Update #44: The Rules of (Melee) Engagement

[Tamerlane]

I gotta say, the book slam thing sounds like it was lifted wholesale from a comedy fantasy game. Totally doesn't fit in with any of the other stuff that I have heard about this game. Just sounds... silly. Ridiculous. "Oh, well, I'm in combat, I'll just close my spell book, thus losing my place, and then bash someone in the face with it, thus destroying the spine. And of course I have to get close enough to do so, which means getting inside that sword-wielder's reach, which is something you ordinarily would want to avoid trying because it's, you know, suicidal. And anyway, it's a huge book weighing five pounds and is made mostly of paper, which makes it an AWFUL weapon (awkwardly shaped, unbalanced, hard to swing, well-padded, etc), so awful that you would be an order of magnitude better off charging someone's fist with magical energy and hitting them with that."

 

I think it's not so much something that a mage gets close to a fighter to do as it is something that a mage does if a fighter gets close.

 

And it's also not so much that they're hitting a fighter with a book as it is that they're hitting fighter with a big pile of magical exploding stuff that happens to be contained within a book.

 

[Tsuga C]

Update by Josh Sawyer, Project Director and Lead Designer

 

I am melee Cornholio!

 

 

Edit: And a very fine update this was, Josh.  Thanks from the peanut gallery!

 

Edited February 27, 2013 by Tsuga C

[Keyrock]

As long as this wouldn't be too complex for the designers, I think the melee engagement zone should change based on whether said melee character is already engaged with another enemy.  For example a melee character not engaged with anyone should be free to spin around in any direction, thus his engagement zone should be a full 360 degrees around them (note the masterful artwork):

 

 

However if said warrior is already engaged with another enemy, then his engagement zone may only cover a cone in front of them, say somewhere between 90 and 120 degrees or so, because turning around to engage an enemy trying to get past them outside that cone would mean giving up their flank or rear to the enemy, which is generally unadvisable:

 

 

Of course there may be skills/spells/items that allow the cone while engaged to expand.

[Tsuga C]

^ 120 degrees sounds about right.  Solid suggestion.

[scolexx]

I must admit that I'm not big fan of attack of opportunity

 

Doing fencing when I was younger and have attended self defense training against knife I've been learned that run away from combat is good and reasonable way to avoid injury. Escape from combat can be achieved quite easily without being hit. Problem is when the attacker try to chase defender. When he catch up he can hit defenders back without able to defend effectively. This mechanism was used in Warhammer fantasy battle.

 

Although I'm not martial art expert and someone more experienced could provide different opinion.  

 

In my pen and paper sessions I try to learn my players that retreat can be good option. Same thing pays for their enemies. There is no reason to get massacred by much more powerful opponent. 

 

I believe that attack of opportunity was introduced to D&D to streamline and speed up combat when character are being hit when they try to retreat. There are also no exp for runaway monsters

 

But I think Engagement mechanism will work fine for the Ethernity and make the combat fun and tactical, although I think its not very realistic  

 

I really recommends you to check the The One ring pen and paper rpg http://www.cubicle7.co.uk/our-games/the-one-ring/. They have really nice, fun and simple system of combat stances. 

 

 

[redneckdevil]

The book thing I see more as someone getting close to a caster, not a caster running into battle armed with a magically enchanted grimoire rdy to slap to slap beasts down. As far as losing place or dmging book, I picture caster basicqlly leaving a finger on their page and the book being magically imbuened to where its not just a regular book but has a magical arua to protect against physical violence or wear and tear.

[n0mDePlume]

n0mDePlume, it sounds like you prefer a different type of micromanagement, not that you don't want to micromanage.  Users on this forum have consistently (though obviously not universally) expressed a general dislike of standard aggro/tanking mechanics.  Engagement is not conceptually much different from threatened area/AoOs, which we already had experience with in NWN2.  It's also a mechanic that many tabletop D&D players are familiar with if they've played 3E, 3.5, and 4E.

 

In IE games, you typically don't move full parties through melee.  You typically move a few characters to ranged positions and a select number of characters into melee with specific targets.  It's true that if you click past melee enemies who are not engaged and run past them, they will engage you.  That's entirely the point of the mechanic.  I certainly understand if not everyone likes it, but it feels more in the spirit of D&D mechanics as well as mechanics that were in NWN and NWN2.

 

 

I know everyone hates tank mechanics, I do too.  They're inherently unrealistic, and when you base your entire gameplay on them like WOW does, combat turns into a simplistic tank-and-spank.  I would never suggest putting hardcore tank mechanics into PE.

 

Tank mechanics are used because they work extremely well with point-and-click real time games.  One reason is that it keeps movement controls simple, you don't have to fight with a click-to-move UI to circle around an enemy or try to "go sticky" in the middle of a group of enemies or something.

 

Another reason is it doesn't limit movement.  The first thing D&D 4 style AOO does is lock down everyone's movement and turn it into a subtle positional game.  This is great for a tabletop game on a small mat... but why would you want to reduce the amount of movement in a real time CRPG?

 

 

Just because people complain about WOW tank-and-spank doesn't mean they'd like a game where they were frequently unable to move where they wanted to for fear of AOO.  Feeling unable to move is a particularly punishing feeling for most players.

 

 

I'd like to clarify I'm not mainly worried about the "bonus attack if your target flees" thing, I'm worried about the "engament triggers against adjacent enemies" thing.  It's based on proximity.  Producing or avoiding proximity is easy in many games, but with a click-to-move system it's something the UI is not designed for.

 

Here's an example of my problem.  I want to get my wizard past an enemy who is in a fight but sticky and still capable of engaging someone extra.  I have the enemy under attack but that doesn't make him safe to walk by.  There is space around his engagement zone, but my wizard still has to circle around him, there's no straight line through it.  How do I do that with the click-to-move system?  Do I micromanage the wizard's movement, pausing every few seconds?

 

And from the point of view of someone trying to lock down enemies, it's far easier to target the ones I want to lock down than to try and micromanage movement so that my guy gets in the way of what are presumably moving enemy targets.  A moving fighter moving into a patrol of moving enemies using click-to-move?  It's what the control system was designed to avoid.

[Ulquiorra]

 

Update by Josh Sawyer, Project Director and Lead Designer

• The wizards' Grimoire Slam allows them to attack an enemy in melee with their magically-charged grimoires, unleashing a concussive wave of energy on contact. If it hits, the attack knocks the target back, usually far enough to break Engagement in the process.

 

I think that "Force push" or "Magick push" sound less nerdy. Book slam in my opinion sound like school nerd atacking some bully ... hahaha.

 

 

BTW.

 

Are grimoires the source of wizards power ? Wizard without this book is nothing ? Ok i understqand that thay have no "REMEMBERING" skill like my self after 1L of vodka ... but seriusly ? i'm 100 lvl wizard and i can't remember 1 spell wthout my book ?

 

Then i supose wariors also can't remember all of their atacks ... what's next ... Warriors phone book ? haha

 

beside that  best update by the way ...

Edited February 27, 2013 by Ulquiorra

[Elerond]

Here's an example of my problem.  I want to get my wizard past an enemy who is in a fight but sticky and still capable of engaging someone extra.  I have the enemy under attack but that doesn't make him safe to walk by.  There is space around his engagement zone, but my wizard still has to circle around him, there's no straight line through it.  How do I do that with the click-to-move system?  Do I micromanage the wizard's movement, pausing every few seconds?

Way-points are probably best and easiest answer such path finding problems.

Edited February 27, 2013 by Elerond

[Giltspur]

This is cool and something I've wanted to see added to "tactical" RPG's as you should have to account for melee characters instead of just cheesing them. 

 

I get the impression from the video (but not from the text) that engagement will stop your character if he was moving from "click to move".  You could then click again to leave the engagement area and be at risk of an AoO (hmm, that acronym makes me think of a wolf howl).  If so that's pretty cool.  I'd like to movement to be interrupted by engagement.  Without that interruption you might just get smacked by someone you thought you wouldn't come in range of.  And the interruption makes sense.  I have to delegate some thinking to my party since I'm not really there on the battlefield in six bodies.  So I'd want them to have the sense to go "Crap, engagement." and stop.  Or auto-attack.  But not just wander willy nilly as some jackal nips his heels with increased bite damage.

 

Does anyone know if that's the plan, for engagement to interrupt click-to-move so that you'd have to make an explicit decision to move out of engagement? 

 

Regardless of the details of default character action upon engagement, it all seems like a natural and interesting evolution of IE-era gameplay.

[Somna]

 

The book thing I see more as someone getting close to a caster, not a caster running into battle armed with a magically enchanted grimoire rdy to slap to slap beasts down. As far as losing place or dmging book, I picture caster basicqlly leaving a finger on their page and the book being magically imbuened to where its not just a regular book but has a magical arua to protect against physical violence or wear and tear.

Pretty much. If you're going to be regularly using a book in combat, you're not going to be bringing in a valuable, fragile object. That would be the equivalent of expecting a warrior to fight in ceremonial gold armor. You're going to be using something that is made to stand the rigors of combat more.

[AGX-17]

 

 

I would like the Grimoire to at least double as a weapon so you can phonebook people for 1 damage, even if just for the comical factor.

The wizards' Grimoire Slam allows them to attack an enemy in melee with their magically-charged grimoires, unleashing a concussive wave of energy on contact. If it hits, the attack knocks the target back, usually far enough to break Engagement in the process.

YEHHHHHHHHHHHHHHHHHHHHHHH

 

 

So now we finally get to RP Klarth in a WRPG.

[J.E. Sawyer]

Another reason is it doesn't limit movement.  The first thing D&D 4 style AOO does is lock down everyone's movement and turn it into a subtle positional game.  This is great for a tabletop game on a small mat... but why would you want to reduce the amount of movement in a real time CRPG?

 

Here's an example of my problem.  I want to get my wizard past an enemy who is in a fight but sticky and still capable of engaging someone extra.  I have the enemy under attack but that doesn't make him safe to walk by.  There is space around his engagement zone, but my wizard still has to circle around him, there's no straight line through it.  How do I do that with the click-to-move system?  Do I micromanage the wizard's movement, pausing every few seconds?

 

And from the point of view of someone trying to lock down enemies, it's far easier to target the ones I want to lock down than to try and micromanage movement so that my guy gets in the way of what are presumably moving enemy targets.  A moving fighter moving into a patrol of moving enemies using click-to-move?  It's what the control system was designed to avoid.

 

I don't agree that 4E really locks down everyone's movement.  I played a warden for a few years and even he couldn't really lock down movement for more than a few characters at a time and that's literally what I built him to do.  I agree that 4E does make movement and positioning more important, and that is ultimately the goal.  We are implementing our Engagement system with the understanding that we may need to modify elements of it to find the right balance of tactical movement requirements and freedom of movement in our combat environments.

 

In the example of the wizard, I don't believe that's any different than dealing with a persistent hazard, something that comes up with great frequency in the IE games.  I do think it will probably require more than one click to move around such a hazard, but I don't believe it will devolve into endless multi-clicking.

 

In the last example, melee enemies moving through melee enemies, moving characters will not trigger Engagement.  If two opposed melee fighters rush by each other and they're not stopping to attack, they won't engage the other.  Also, if a fighter is not in his or her Defensive mode, he is likely to only be capable of engaging a single target at a time (his or her current melee target).  If some of those melee enemies are targeting someone rushing past, they will stop to attack and that will result in Engagement, but that's a typical and intentional result.

[Sannom]

Josh may have already answered this, but I will ask anyway : what happens to a fighter who tries to go help another fighter locked in an engagement with an enemy fighter? Does the enemy fighter gets a 'disengagement' attack as the other fighter gets close enough, or did I completely miss something here?

Here's an example of my problem. I want to get my wizard past an enemy who is in a fight but sticky and still capable of engaging someone extra. I have the enemy under attack but that doesn't make him safe to walk by.  There is space around his engagement zone, but my wizard still has to circle around him, there's no straight line through it.  How do I do that with the click-to-move system?  Do I micromanage the wizard's movement, pausing every few seconds?

I want to say 'Duh'. That's how firing breath-like spells work too, like in NWN2 or Dragon Age, you have to move the characters to be in a very precise position to maximize the effectiveness of the spell. Depending on the size of the map or the room the fight is taking place in, going around the big guys beating each other with clubs shouldn't take more than three successive movements.

[Lephys]

Keyrock totally beat me to it (with number 1 below), but I just want to emphasize 2 things:

 

1) Cones of engagement (once a character is already engaged with at least one target), with maybe feats for a full 360 degrees of engagement (so only melee fighters specialized in peripheral engagement could possibly engage in a full 360 degrees, and maybe their maximum number of targets-of-engagement would be limited to 2, whereas the cone-people could still hold up to 4 or 5 (Barbarians, perhaps?) in up to a 180-degree (or maybe a little smaller) cone in front of them? *shrug*

 

2) Click-and-drag path "drawing" for movement commands. OR, as an alternative, what if you could tether movement waypoints to enemies? Maybe click and drag a certain distance from that enemy, and that's how far away from that enemy your character will try to stay whilst moving as close to the initial waypoint (Where you released the mouse click-and-drag) as possible? It's like a restraining order for RPGs, 8D.

 

Just some thoughts.

[n0mDePlume]

I don't agree that 4E really locks down everyone's movement.  I played a warden for a few years and even he couldn't really lock down movement for more than a few characters at a time and that's literally what I built him to do.  I agree that 4E does make movement and positioning more important, and that is ultimately the goal.  We are implementing our Engagement system with the understanding that we may need to modify elements of it to find the right balance of tactical movement requirements and freedom of movement in our combat environments.

 

In the example of the wizard, I don't believe that's any different than dealing with a persistent hazard, something that comes up with great frequency in the IE games.  I do think it will probably require more than one click to move around such a hazard, but I don't believe it will devolve into endless multi-clicking.

 

In the last example, melee enemies moving through melee enemies, moving characters will not trigger Engagement.  If two opposed melee fighters rush by each other and they're not stopping to attack, they won't engage the other.  Also, if a fighter is not in his or her Defensive mode, he is likely to only be capable of engaging a single target at a time (his or her current melee target).  If some of those melee enemies are targeting someone rushing past, they will stop to attack and that will result in Engagement, but that's a typical and intentional result.

 

 

I'm surprised you're actually using 4E as an example of something you want to emulate in PE.  I would have hoped you'd say "don't worry, 4E is a rules heavy, turn based game which many people consider very slow, and we don't want our engagement mechanic to produce that kind of game".

 

You're saying that making movement and positioning more important is ultimately the goal, which sounds like a design decision bigger than just engagement.  Can you understand my concern about trying to do this with a click-to-move CRPG?

 

Party movement and pathfinding AI has traditionally been one of the weakest spots of these games, showing the limits of the engines and generating a lot of complaints.  This is a genre where NPCs can frequently teleport because pathfinding is so easily broken, and where post-IE games avoided full party control because it was harder to deal with.  I remember times - many times - when I tried to move a character halfway across the screen in combat and they decided to take a detour halfway around the level because a moving enemy temporarily blocked their path.

 

 

Having engagement as a specific feature is one thing, and can obviously be debated.  But if you want to emphasize movement and position in general, do you plan any engine/UI improvements to go along with that?

 

Someone gave waypoints as an example.  I'd hate to have game mechanics that required waypoints, but they are the classic way of allowing multiple units to execute complicated moves without having to madly scroll back and forth and babysit all of them at the same time.

[Lephys]

...

 

Party movement and pathfinding AI has traditionally been one of the weakest spots of these games...

Perhaps it's time to break tradition, rather than breaking complexity for tradition's sake?

[some guy]

 

Here's an example of my problem. I want to get my wizard past an enemy who is in a fight but sticky and still capable of engaging someone extra. I have the enemy under attack but that doesn't make him safe to walk by. There is space around his engagement zone, but my wizard still has to circle around him, there's no straight line through it. How do I do that with the click-to-move system? Do I micromanage the wizard's movement, pausing every few seconds?

Way-points are probably best and easiest answer such path finding problems.

Or an "avoid engagement zone" party-AI toggle combined with holding an "show future path" key... or some such.

 

I see were the concerns lays on how messy this could be if enemies and party-members gets randomly spread, and how hectic enemy AI will be on disarranging player formations. For me though, this is an interesting challenge in that it put focus on prevent losing a stable formation while disarranging your enemies'. Part of the reason why a lot of battles in most games becomes a mosh pit is, I believe, the sheer lack of restrains for moving around. Without a firm, basic way of creating a manageable blockade it's just bound to become a mess.

Edited February 28, 2013 by some guy

[Lephys]

Also, what if, in addition to key, active-use "disengage without penalty" abilities, each character had a passive disengagement ability attached to their standard attack cycle? For example, once a Fighter becomes engaged with an enemy, his 3rd standard attack will allow him to disengage without penalty (with some kind of visual mini-stun animation on the opponent, and/or other visual indicators of this). However, a Wizard would only be able to manage some sort of disengagement maneuver in standard combat attacks after, oh... 6 attacks?

 

Maybe that's too complex for the resources/timeframe they're working with, but I think it could add in a nice buffer layer. Basically, the more adept a character is at melee combat, the more often they should be capable of voluntarily disengaging from an enemy. I think it would probably only work with single-engagements, though. Only active abilities (like the Rogue's Escape ability) should disengage you from more than one eopponent without penalty.

 

This is akin to a "critical defense," a bit. So, maybe they could be one in the same? Maybe if your enemy misses, you get an opportunity to disengage? I dunno... like I said, it seems like melee skill should contribute somehow. I mean, the better I am at melee combat, the more easily I could strike at you/your weapon in such a way as to get away to handle a target of more strategic interest rather than dealing with you at the moment.

 

Of course, giving Fighters 5 disengagement moves and Wizards only 1 would accomplish the same representation of melee skill affecting disengagement. *shrug*. So maybe there's absolutely no need for an extra, passive layer.

[anek]

I'm surprised you're actually using 4E as an example of something you want to emulate in PE.  I would have hoped you'd say "don't worry, 4E is a rules heavy, turn based game which many people consider very slow, and we don't want our engagement mechanic to produce that kind of game".

 

 

Don't forget that PE will be a computer game, not pen & paper.

 

Things that might have been too "rules heavy" and "slow" for P&P, might very well work great when the computer takes care of evaluating all the rules for a particular action in a fraction of a second.

[Gumbercules]

I enjoyed all the silliness in the video for this update, especially the hawk screech.

 

I like the idea of the engagement system in theory, although I share some of the concerns that other people have already voiced. I really hope that the pathfinding and basic AI is VASTLY improved from what we're used to with these types of games, or else things will get very fiddly and frustrating in a hurry. If it is improved however, I can see this system working quite nicely.

 

Party members should probably know by default, whenever you send them somewhere, to avoid enemies' engagement zones unless absolutely necessary. Clicking while holding down some assigned key should override this and send them in a straight line instead, and there should be a waypoint option as well. For enemy AI, it can depend on the enemy type. Large, mindless hordes could ignore engagement zones and get tied up easily by fighters, while smaller groups of elite enemies could have the same avoidant behavior as party members and try to circumvent.

[Somna]

The more I think about that Grimoire Slam ability, the more it feels like there might be a cooldown, if indirectly.

 

The ability and description pretty much imply that the book is going to be closed, physically swung at someone, and cause an explosion/shockwave to knock the target back.

 

What's going to shift that belief about a cooldown around is the following thought:

 

Is the "closed" state of the Grimoire going to have any effect on normal usage of the Grimoire?

 

For example, there's a couple different thoughts on possible in-game reasons for a Grimoire swapping cooldown.

 

1) The book needs to be open in order to awaken and use the magic stored in the book. The bulk of the cooldown is waking up the magic.

- This would make the Grimoire Slam closing the book result in a cooldown.

 

2) You need to re-create triggers for the active Grimoire to access the spells you need each time you swap out a Grimoire. You can only have one set of triggers ready at a time, so taking out another Grimoire means setting up another set of triggers, but opening and closing the book won't change the state of the triggers.

- This would make the Grimoire Slam closing the book irrelevant. You just wiggle your finger (or whatever other triggers are actually set up) and keep on using magic, kind of like an arcane puppetmaster feel.

 

I'm sure there's plenty of other ways it can work, but the question would still boil down to whether closing that book means anything.

Edited February 28, 2013 by Somna

[Eleneithel]

1) I also find rather... inadequate the Book Slam Technique of the wizards. I do not find logical or realistic for a wizard to wield (and read) a book in combat, even if he fights away from the melee, because of his need to remain aware of his surroundings and of his ennemies in an chaotic situation. An adventurer wizard would more likely hold a staff or a melee weapon (like a sword) when casting at a range and use it in close quarters to defend himself. Why? Because an hostile fighter can engage him in a matter of seconds. So, it would be more coherent for a wizard to use an melee weapon imbued with his own power to compensate his vulnerability, but maybe this point of view conflict with the magic system... I do not know.

 

Concerning the Defense ability of the fighter, I think the number of ennemies engaged should be a matter of skill and experience. By experience, I can tell that it is very difficult to engage two ennemies at the same time, even when they aren't trained to fight efficiently together. Thus, I would advise that a novice fighter be able to engage only two ennemies when defending, but he should then be able to raise that limitation when leveling and "training". Otherwise, I like the principle of this rule of engagement.

 

2) Concerning fighting skills and feature, I always found D&D video games a bit irrealistic and I would propose another vision for the fighting skills system. For example, take a fighter like a swordman in real life. A novice fighter IRL would learn the basics, which are fencing movements, fundamental stances (which serve equally as offensive and defensive moves), and some desarming, grappling, and projections techniques amongst the most easiest and basics fencing techniques. This, I would qualify as the minimal training to have to survive a swordfight against another opponent. In D&D games (especialy NWN), I always thought that level 1 fighters were seen almost as beginners: they can swing a sword, they can move, but they haven't really received the minimal training of the novice fighter and you must choose wether he can disarm, knock down, or else... which makes combat with low level characters rather uninteresting.

 

So, my suggestion for a level 1 fighter, for example, would be to give him some basic abilities such as disarm, knock on the ground, shield bashing /pommeling, basic dodge... and a limited selection of weapons he has trained with (a primary melee weapon like the broadsword or the battleaxe, a secondary like a dagger or a hatchet, a ranged weapon like the bow or another melee weapon, and eventually basic unarmed combat). When leveling, the fighter can train to improve these basics abilities (better disarm, for example) or to gain new abilities (wielding a new weapon, dual-wield, ens...). I think this make more sense than the D&D way of having a level 1 fighter knowing how to wield much different weapon, but knowing few or none basic fighting techniques. The logical progression for a fighter is to learn the basics of a few weapons and, later eventualy, specialise in particulars weapons or learn other ones. Of course, such fighting system would be transposed to the other classes (fighting or non-fighting), with variations. For example, a ranger /archer wouldn't learn melee techniques as disarming, but learn rapid firing and precise shooting instead. A rogue classe would trade certain frontal fighting techniques for more disloyal and sneaky ones to be able to inflict swift attacks. A wizard would know less agressive fighting techniques and concentrate on defensive ones to be able to survive and retreat behind his allies. The amount of basic abilities retained by each classe would depend on their amount of martial training (high, medium, or low).

 

I think such system would be more interesting and more realistic on a tactical point of view than the D&D system. Most characters do not need to be able to wield a large selection of weapons as the player often specialises them in a particular weapon for tactical reasons or personal tastes. Especialy if the player can choose the weapons of his character during his creation. (The NPC would more likely wield weapons accordingly to their culture and preferences)  What do you think?

 

 

PS: I am not used to speak /write English (as French is my native language), so I hope I am understandable...^^

Edited February 28, 2013 by Eleneithel

[rjshae]

It sounds like a nice mechanic; somewhat like the old wargaming "zone of control" mechanism. It will make combat much more tactically interesting. However, a few questions occurred to me:

• If a defender is engaged, shouldn't they be able to perform a slow fighting withdrawal, at a defensive bonus, simply by gradually pulling back while keeping their weapon and shield at the ready? I think D&D solved this by allowing a single square of movement without incurring an opportunity attack.
• Suppose an enemy has a reach weapon, such as a swiss pike. Will the radius of engagement be correspondingly enlarged?
• If the enemy has a reach weapon and more than one engagement slot, will he be able to engage an adjacent opponent and one further away?
• If you move into the engagement radius of more than one opponent, do they all engage you and thus collectively decrease their available engagement slots? This seems like a way for a party to pin down opponents, allowing others to slip past.

Thank you for the update!

 

Silence. Okay... I did a little thinking and here is what seems to make sense:

• I'd like to see a 'retreat' mode that puts you in a defensive stance and causes you to retreat out of the engagement radius following each melee attack. If the enemy doesn't follow up the attack and there are no other engagements, then you become disengaged. If you are not attacked during the "round", then you automatically back away from the engagements (if possible) without penalty.
• For a reach weapon, it would be logical for engagement to be prioritized by proximity. You want to threaten the most threatening foe, which is usually the nearest.
• The nearest enemy contacted (with a free engagement slot) incurs the automatic engagement. Enemy AI should then check for an existing engagement before automatically engaging an opponent. That leaves engagement slots open for other opponents.

[Sensuki]

Party movement and pathfinding AI has traditionally been one of the weakest spots of these games, showing the limits of the engines and generating a lot of complaints. This is a genre where NPCs can frequently teleport because pathfinding is so easily broken, and where post-IE games avoided full party control because it was harder to deal with. I remember times - many times - when I tried to move a character halfway across the screen in combat and they decided to take a detour halfway around the level because a moving enemy temporarily blocked their path.

Pathfinding has come a long way since Baldur's Gate 1. Dota 2 is probably a good example, and to avoid your problem there are a few things you can do to avoid that, such as taking units out of the equation when figuring out the most optimal path.