Welcome

Please register for Total War Access to use the forums. If you're an existing user, your forum details will be merged with Total War Access if you register with the same email or username. For more information please read our FAQ’s here.

Categories

An overview of the problem with sieges and potential solutions.

2»

Comments

  • nrdilodoviconrdilodovico Registered Users Posts: 59
    All this talk of siege AI will be laughable in a year or so. Machines are taking over, 40k is closer then it seems. Agree with original poster with caveat...simple is better.
  • Galvinized_IronGalvinized_Iron Registered Users Posts: 830
    Realistically siege battles should simply be all about capturing the walls. Having street battles is kinda unnecessary so I would be in favor of reducing them to just the wall.

    Reason is of course that a commander can't realistically issue orders anymore when it breaks down into streetfighting. Plus historically defenders pretty much always lost once walls were breached.
  • LayzanLayzan Registered Users Posts: 954
    I get more results from not manning the wall.
  • LayzanLayzan Registered Users Posts: 954
    edited September 2017
    We at least need to be able to attack siege towers with more than the lame defense towers outside of manning the arty if that even works.
  • Vanilla_GorillaVanilla_Gorilla Registered Users Posts: 19,331
    Layzan said:

    Someone once wrote here that sieges are are more like a road bump to the enemy, I'm inclined to agree.

    I wouldn't call a 1.5 - 2x force multiplier a bump, but each to their own.
    Malakai is the best choice for a Dwarf LP. Give us Slayer lords so we may form a Slayer host and revel in our destruction!
  • Grimgor_the_CAkeGrimgor_the_CAke Registered Users Posts: 1,703
    One day CA not setting option(s) for players to tailor-made battles in campaign like custom battles using 'excuse' like angry crowds from the downtown, suburb(s), campaign battles can't reach its fullest potential and still suffer the cycling black-hole of auto-resolve. The auto-resolve black-hole not only a waste of efforts from the talented artists from CA, but also deprives the Total War Warhammer tourists the right to enjoy breath-taking scenes of battles in the Old World and the New World.

    As an angry tourist, I demand that a free for all Game 3 release by tomorrow morning, or I will consider a refund of the bloody hell DLC.

    Well then... what fun in a big well crafted city garrisoned by 300 models (2-4 units lord / hero included) defending against 1000 models? And, how realistic it is?

    I like some ideas pointed up above. This one particular: no one enjoys models crumbling together as most of them are not fighting but watching other frontline models fighting. But such mechanics has been there since older titles.

    I would suggest one thing.

    An option to "dismount" (i.e. divide) an unit into 2, or even 3 in game, subject to a max. controllable card limit.

    And, let's go back to the "excuse" as everyone likes excuse.

    "The "excuse" to increase / decrease armies participating a battle / siege in this discussion can be many, to name a few:

    To increase defense for a settlement:

    1, You are sieging one settlement, not the region;
    2, You are sieging part / a corner of a settlement, not the whole settlement;
    3, You cannot completely siege a settlement due to your army size in terms of no. of models (say for example, it is not reasonable to siege a settlement by 20 models - 1 x lord, 3 heroes, 16 monsters); and,
    4, There are hundreds of thousands population of the defending race residing inside a settlement, a portion of them can be enlisted because they may face raze of their homes. Many of them are ready to be expendable units, to say the least.

    To decrease defense models (I prefer to use "models" as no. of models involved in a battle is a more important indicator for measurement of the scale of a battle) for a settlement:

    1, shortage of food; and,
    2, scared by sieging faction.

    To increase models for sieging:

    1, assistance from nearby controlled regions;
    2, the region of the sieged settlement has been your race's traditional area of residence; and,
    3, unhappy populations within a region / settlement due to race conflict / political causes (that may mean you can have some surplus models of other faction).

    To decrease models for sieging;

    1, the sieging faction lost some key battles recently in other (nearby) places;
    2, supply line problem (hostile locals disrupting supply); and,
    3, logistic problem (your force is cut-off-ed from nearest force / region under your control).

    To avoid controlling too many cards and still maintaining the no. of models in battle same time, surplus models for human players can be directly controlled by AI (optional would be nicer).

    Written by an angry man with righteous fury and consented by Sigma and the Grimgor without mount!
  • FloppingerFloppinger Registered Users Posts: 418
    Itharus said:

    ....

    Somehow Shogun 2 managed to get a good balance going on. Like, what slipped between Shogun 2 and Rome 2?

    Rome 2 had many issues with sieges. Huge cities that where hard to navigate for the AI, derpy phalanxes, factions that had sword infantry capable of taking/defending walls properly and factions that didn´t (oh look they´re greeks and I´m not, so I get to auto massacre everything they have on walls, gg I win).

    The problem with M2TW and Rome sieges was that the AI couldn´t negotiate chokepoints in any reasonably capable way.

  • endurendur Registered Users Posts: 3,601
    My suggestions.

    1. Tower Improvements
    Wider arcs for towers, so that one tower can cover double or triple the current width to prevent corner camping outside of tower arcs. Improved AI so that the AI mans the tower garrison spaces more often. Increased map size/reduced range for towers, so that attackers can have a reserve space out of range of towers (and also out of range for attacking siege engines). Towers should have range equal or greater than attacking siege engines.

    2. Improved AI for targeting enemies. Instead of firing on closest enemy, priority should be on AI targeting most dangerous enemies: i.e. target enemy siege engines first. Then casters. Then generals. etc.

    3. Fix battering rams. Currently useless, too slow, and units do plenty of damage to gates with swords. Even trolls should benefit from having a battering ram against a gate.

    4. Improve AI priority for defending walls, gate, and other key locations. Don't always put the general behind the main gate. Instead, put the best unit, whatever that best unit is, behind the gate. i.e. a unit of Knights of the Realm should be behind the main gate instead of a unit of bretonnian swordsmen.

  • Arthas_MenethilArthas_Menethil Senior Member Registered Users Posts: 5,994

    Itharus said:

    ....

    Somehow Shogun 2 managed to get a good balance going on. Like, what slipped between Shogun 2 and Rome 2?

    Rome 2 had many issues with sieges. Huge cities that where hard to navigate for the AI, derpy phalanxes, factions that had sword infantry capable of taking/defending walls properly and factions that didn´t (oh look they´re greeks and I´m not, so I get to auto massacre everything they have on walls, gg I win).

    The problem with M2TW and Rome sieges was that the AI couldn´t negotiate chokepoints in any reasonably capable way.

    Rome 1 and the path finding on any gate where you MUST walk though the gate or take a very long time getting a unit though is made worse with the oil thing at gates honestly can't remember its proper name.
    So...the Light's vaunted justice has finally arrived. Shall I lay down Frostmourne and throw myself at your mercy, Fordring?

  • TheGuardianOfMetalTheGuardianOfMetal Senior Member Registered Users Posts: 11,228
    edited September 2017

    Itharus said:

    ....

    Somehow Shogun 2 managed to get a good balance going on. Like, what slipped between Shogun 2 and Rome 2?

    Rome 2 had many issues with sieges. Huge cities that where hard to navigate for the AI, derpy phalanxes, factions that had sword infantry capable of taking/defending walls properly and factions that didn´t (oh look they´re greeks and I´m not, so I get to auto massacre everything they have on walls, gg I win).

    The problem with M2TW and Rome sieges was that the AI couldn´t negotiate chokepoints in any reasonably capable way.

    Rome 1 and the path finding on any gate where you MUST walk though the gate or take a very long time getting a unit though is made worse with the oil thing at gates honestly can't remember its proper name.
    Murder Hole? Machicolation?

    I wonder how utterly annoying a Proper Gate House would be in TW... or even a proper Castle Layout that is completely geared towards defense... 2 Gate houses, one is around a corner and there is not enough space for the battering ram... murder holes in the gatehouse so your units are getting hit with hot sand/oil/water and arrows... and rocks thrown from above... and so on
    Every wrong is recorded. Every slight against us, page after page, ETCHED IN BLOOD! Clan Gunnisson! Karak Eight Peaks! JOSEF BUGMAN! TOTAL WAR TROY FOR ONE YEAR EXCLUSIVELY ON THE EPIC GAMES STORE!"

    The Empire still hasn't gotten their FLC LL. We need Marius Leitdorf of Averland!

    Where is Boris Todbringer? Have you seen him? For a Middenland DLC with Boris and the Ar-Ulric!

    Queek could smell their hatred, ratcheted to a degree that even he could not evoke in their simple hearts. He stepped over the old orange-fur’s body, eager to see for himself what it was they saw. But he heard it first.
    'Waaaaaaaggh! Gorfang!'
  • ItharusItharus Senior Member Registered Users Posts: 9,374
    Regarding #3, the problem is the gate itself I think. They need to be a LOT tougher. And have boiling oil, portculis, murder holes, arrow slits...

    #1, yeah why towers don't have full coverage is beyond me.

    #2, that might not be acceptable for CA... they are kind of pushing the whole "chaff" units for soaking up fire. like peasants/goblins/clanrats.

  • EaglePhoenixEaglePhoenix Senior Member Registered Users Posts: 1,270
    I quite like the idea of having several stages of the siege for unique cities.

    First the walls, then a fight for the city - then it's at least more than a rush for the walls.

    That being said, I have yet to fight the ai as the defender myself. Like, honestly, they never attack but just wait and starve out. Perhaps a valid tactic but boring AF. Missed many possible good fights like that. The only way to fight as the city defender is through custom battles.

    Plus if you move another army in, they just retreat anyway, or it becomes just another normal map battle.

    Not much in difference to me sieging them, because no matter the odds - they will never Sally out and will starve instead. Even if they outnumber me 3 to 1.

    Might as well remove sieges all together.
    Team Empire!..and also tiny lil bit VC - but shhh
  • Vanilla_GorillaVanilla_Gorilla Registered Users Posts: 19,331
    @EaglePhoenix If anything there are too many siege battles as is, doubling that number seems like an odd choice.

    Do you have mods on? The AI certainly attacks cities and walled settlements, and they do indeed sally out.
    Malakai is the best choice for a Dwarf LP. Give us Slayer lords so we may form a Slayer host and revel in our destruction!
  • GwydionGwydion Senior Member Registered Users Posts: 2,203
    agreed with almost everything. Sieges are the worst part of this game hands down.
    PLEASE CA!!! Chaos Warriors need a faction that is not horde only by the time the trilogy is finished! We beg of you!

    Just Realized this topic has been viewed more, the topic of " Limiting Race Expansion/ Colonization Expansion" alone than more than half of the stickied things at the top of the forum... I hope you are too CA and I mean that in a positive way from a huge fan!

    Please improve sieges! Add racial flavor and ACTUALLY make them "deeper rather than wide" copy and paste with different art needs to go!
  • blaatblaat Junior Member Registered Users Posts: 3,114

    @EaglePhoenix If anything there are too many siege battles as is, doubling that number seems like an odd choice.

    Do you have mods on? The AI certainly attacks cities and walled settlements, and they do indeed sally out.

    I need mods to get them to do that

    snip

    It's much easier and more fun to get engrossed in lore that takes itself seriously and tries to make sense within its own frame of reference.

    the reason I prefer LOTR over warhammer fantasy and 40k

    I am dutch so if you like to have a talk in dutch shoot me a PM :)
  • DebaucheeDebauchee Junior Member Registered Users Posts: 1,406
    With so many regressive innovations in TW:WH I wouldn't be surprised if siege battles are autoresolve only in the next historical title.
  • weefrazeweefraze Registered Users Posts: 31
    endur said:

    My suggestions.

    4. Improve AI priority for defending walls, gate, and other key locations. Don't always put the general behind the main gate. Instead, put the best unit, whatever that best unit is, behind the gate. i.e. a unit of Knights of the Realm should be behind the main gate instead of a unit of bretonnian swordsmen.

    I mentioned something similar in my initial post, about the ai assigning a small force to defend the gate. You raise a further point that's actually really important, one of the units assigned to the gate seems to be (very often, always?) the general and rarely does he have a force within range to defend him if he gets swarmed.
  • FloppingerFloppinger Registered Users Posts: 418
    endur said:

    My suggestions.
    ......

    4. Improve AI priority for defending walls, gate, and other key locations. Don't always put the general behind the main gate. Instead, put the best unit, whatever that best unit is, behind the gate. i.e. a unit of Knights of the Realm should be behind the main gate instead of a unit of bretonnian swordsmen.

    I´d even go a bit further. There´s some basic things players have been doing for ages in TW sieges, that the AI imo could do aswell.
    Like defending the gate not by putting a unit directly behind it, but by using a unit to the left and to the right of the gate to sandwich whatever comes through. At least if the garrison is large enough.

    Or moving units that are under fire and take unnecessary damage off the walls and to the street directly below, thus hiding it from the vast majority of missile fire, and then moving it back up again when the enemy is sufficiently close.


  • uriakuriak Registered Users Posts: 3,421
    The AI hides many units behind the walls, but doesnt't really correctly cheks whether it's good to put/remove a unit. Let's be honest nothing is trivial there and any scripted behavious is susceptible to be "gamed" by players.

    I don't mind the AI not doing some things against "cheesy" strategy. I do mind when it seems to do silly stuff when the players doesn't try to game it. For instance the whole not sending the army on the walls at once, or trying to open a second breach. I understand spreading is considered unsound to avoid tower damage, but sending only portions of the army opens to ranged fire.

    There may be a self fulfilling prophecy here : people don't like sieges because the design is repetitive beaucause the AI needs way more scripting but at the same time it's time spent on a battle mode avoided by the players so it's not as important as regular battles etc

  • Grimgor_the_CAkeGrimgor_the_CAke Registered Users Posts: 1,703
    edited September 2017
    Part 2 comments

    Is AI bad a fact / myth?

    Many people have been saying the AI is bad both in campaigns and in battles. However, according to my experience, the AI has been outsmart and surprising me on many instances. For examples:

    1, placing a tanky unit behind gate and an artillary unit somewhere behind the that tanky unit. Opened the gate and used artillary to shot my units. When I moved another unit to rush thru. It closed the gate.

    Although this scenario occurred only once (partly due to the charms of auto-resolve), it led me to conclude that the AI can think like a normal human if it is given the tools - enough (good) models.

    2, the AI specifically rushed to my caster and knocked him out in field battles. Probably I have slow hands (and brain), these happened more than a few times. That is, given enough space, the AI can outperform human due to its multi-core nature.

    What does this mean for siege battles? While choke point design may help AI defender, more open field behind the wall does not necessarily bad for the AI.

    Choke point, good or bad?

    One of the worst things in design of wargames, MOBA and strategy games is choke point. Parking a tanky unit somewhere = auto-win. What fun is that? How strategy is that?

    However, real life choke-point does not really provide such overwhelming advantage to the defenders. So, what have been gone wrong isn`t choke point the concept but in fact the way things have been simulated.

    From this, we can come up to a few assumptions:

    1, some tanky units are unrealistically too tanky;
    2, AI fails to arrange its most suitable units to take down the human defender at choke point and wastes too many other units in grinding (compared to open field AI does not have such path / target finding problem);
    3, capture point location assignment badly chosen at times.

    Possible Solution

    Setting choke point as a terrain and giving de-buff to defending models. Time related de-buff can work as something like fatigue (maybe an accelerated version).

    The de-buff timer resets after the defending models leaving the choke point zone, or when they are not suffering damages for 10 seconds (say). The de-buff gradually fades away when either condition is not satisfied.

    The timer set to initiate when models start engaging enemies (getting fired / suffering magic damages alike) at which time within 20m (say*) there are 2-3 (say**) higher/equal tier enemy units and those units have over 60% (say***) remaining models of the unit size.
    * 20m to include a reasonable cluster of nearby attackers
    ** 2-3 higher/equal tier units to reflect the fact that the enemies should reasonably be able to take down the weaker choke point defenders
    *** 60% to avoid too few attacking models being able to take down defenders unreasonably

    A bit complicated but this solves ages of choke point problem of total war games.
    Post edited by Grimgor_the_CAke on
  • ItharusItharus Senior Member Registered Users Posts: 9,374

    @EaglePhoenix If anything there are too many siege battles as is, doubling that number seems like an odd choice.

    Do you have mods on? The AI certainly attacks cities and walled settlements, and they do indeed sally out.

    There are only too many because they are unsatisfying. Take some lessons from Medieval and you might be begging for more :-P
  • Combat_WombatCombat_Wombat Registered Users Posts: 4,092
    Pretty good ideas from the OP. Although their suggestions from improvements aren't extreme enough in my opinion.
Sign In or Register to comment.