easy, they should fight it out.

There is one side case, extreme enough that I don't think it's a big deal: a fast defender against a flying only slow attacker would still be allowed to go with this trick. It's basically only an issue against draconians, but these are powerful on their own and this would be a very minor nerf to them, so I don't think it's a big deal. Keep in mind that in this case the slow flyers are still damaging the city. Despite this, I would love to solve this side case too, and I can think of ways but not with complicating the algorithm excessively.

* on invisibility I'm not 100% sure. If it's inside the walls then it can probably be seen quite easily, all the AI needs is an adjacent unit. It might be worth it testing the solution without invisibility first.

Are there any other ways to stand around without engaging?

I'm strongly against excluding flying creatures from the "conquest" trigger. It's unfair.  If I'm attacking a city and my creatures are flying, why should I be penalized for that and forced to engage the enemy when I can just move into the city tiles and conquer it. What, the enemy can't attack me? That's their problem, they should have researched web. If my cavalry can do it, my death knights or doom drakes should be able to as well. Again, this is subjective, for me this would be a greater deal breaker than the city not becoming mine even though I enter it. At least the current rules are consistent : You don't get the city until you kill all enemies. New rule would be : You don't get the city until you kill all enemies if your units are flying, otherwise you do. Well, that's just bullshit. But if we don't exclude flying, then I can abuse that to steal cities. Which is even worse.

No this doesn't work. You cannot conquer it without effort, because the enemy knows that they need to actively defend it and can't just run in circles, so flyers need to engage just like land troops. In the case that the enemy decides to flee, and is faster than you, then there is a small reduction in the huge advantage that flying currently has - but this is my opinion. Even if this opinion is not shared, it doesn't matter: the case is very minor as it only concerns the cities the player has decided to use the trick and the AI has brought slow flying attackers only, how often does happen...

If you're worried about the AI not killing enough stuff by turn 10 in tactical, then raise from 10 to 20, with this trick it won't matter so much. Or even better, make a check and instead of a fixed turn, stop pursuing and go for the city as soon as all enemies are faster than you, from round 1 if needed. This will make tactical AI a bit less dumb. It might be too difficult; in this case, ignore this, a situation where only slow flying AI stacks can be tricked like this is already much better than the current situation. Fairness to AI doesn't matter....

[Suriname]

And, finally....
Yeah.... It's still not a problem
Well, then you are not convinced. But let me you ask the reverse, if you aren't, and please answer just for the effort I've put in this. If it isn't a problem, is it a feature? What does it add to the game? Why do you like it? Would you still like it if the AI did it?

... Can you convince me that you're not just after the cheap satisfaction of beating a game that is famous for its difficulty using a trick that has been by now clearly outed as easy?

[Suriname]

Just a last thought I had while eating: this change brings tactical closer to strategic (albeit I'd like the other way around) by making it more difficult to have combats without consequence. As far as I know strategic always ends in a defeat for one of the sides, and never in a retreat, right?

[zitro1987]

The problem is that you bring the same kind of proposal with conquest trigger, which combines difficult code with opening up new abuses that would be prevented with more complex rules and code.

Let's go back to the feature and see if there is something that can reduce the annoyance of the escaping defender or reduce the reward of it.

Idea #1 (radical-esque because of a feature change) -
*We set some reasonable algorithm that makes it obvious you're escaping and your town got trashed, with almost no cases where that is not the case (a basic one may be troops outside city walls and city damage reached 75% cap)
*If this algorithm kicks when turn reaches 25, parties are exhausted, your defending units are pushed an adjacent square overland, if available (some or all).
So what's the result? Some/all of your units get kicked out, enemy does a second attack with an available stack, and you have a higher chance of losing. On the other hand, you have a chance of surviving a turn or so. So the concept is a bit more moderated.

Idea #2 (attacking rewards, my favorite of the 3)
*Attacking party, if reaching the cap city destruction, gains gold based on what the town has, similar to razing. This amount of gold can be significant for big cities. Maybe win some fame as well!
*Population destruction cap being 99% rounded down.

Idea #3 (minor)
*if algorithm that AI thinks player is not too threatening if escaping (horse archers, etc), make it more likely to go in town and reach destruction cap first.