Page 1 of 2 12 LastLast
Results 1 to 20 of 25

Thread: Records Mode Gameplay Reveal

  1. #1

    Default Records Mode Gameplay Reveal

    They have recently published a video showing Records mode gameplay:



    The armies are similarly structured. Each 6 unit has a general unit. Each general has accompanying soldiers based on their style.
    Last edited by PointOfViewGun; February 19, 2019 at 09:09 AM.
    The Armenian Issue

  2. #2
    ♔Greek Strategos♔'s Avatar THE BEARDED MACE
    Artifex Moderator Emeritus

    Join Date
    Feb 2008
    Location
    Athens, Greece
    Posts
    11,588

    Default Re: Records Mode Gameplay Reveal

    Τhe Sepia colored map looks fitting, nonetheless

  3. #3

    Default Re: Records Mode Gameplay Reveal

    Does anyone know how it is possible that archers shoot at units they can't see because landscape blocks their line of sight?

    And the horses in the attack probably have nitro or something similar turned on

  4. #4

    Default Re: Records Mode Gameplay Reveal

    Quote Originally Posted by musketer View Post
    Does anyone know how it is possible that archers shoot at units they can't see because landscape blocks their line of sight?

    And the horses in the attack probably have nitro or something similar turned on
    The games have been letting archers shoot at targets they can't actually see for years so long as there's a valid firing arc. It's their biggest advantage over gunpowder units in later titles.

    What matters (as of Rome 2) is if something in your army can see it. LOS is army wide, so as long as someone can see the unit, they presumably signal to the archers in what general direction to fire. Dunno if blind firing or high arcs has ever affected accuracy in game though.
    Last edited by zoner16; February 19, 2019 at 01:42 PM.
    My Three Kingdoms Military History Blog / Military Map Project - https://zirroxas.tumblr.com/
    Ask me a question!

  5. #5
    Seether's Avatar RoTK Workhorse
    Join Date
    Oct 2005
    Location
    FloRida
    Posts
    5,404

    Default Re: Records Mode Gameplay Reveal

    Highly disappointed.

    Records mode is just Romance mode except generals have bodyguards and can’t demolish everything in sight. Whoopty doo...

    And “Shu-Han Kingdom” makes zero sense, especially in eastern China and in the “historical” records mode.
    1) “Shu-Han” was never the name of Liu Bei’s kingdom or empire. Ever. Neither was it ever called ”Shu”.
    2) Liu Bei’s territory was called “Han”, and as a continuation of the Han dynasty usurped by Cao Pi and the state of Wei.
    3) Shu was the name of an older kingdom in western China, which was conquered by Han when it united the country after Qin, where Liu Bei eventually became based. Having Shu in the east is just lazy.
    4) ”Shu-Han” is literally not a land or title or anything. It’s contemporary made-up nonsense.
    5) Liu Bei’s Han is incorrectly referred to as Shu or Shu-Han only to differentiate his dynasty from the one usurped by Cao Pi.
    6) Liu Bei was King of Hanzhong, not King of Shu; the latter title never existed in this era. So “Shu-Han kingdom” makes zero sense.

    So having a “Shu-Han kingdom” led by Liu Bei, in eastern China, not even remotely close to the actual land of Shu, is just sloppiness on CA’s part. It would not be hard to make King/Emperor titles to be dynamic. The names literally would be wherever the clan leadership has their primary title and base of power. Hence Wei, Wu and Jin were named so because the ruling family’s base of power was located in those older kingdoms and, per tradition and precedence, named their dynasty after it.

    C’mon CA, this was supposed to be an historical title. Again, highly disappointed.
    Last edited by Seether; February 19, 2019 at 02:18 PM.
    Member of the Imperial House of Hader - Under the Benevolent Patronage of y2day
    A Wolf Among Sheep: A Rise of Three Kingdoms AAR

  6. #6

    Default Re: Records Mode Gameplay Reveal

    I don't have a problem with them using the term "Shu-Han" persay, Chinese historiography has always used qualifiers like that to distinguish different epochs (ie, the Later Han, the Northern Song, etc). It's like using "Byzantine" as a term - not accurate to what the people themselves said but perfectly fine as a shorthand term.

    That said, you should have to have your base actually in Shu to give you that title. This effects more than Liu Bei and his claim on the Han Dynasty, as every Dynasty that rises is going to take its name from an analogous title.

    Anyways, the actual gameplay looks fun. I know some people will probably be disappointed that it looks mostly like Romance mode only with bodyguards and longer battles but that's really all I expected.

  7. #7
    Seether's Avatar RoTK Workhorse
    Join Date
    Oct 2005
    Location
    FloRida
    Posts
    5,404

    Default Re: Records Mode Gameplay Reveal

    I can agree with that - as I noted above - about actually having to be in Shu to be called Shu-Han. From where his base was in the video, he should’ve been called “Lu-Han” or “Song-Han”.

    As far as gameplay, the unit collision looks awkward. I don’t remember the minute mark, but the collision of the two infantry units close-up looked very strange and the animations were weird.
    Member of the Imperial House of Hader - Under the Benevolent Patronage of y2day
    A Wolf Among Sheep: A Rise of Three Kingdoms AAR

  8. #8

    Default Re: Records Mode Gameplay Reveal

    What's interesting is that this is the closest we've gotten to the old TW games since Medieval 2. No matched combat, slow as , and no general abilities at all, not even morale buffs or brace.


    ​Scoodlypooper Numero Uno

  9. #9

    Default Re: Records Mode Gameplay Reveal

    Quote Originally Posted by Mantis42 View Post
    Anyways, the actual gameplay looks fun. I know some people will probably be disappointed that it looks mostly like Romance mode only with bodyguards and longer battles but that's really all I expected.
    Basically my thoughts on the matter. It's all I was asking for TBH. Everything else would be a straight quality improvement, and I don't see why Romance mode can't be better in general too just because I'm barely going to play it.

    Quote Originally Posted by Seether View Post
    I can agree with that - as I noted above - about actually having to be in Shu to be called Shu-Han. From where his base was in the video, he should’ve been called “Lu-Han” or “Song-Han”.

    As far as gameplay, the unit collision looks awkward. I don’t remember the minute mark, but the collision of the two infantry units close-up looked very strange and the animations were weird.
    Since the Han empire is still in the game, I think he probably should just be the King/Prince of Lu (since Yuan Tan is apparently Duke of Song) since he'd be playing the role of a vassal king up until he pulls a Liu Xiu and screws over his liege.

    It's a really minor thing, and I get that people expect Liu Bei to be "Shu" because that's what they know him by and only a few people over here are well versed in dynasty naming connections to Spring and Autumn era states. If it's possible, I'd like to be able to mod that to be more region determined, with the exception that Yuan Shu has to name his state Zhong no matter where he is.

    Infantry colliding is at 6:40. It does look a bit awkward, but lines are coherent. Air swording abounds, animations are too quick, and people are still flying too far back when getting knocked around, but units aren't spilling into each other and I don't think I'd squint at it if watching from above.
    Quote Originally Posted by ptoss1 View Post
    What's interesting is that this is the closest we've gotten to the old TW games since Medieval 2. No matched combat, slow as , and no general abilities at all, not even morale buffs or brace.
    Med 2 still had rally if I remember correctly. I think the Rally/Inspire/Stand and Fight combo that Shogun 2 had was the best. Nothing crazy, but still good old charisma.

    I remember asking Grace on Reddit if there were any abilities in Records mode and she seemed to indicate that Rally was going to be there, so either they changed that plan or it's unlocked somehow.
    My Three Kingdoms Military History Blog / Military Map Project - https://zirroxas.tumblr.com/
    Ask me a question!

  10. #10

    Default Re: Records Mode Gameplay Reveal

    Quote Originally Posted by Seether View Post
    Highly disappointed.

    Records mode is just Romance mode except generals have bodyguards and can’t demolish everything in sight. Whoopty doo...
    This is probably a title to skip, or wait on sale. On top of that apparently the ai still looks stupid.

  11. #11
    Seether's Avatar RoTK Workhorse
    Join Date
    Oct 2005
    Location
    FloRida
    Posts
    5,404

    Default Re: Records Mode Gameplay Reveal

    More than likely will wait for a good sale. Although I’m still hoping in a few years it will be fixed and polished, much like Rome 2 eventually became, I’m not holding out hope. Until then I always have my M2TW mod
    Member of the Imperial House of Hader - Under the Benevolent Patronage of y2day
    A Wolf Among Sheep: A Rise of Three Kingdoms AAR

  12. #12
    Skotos of Sinope's Avatar Macstre Gaposal
    Join Date
    Sep 2017
    Location
    The Republic of Letters
    Posts
    789

    Default Re: Records Mode Gameplay Reveal

    Waiting until a Warring States DLC...

  13. #13

    Default Re: Records Mode Gameplay Reveal

    I don't know what people were expecting. Romance mode without heroes is basically just the usual TW game. Wasn't that what everybody wanted? Ehhh...


    ​Scoodlypooper Numero Uno

  14. #14

    Default Re: Records Mode Gameplay Reveal

    So I had some time to go through the video some more and do some more frame-by-frame observation now that my initial impressions have settled (asterisk(*) is time-stamp, quote(") means same as previous):

    Map Stuff

    • I was wondering about the marks on the diplomacy/grand-strategy map. I figured out that the shield icon+portrait means a unit is garrisoned in a settlement and the triangles are ungarrisoned armies. Why only the garrisoned troops have portraits is beyond me, but it does clear that up(*).
    • The "Kingdom of Shu-Han" is still dumb for reasons already explained. However, yes, it is ultimately a minor complaint, and most of those without specific familiarity with that aspect of Chinese nomenclature will just want the big three to have the kingdom names that history gave them(").
    • The dragon symbol and the gold text might be a new way of denoting the imperial protector. It wasn't there in previous videos(").
    • Bogui's looking spry for a 68 year old(*).
    • If there's something that this game has down pat, it's interesting minor settlement battle maps. This is definitely a trend I want to see going forward. Rome 2's samey generic urban grids without walls and Throne's ignoreable huts in a corner just don't provoke the same sense of anticipating for fighting on them. The battlefield looks great without being cluttered, and the terrain is varied while still being coherent(*).


    Unit Stuff

    • It only occurred to me while viewing the garrison troops' unit cards, but it seems that CA has decided that everyone who wields a dao (sabre) gets a round shield and everyone who wields a jian (straight sword) gets a double arc shield(*). I used to be under the impression that the round rattan/wicker shield was a later adoption, but it's been cropping up more and more in discussions and media I have on classical era China, so maybe there's been new discoveries.
    • Ma Chao seems to have two crossbow units that are absolutely supposed to be Kong Rong's unique units (Fury of Beihai, Thunder of Jian'an). Dunno if he did a stint in Kong Rong's army and then came back or if this is another bug(*).
    • There was some consternation on Teiba that soldiers were too thin, and I can kinda see it, but if they hadn't pointed it out, I wouldn't have noticed. It's probably alright(*).
    • Jian Swordguard are kind of a mess. The jian are too thick, the notches on the shields are too pronounced, and their use of a plain leather cuirass over a lamellar one doesn't fit a supposedly professional unit. Of all the units in the game, these guys look like they're straight out of the late Warring States rather than the Later Han(*).
    • Similarly, Lance Cavalry should have some armor. Being unarmored should be for peasants/bandit units(*).
    • Ji Infantry look alright. The halberd heads are a bit out of date, but they're at least imperial era(*).
    • Crossbows of both types look good(*).
    • The White Horse Fellows look rather cheap. I get that they split the White Horse Volunteers into two units (Fellows and Banner) to prevent them from being overpowered like Attila's Hunnic cavalry were, but they could still have made them both heavy cavalry, just one with a bow and the other with a lance(*).
    • Gongsun Zan's bodyguards look like a proper bodyguard(*).


    Charge Stuff

    • Cavalry charges on infantry are mostly good. Most units only go flying a couple feet if they get knocked aside. A couple go a few meters but not to any degree that I'd be taken aback if not zoomed in. Additionally, they go straight through to the unit behind when not in formation, which feels like cavalry charges finally have proper impact in a historical game, even if the animations are a bit much(*).
    • Cavalry charges on cavalry look very good (no close up unfortunately). No more Rome 2 era nonsense where they just stick into each other. Horses and riders actually die off the charge and units are fluid enough to pass through gaps. There's a little bit of shove and stick, but only a couple models out of the unit get caught by this, and the riderless horses running off afterwards is a great touch(*).
    • Infantry on infantry charges are alright. As I said earlier, the air-slashing, high knockback, and overly fast attack animations are still there, but less noticeable and the unit are fairly coherent and not blobbling together(*).


    Combat Stuff

    • Fatigue does seem to be important, which is good, but I worry that it might be draining a bit too fast. Gongsun Zan went down to exhausted after two charges+melees. Battles shouldn't be lengthened just by people running out of breath almost immediately. Vigor works if it becomes an issue later in the battle when you have to pick and choose what units get committed near the end (*).
    • The ax band ultimately fight the militia at the south entrance for over six minutes. Not bad for low tier units(*).
    • We see the effects of charge reflection eventually when Gongsun Zan and Zhao Yun charge a unit that turned to face them at the last second. I fear this might be too powerful if a unit doesn't even have to be braced or in formation to enable charge reflection on cavalry. Gongsun Zan loses two thirds of his men in that charge and Zhao Yun loses half. I don't even think they were good units. At least two of them were Peasant Band and the other was just Ji Infantry(*).
    • Chain rout from army losses and capture point loss at about 3/4ths of the balance bar(*).
    • Total battle length for a minor settlement between a stack and a stack and change was about 17 minutes(*).
    • The final attack is Ma Chao getting knocked off his horse by...a gust of wind it seems. Nobody swung their swords. I'd complain about animations and whatever, but Ma Chao is a douchebag and probably deserves it(*).
    • Persistent corpses look good. Very Shogun 2-esque, which has always been my favorite in terms of post-battle carnage(*).


    AI Stuff

    • The AI is something I can still very clearly see having problems, but I can at least understand why they're happening. It's usually having trouble executing some maneuver that, if the flaws were fixed, would be a step ahead of it's old capabilities:



    1. When Gongsun Zan's cavalry break through to the north, it moves its cavalry that was backing up the infantry at the other entrances to plug the breach. A good idea, since the cavalry would get there the fastest, and wouldn't be of much help in the infantry grind. It does then have some issues pulling this off completely, as one of the units either has second thoughts upon getting into range of the horse archers, or wants to go back to help the infantry at another entrance. That hesitation is what gets it caught by a charge from said horse archers. The AI's difficulties committing to a single maneuver are something that comes up a lot in the video (and other videos).
    2. The AI isn't pulling out of melees that it has no hope of escaping or melees its doing well in just because breaches are happening elsewhere, which is good. That was a big complaint of mine from January videos. It could be suboptimal in some situations, when it'd technically be better to attempt to pull back, but I'd rather the AI gave a good fight rather than constantly retreat to consolidate.
    3. That pathfinding bug comes up, but they pointed it out and it doesn't appear to be a systemic issue, so I won't go into detail.
    4. We can kind of see the beginnings of the issue the AI is going to have in the late stage of the battle and why it's happening. The reinforcement army is gathering at the center of the settlement in a defensive arrangement, which isn't a bad thing, but quite obviously, the cavalry arrived first and the infantry is in a column. The horse archers are about to aggro the enemy foot archers, who will stand their ground and shoot, but the melee infantry will keep moving because they're not in range yet. This is ultimately what causes the archers to get separated from the rest of the enemy army and never reach their destination. Immediately after beating off the cavalry, the enemy archers stay where they are and switch targets to the ji infantry now in their range. Again, not a bad idea on its own, but now they're exposed (their melee support either heading to the rendezvous or engaging said Ji Infantry) and this is what gets them all killed. A big mistake, but one I've also seen out of human players, so I won't knock the AI too hard here.
    5. The next issue is more problematic. Ma Chao and the units that made it to their gathering point begin being very indecisive and running back and forth. A unit of cavalry wanders into range of the crossbows, but then just canters forward rather than either charging or fleeing right away, while Ma Chao and the rest of the cavalry seem to be unable to decide if they want to go fight Gongsun Zan or not, running back and forth. This leads to a lull that allows Gongsun Zan to finish killing the archers while the other unit of cavalry eventually charges the Jian Swordguard wall on their own. I can't tell if they were trying to rearrange their troops to defend the area they were in or trying to pick a target to attack.
    6. The AI switches tactics again after the archers are all dead. Ma Chao and his cavalry attempt to flank Gongsun Zan's infantry, while about half of his infantry attempt to attack Gongsun Zan's own flanking force.
    7. The first maneuver goes poorly almost immediately. You can see that Ma Chao is attempting to move around via the side street, but he only gets about halfway before trying to come back for some reason. It's a terrible mistake because that direction was only guarded by ammo depleted archer units. This may have been due to the AI not wanting to leave the way into the settlement center undefended, but even after Ma Chao and some of his units go back to plug the gap, about half the remainder begin going on the flank attack again, only to turn right around at the halfway point once more and return to fight the shieldwall with Ma Chao or sit in the settlement center. Only one unit of halberds eventually completes the maneuver, and quickly overwhelms the defenders. If they had been joined by other units, they might've turned that flank.
    8. The other maneuver goes a bit better, but the spear unit closest to Gongsun Zan's cavalry initially quits it charge and falls back instead of advances, leaving the sword unit at the front unsupported. It eventually turns back towards the fight after the other two units advance past it, but by that time it's too late. It might've been able to catch some of the cavalry if it had continued forward. Again, I don't know if it wanted to defend the square or was trying to link up with reinforcements, but it needed to commit to its attack rather than second guess itself.
    9. When both those maneuvers fail, the AI is essentially doomed and there's no way left to bail itself out, so there's nothing more to say.



    Overall, if CA clean up the AI issues with committing to plans and maybe improve infantry combat animations some more, I don't think I'd complain too much on launch. Some issues I know they're not going to fix, like the state naming and the anachronistic trebuchets, because not enough people seem to care, but those are things I can live with, or at the very least mod around. The core game needs to be solid for mods to have a chance of smoothing the details. Since it looks passable right now and they just gave themselves three more months of dev time, I'll dare to hope it's a decent release that I don't have to wait to be fun.
    Last edited by zoner16; February 20, 2019 at 03:22 AM.
    My Three Kingdoms Military History Blog / Military Map Project - https://zirroxas.tumblr.com/
    Ask me a question!

  15. #15
    ♔Greek Strategos♔'s Avatar THE BEARDED MACE
    Artifex Moderator Emeritus

    Join Date
    Feb 2008
    Location
    Athens, Greece
    Posts
    11,588

    Default Re: Records Mode Gameplay Reveal

    Quote Originally Posted by Seether View Post
    I can agree with that - as I noted above - about actually having to be in Shu to be called Shu-Han. From where his base was in the video, he should’ve been called “Lu-Han” or “Song-Han”.

    As far as gameplay, the unit collision looks awkward. I don’t remember the minute mark, but the collision of the two infantry units close-up looked very strange and the animations were weird.
    I find it a bit weird as well.

    Quote Originally Posted by Seether View Post
    More than likely will wait for a good sale. Although I’m still hoping in a few years it will be fixed and polished, much like Rome 2 eventually became, I’m not holding out hope. Until then I always have my M2TW mod
    AFAIK 3K has already the most pre orders in total war history and in my book that's not a good thing. It'll make the devs less ''active'' IMO.

    Quote Originally Posted by zoner16 View Post
    So I had some time to go through the video some more and do some more frame-by-frame observation now that my initial impressions have settled (asterisk(*) is time-stamp, quote(") means same as previous):

    Spoiler Alert, click show to read: 
    Map Stuff


    • I was wondering about the marks on the diplomacy/grand-strategy map. I figured out that the shield icon+portrait means a unit is garrisoned in a settlement and the triangles are ungarrisoned armies. Why only the garrisoned troops have portraits is beyond me, but it does clear that up(*).
    • The "Kingdom of Shu-Han" is still dumb for reasons already explained. However, yes, it is ultimately a minor complaint, and most of those without specific familiarity with that aspect of Chinese nomenclature will just want the big three to have the kingdom names that history gave them(").
    • The dragon symbol and the gold text might be a new way of denoting the imperial protector. It wasn't there in previous videos(").
    • Bogui's looking spry for a 68 year old(*).
    • If there's something that this game has down pat, it's interesting minor settlement battle maps. This is definitely a trend I want to see going forward. Rome 2's samey generic urban grids without walls and Throne's ignoreable huts in a corner just don't provoke the same sense of anticipating for fighting on them. The battlefield looks great without being cluttered, and the terrain is varied while still being coherent(*).


    Unit Stuff

    • It only occurred to me while viewing the garrison troops' unit cards, but it seems that CA has decided that everyone who wields a dao (sabre) gets a round shield and everyone who wields a jian (straight sword) gets a double arc shield(*). I used to be under the impression that the round rattan/wicker shield was a later adoption, but it's been cropping up more and more in discussions and media I have on classical era China, so maybe there's been new discoveries.
    • Ma Chao seems to have two crossbow units that are absolutely supposed to be Kong Rong's unique units (Fury of Beihai, Thunder of Jian'an). Dunno if he did a stint in Kong Rong's army and then came back or if this is another bug(*).
    • There was some consternation on Teiba that soldiers were too thin, and I can kinda see it, but if they hadn't pointed it out, I wouldn't have noticed. It's probably alright(*).
    • Jian Swordguard are kind of a mess. The jian are too thick, the notches on the shields are too pronounced, and their use of a plain leather cuirass over a lamellar one doesn't fit a supposedly professional unit. Of all the units in the game, these guys look like they're straight out of the late Warring States rather than the Later Han(*).
    • Similarly, Lance Cavalry should have some armor. Being unarmored should be for peasants/bandit units(*).
    • Ji Infantry look alright. The halberd heads are a bit out of date, but they're at least imperial era(*).
    • Crossbows of both types look good(*).
    • The White Horse Fellows look rather cheap. I get that they split the White Horse Volunteers into two units (Fellows and Banner) to prevent them from being overpowered like Attila's Hunnic cavalry were, but they could still have made them both heavy cavalry, just one with a bow and the other with a lance(*).
    • Gongsun Zan's bodyguards look like a proper bodyguard(*).


    Charge Stuff

    • Cavalry charges on infantry are mostly good. Most units only go flying a couple feet if they get knocked aside. A couple go a few meters but not to any degree that I'd be taken aback if not zoomed in. Additionally, they go straight through to the unit behind when not in formation, which feels like cavalry charges finally have proper impact in a historical game, even if the animations are a bit much(*).
    • Cavalry charges on cavalry look very good (no close up unfortunately). No more Rome 2 era nonsense where they just stick into each other. Horses and riders actually die off the charge and units are fluid enough to pass through gaps. There's a little bit of shove and stick, but only a couple models out of the unit get caught by this, and the riderless horses running off afterwards is a great touch(*).
    • Infantry on infantry charges are alright. As I said earlier, the air-slashing, high knockback, and overly fast attack animations are still there, but less noticeable and the unit are fairly coherent and not blobbling together(*).


    Combat Stuff

    • Fatigue does seem to be important, which is good, but I worry that it might be draining a bit too fast. Gongsun Zan went down to exhausted after two charges+melees. Battles shouldn't be lengthened just by people running out of breath almost immediately. Vigor works if it becomes an issue later in the battle when you have to pick and choose what units get committed near the end (*).
    • The ax band ultimately fight the militia at the south entrance for over six minutes. Not bad for low tier units(*).
    • We see the effects of charge reflection eventually when Gongsun Zan and Zhao Yun charge a unit that turned to face them at the last second. I fear this might be too powerful if a unit doesn't even have to be braced or in formation to enable charge reflection on cavalry. Gongsun Zan loses two thirds of his men in that charge and Zhao Yun loses half. I don't even think they were good units. At least two of them were Peasant Band and the other was just Ji Infantry(*).
    • Chain rout from army losses and capture point loss at about 3/4ths of the balance bar(*).
    • Total battle length for a minor settlement between a stack and a stack and change was about 17 minutes(*).
    • The final attack is Ma Chao getting knocked off his horse by...a gust of wind it seems. Nobody swung their swords. I'd complain about animations and whatever, but Ma Chao is a douchebag and probably deserves it(*).
    • Persistent corpses look good. Very Shogun 2-esque, which has always been my favorite in terms of post-battle carnage(*).


    AI Stuff

    • The AI is something I can still very clearly see having problems, but I can at least understand why they're happening. It's usually having trouble executing some maneuver that, if the flaws were fixed, would be a step ahead of it's old capabilities:



    1. When Gongsun Zan's cavalry break through to the north, it moves its cavalry that was backing up the infantry at the other entrances to plug the breach. A good idea, since the cavalry would get there the fastest, and wouldn't be of much help in the infantry grind. It does then have some issues pulling this off completely, as one of the units either has second thoughts upon getting into range of the horse archers, or wants to go back to help the infantry at another entrance. That hesitation is what gets it caught by a charge from said horse archers. The AI's difficulties committing to a single maneuver are something that comes up a lot in the video (and other videos).
    2. The AI isn't pulling out of melees that it has no hope of escaping or melees its doing well in just because breaches are happening elsewhere, which is good. That was a big complaint of mine from January videos. It could be suboptimal in some situations, when it'd technically be better to attempt to pull back, but I'd rather the AI gave a good fight rather than constantly retreat to consolidate.
    3. That pathfinding bug comes up, but they pointed it out and it doesn't appear to be a systemic issue, so I won't go into detail.
    4. We can kind of see the beginnings of the issue the AI is going to have in the late stage of the battle and why it's happening. The reinforcement army is gathering at the center of the settlement in a defensive arrangement, which isn't a bad thing, but quite obviously, the cavalry arrived first and the infantry is in a column. The horse archers are about to aggro the enemy foot archers, who will stand their ground and shoot, but the melee infantry will keep moving because they're not in range yet. This is ultimately what causes the archers to get separated from the rest of the enemy army and never reach their destination. Immediately after beating off the cavalry, the enemy archers stay where they are and switch targets to the ji infantry now in their range. Again, not a bad idea on its own, but now they're exposed (their melee support either heading to the rendezvous or engaging said Ji Infantry) and this is what gets them all killed. A big mistake, but one I've also seen out of human players, so I won't knock the AI too hard here.
    5. The next issue is more problematic. Ma Chao and the units that made it to their gathering point begin being very indecisive and running back and forth. A unit of cavalry wanders into range of the crossbows, but then just canters forward rather than either charging or fleeing right away, while Ma Chao and the rest of the cavalry seem to be unable to decide if they want to go fight Gongsun Zan or not, running back and forth. This leads to a lull that allows Gongsun Zan to finish killing the archers while the other unit of cavalry eventually charges the Jian Swordguard wall on their own. I can't tell if they were trying to rearrange their troops to defend the area they were in or trying to pick a target to attack.
    6. The AI switches tactics again after the archers are all dead. Ma Chao and his cavalry attempt to flank Gongsun Zan's infantry, while about half of his infantry attempt to attack Gongsun Zan's own flanking force.
    7. The first maneuver goes poorly almost immediately. You can see that Ma Chao is attempting to move around via the side street, but he only gets about halfway before trying to come back for some reason. It's a terrible mistake because that direction was only guarded by ammo depleted archer units. This may have been due to the AI not wanting to leave the way into the settlement center undefended, but even after Ma Chao and some of his units go back to plug the gap, about half the remainder begin going on the flank attack again, only to turn right around at the halfway point once more and return to fight the shieldwall with Ma Chao or sit in the settlement center. Only one unit of halberds eventually completes the maneuver, and quickly overwhelms the defenders. If they had been joined by other units, they might've turned that flank.
    8. The other maneuver goes a bit better, but the spear unit closest to Gongsun Zan's cavalry initially quits it charge and falls back instead of advances, leaving the sword unit at the front unsupported. It eventually turns back towards the fight after the other two units advance past it, but by that time it's too late. It might've been able to catch some of the cavalry if it had continued forward. Again, I don't know if it wanted to defend the square or was trying to link up with reinforcements, but it needed to commit to its attack rather than second guess itself.
    9. When both those maneuvers fail, the AI is essentially doomed and there's no way left to bail itself out, so there's nothing more to say.



    Overall, if CA clean up the AI issues with committing to plans and maybe improve infantry combat animations some more, I don't think I'd complain too much on launch. Some issues I know they're not going to fix, like the state naming and the anachronistic trebuchets, because not enough people seem to care, but those are things I can live with, or at the very least mod around. The core game needs to be solid for mods to have a chance of smoothing the details. Since it looks passable right now and they just gave themselves three more months of dev time, I'll dare to hope it's a decent release that I don't have to wait to be fun.
    Excellent post mate. Very useful analysis

  16. #16
    Seether's Avatar RoTK Workhorse
    Join Date
    Oct 2005
    Location
    FloRida
    Posts
    5,404

    Default Re: Records Mode Gameplay Reveal

    Great post zoner! For me, the battle AI looks as though it is in shambles. There appears to be a complete lack of "communication" between the AI's units, in that every unit acts entirely independent of one another without a "plan" of what the "army" wants to do or how to act. In addition to that, the AI's units spend a lot of time just standing around and doing nothing - much of that is allowing allied units to get overwhelmed and destroyed piecemeal, while other units that could help out (and put up an intelligent defense) stand around and guard (essentially) somewhere tactically unimportant. Units, if they do move, appear do so in a manner that makes it seem as though they are confused and don't know what to do, such as running back and forth without any sort of cause or purpose. While I have mixed feelings about the 2 month delay, I am not at all confident that the battle AI is something that will or can be fixed by CA in that time frame.
    Member of the Imperial House of Hader - Under the Benevolent Patronage of y2day
    A Wolf Among Sheep: A Rise of Three Kingdoms AAR

  17. #17
    Huberto's Avatar Praepositus
    Join Date
    Jun 2008
    Location
    USA
    Posts
    5,313

    Default Re: Records Mode Gameplay Reveal

    This gameplay reveal was really disappointing.

    1.) Now that I've seen two town battles, the maps seem more MOBA-like than any I've seen before in TW -- a network of huge passageways separated by rock formations. I liked Rome 2's and Thrones much better -- becasue they look like actual towns.

    2.) I detest the jokey banter between disembodied generals portraits flying over the battle. Maybe this could work during duels, but never like it's shown here.

    3.) Projectiles seem over done. Can 120 men shoot that many arrows that high and accurately over the battlefield? Just doesn't look realistic.

    4.) Cavalry seems to have a turbo mode where suddently they start moving at unrealistic speeds.

    5.) The unit cards and markers have been discussed elsewhere, and they just don't look good.

    The overall effect of all these aesthetics reminds me it's a video game, not a battle. So it kills immersion, I can't suspend my disbelief, and I'm bored. I literally fell asleep watching this battle. Contrast that to a recent NTW3 battle that lasted almost an hour and I watched it twice.

    Need to wait and see if mods can bring this strange, sad little game to life somewhere down the road. It's not good right now.

    Note how I didn't even discuss the AI.
    Last edited by Huberto; February 20, 2019 at 12:30 PM.

  18. #18

    Default Re: Records Mode Gameplay Reveal

    Quote Originally Posted by Seether View Post
    Great post zoner! For me, the battle AI looks as though it is in shambles. There appears to be a complete lack of "communication" between the AI's units, in that every unit acts entirely independent of one another without a "plan" of what the "army" wants to do or how to act. In addition to that, the AI's units spend a lot of time just standing around and doing nothing - much of that is allowing allied units to get overwhelmed and destroyed piecemeal, while other units that could help out (and put up an intelligent defense) stand around and guard (essentially) somewhere tactically unimportant. Units, if they do move, appear do so in a manner that makes it seem as though they are confused and don't know what to do, such as running back and forth without any sort of cause or purpose. While I have mixed feelings about the 2 month delay, I am not at all confident that the battle AI is something that will or can be fixed by CA in that time frame.
    I think it's because the units are mostly all just fodder trash, and the game is meant to revolve around the heroes. It's designed for them, though it looks a little too dumb even if that was the case...

    I'm pretty disappointed. I wouldn't have minded the game even with how it is, but the battle ai looks like it needs a lot of work (or mods) before it will be worth it.

  19. #19

    Default Re: Records Mode Gameplay Reveal

    Quote Originally Posted by Seether View Post
    Great post zoner! For me, the battle AI looks as though it is in shambles. There appears to be a complete lack of "communication" between the AI's units, in that every unit acts entirely independent of one another without a "plan" of what the "army" wants to do or how to act. In addition to that, the AI's units spend a lot of time just standing around and doing nothing - much of that is allowing allied units to get overwhelmed and destroyed piecemeal, while other units that could help out (and put up an intelligent defense) stand around and guard (essentially) somewhere tactically unimportant. Units, if they do move, appear do so in a manner that makes it seem as though they are confused and don't know what to do, such as running back and forth without any sort of cause or purpose. While I have mixed feelings about the 2 month delay, I am not at all confident that the battle AI is something that will or can be fixed by CA in that time frame.
    So if things haven't changed from past titles, I think you're partially right.

    You're correct that it's not operating as an army, but rather as a city. This minor settlement AI is looks either similar to or the exact same as the city siege AI. Which means that it's actually not trying to destroy the player; it's trying to maintain and exert control of the settlement sectors and the defenses between. The siege AI does this by allocating units to various points of interest (entrances, choke points, victory points) based on its perception of need and feasibility. If there's an enemy unit there, it'll attack it to try to remove it; if not, it just sets up to defend. Anything that isn't allocated to elsewhere is set as a reserve unit and held either on the capture point or some other central location to await a change in the situation. When Ma Chao's force arrives, most of it is evidently set to reserves while it waits for everyone to make it inside.

    Added in on top of this is the AI's new "defense in depth" paradigm which they mentioned back with the first ever gameplay video, which means that the AI doesn't just keep committing more troops to external points of interest by design. The likely reason was that they wanted to have more protracted sieges where the enemy fell back from outer defenses gradually towards the capture point (thus using more of that sparkly map they spent so much time on), rather than just constantly reinforcing the walls or entrances until most everything had died and everything else surrenders from army losses, since that's been a complaint for a while. In recent titles you might get one or two units held on the capture point for the entire game while everything else died at the entrances. A worthwhile goal, but the problems with implementation are obvious.

    Unlike previous titles, the AI will now reassign engaged units back into the reserves or to another point on the fly rather than wait for a section to quiet down. This is somewhat necessary for defense in depth, since you'd need to pull back from a lost cause you can escape from rather than let all the units there die for nothing if some can be saved. However, as we saw, this opens the door for the AI to reassign units in the middle of a counterattack that might actually work or cancelling it entirely.

    All in all, while some of this might be remedied by fuzzing the states a bit more to force the AI to commit until something really changes, the bigger issue that I can see is that the AI is reevaluating itself way too often. Things move back and forth constantly because the AI keeps changing its mind about where it wants units. This could be because the situation is constantly changing and new data is triggering whatever decision making process it has, but the speed at which it happens is way beyond what any human (either player or in-game) would be doing.



    The siege AI doing very little and letting itself get run over isn't anything new, obviously. We've all been dealing with this since Rome 1. The new part is that it's trying to be more active and sort of half-failing, half-succeeding, which can look even more bizarre. Previously, the settlement defense AI was just kind of lethargic and rather one note, but it's simplicity meant it was just dumb, whereas the new one's problems make it seem like the AI is sabotaging itself, which is more frustrating because it has more good ideas than the old one.

    I will say that I do believe that the problem is manageable in the 3 months they have. I believe this because I can understand what the AI is attempting to do and that it sometimes succeeds, which is more than I can say of some of the issues that the AI has had in the past. That means the system and model themselves seem to be on solid ground, with the problem lying in the variables around its execution, which is a lot more reasonable to correct than trying to redesign pieces entirely. Nobody ever "fixes" AI problems, but they can tone them down to the point where most players will either not notice or not be perturbed by them. Whether CA pulls this off is another matter. They managed to do it with Warhammer for the most part, so here's hoping.

    EDIT: Apologies for my lengthy rambling. I'm trying to organize my thoughts on all this and there's too many unknowns that send me in weird directions without having played the game.

    EDIT2: Went through the footage of the weird AI moves again and I noticed that their aimless fidgeting looked very similar to the way that the one routing unit that bugged out near the beginning because it thought all its exits were blocked was behaving. Perhaps the constant reevaluating and pointless movements were related to this borked pathfinding, i.e. the AI was running back and forth because its units kept thinking there were no ways to reach any of their objectives and kept rerouting themselves in place?
    Last edited by zoner16; February 21, 2019 at 08:19 PM.
    My Three Kingdoms Military History Blog / Military Map Project - https://zirroxas.tumblr.com/
    Ask me a question!

  20. #20

    Default Re: Records Mode Gameplay Reveal

    What the hell! Look at this! Romance Mode is just Records Mode with generals as single entities, duels, and slightly tweaked balancing! For shame CA, for shame. Arch and Darren were right, I'm going home.

    100000 times /s if you didn't catch that.


    Anyway, I will second the whole Shu-Han notion, I get WHY they did it but I still don't like it. I'd prefer if there was a list of historical regions you could pick from like Zhao, Zhou, Wei, Jin, Qin, Shu, Wu, Chu etc. and you could pick a title from any areas you control when you rank up to Duke, or King, or Emperor. Liu Bei would do that and add -Han at the end of that.

    So far my biggest problems are some units having anachronistic weapons and armor and the unit cards, neither of which are big problems, but what is a big problem is the AI which looks pretty bad. Hopefully it can be fixed in the three months they gave themselves.

Page 1 of 2 12 LastLast

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •