07 Dec, 2009, Runter wrote in the 1st comment:
Votes: 0
So. I'm developing a threat system to determine who is "tanking." It's not very unique considering how many popular places this type of system has been used. What I'm interested in doing perhaps a little differently is using threat to generate a likelihood of being attacked. Instead of dealing in absolutes. Obviously my game design wouldn't involve requiring a tank receive every blow from boss encounters.

A quick example of my current idea:

Tank has 50 "threat".
dps A has 20.
dps B has 20.
healer has 10.

I picked these numbers because they're easy to understand. 50 in this case is 50% of the total, 10 is 10% of the total, etc. So in my idea when the boss has to decide who to target with a swing or ability there would be a 50% chance to pick the tank, 20% chance each for the DPS to be selected, and a 10% chance for the healer to be selected. Needless to say those numbers could have been anything—but the percent of the total is what is actually important.

What does this really mean on gameplay? As I've already said I will be games damage so that it is manageable for anyone to be hit. Obviously, though, you'd want anyone who isn't supposed to fill tank role to be lower threat. This makes the game a little more interesting for designing your character since you may want to use weaker abilities that per point of damage/heal do perhaps less threat. Or you may decide to take other options to lower threat. Or perhaps you would toss in attacks to lower threat.

If anyone has any ideas about threat systems by all means share. :)

Edit: Yeah, reading over it I kinda skipped explaining what most threat systems do. Kinda expecting the reader already knew. If you didn't most threat systems work by players generating internal numbers of threat based on their damage deal, amount healed, or tank abilities to raise threat. Usually in these systems the intent is to make it so that a tank role is viable without them having to lead in damage. Generally in these systems they make it absolute. For example, if the tank has more threat he receives *every* hit. There's no chance for anyone else to receive selection. That's the key change I'm currently looking at making.
07 Dec, 2009, Mudder wrote in the 2nd comment:
Votes: 0
You could make EQ affect threat. Someone decked out in all hardcore evil looking armor might be more intimidating than a halfling wearing cloth and leather bracers. Thus the dude in evil armor would seem to be a bigger threat.
07 Dec, 2009, Runter wrote in the 3rd comment:
Votes: 0
Mudder said:
You could make EQ affect threat. Someone decked out in all hardcore evil looking armor might be more intimidating than a halfling wearing cloth and leather bracers. Thus the dude in evil armor would seem to be a bigger threat.


Yes, I do see where you are coming from.

Just keep in mind that threat doesn't have to equal intimidation. I could also use a word like "tactical priority assessment points." The mob may feel the little healer constantly full healing his party is the highest priory to kill. (Or highest threat to his living.) :P
07 Dec, 2009, Mudder wrote in the 4th comment:
Votes: 0
Yes, it definitely could. If you adjust a target's threat based on the amount of damage it causes the mob, you could then increase the threat of a character that heals another PC by a similar amount.

In other words…

Bob the warrior hits a goblin rogue for 100 hp. - Bob's threat is increased by 10%
John the healer heals Bob for 100 hp. - John's threat rate could increase by a similar, albeit less amount.

I still think looks play a large part in how people/things react to others. While they may not all react the same - I think it is fair to say that there will be certain "looks" that everyone will read to assume that the target will be the biggest threat.

An example:
This guy: http://www.starstore.com/acatalog/Sauron...
Will be seen by almost everyone to be more of a threat or a bigger priority
than this guy: http://www.conversationmarketing.com/bil...

Now whether or not you want your system to figure out which would be the easiest kill and dispatch them first, is up to you. But I think it fits the theme to take looks into consideration. It would also add another element to combat. People won't simply wear eq because it is the "best" but because it makes them look less scary or whatever. It could even assist/hurt haggling with shopkeepers. :P


EDIT: Also, you could have different types of mobs act differently in these scenarios. Simple beasts would likely just attack the nearest aggressive target, so would remain focused on the enemies using melee hits. While a smarter rogue type mob would try to finish off the weakest opponent before fleeing.
07 Dec, 2009, Lyanic wrote in the 5th comment:
Votes: 0
I like the system I've seen in Final Fantasy games where there are multiple tactics that NPCs use for assessing threat. One might be to go for the party member that looks most intimidating - the archetypal "tank". One might be to go for the healer to decrease overall survivability of your party. One might be to go for the party member dealing the most damage to it. It would be cool to set it up such that you can assign these different tactics to NPCs appropriately. For instance, a dumb NPC like an Ogre might attack the heavily armored tank, whilst allowing a row of Magi in the back to burn him to a crisp, whereas a Demon would go rip the Magi's hearts out before coming back around to the tank.
07 Dec, 2009, Runter wrote in the 6th comment:
Votes: 0
Lyanic said:
I like the system I've seen in Final Fantasy games where there are multiple tactics that NPCs use for assessing threat. One might be to go for the party member that looks most intimidating - the archetypal "tank". One might be to go for the healer to decrease overall survivability of your party. One might be to go for the party member dealing the most damage to it. It would be cool to set it up such that you can assign these different tactics to NPCs appropriately. For instance, a dumb NPC like an Ogre might attack the heavily armored tank, whilst allowing a row of Magi in the back to burn him to a crisp, whereas a Demon would go rip the Magi's hearts out before coming back around to the tank.


It would be easy enough to give different mobiles different threat assessment algorithms. However, I'm curious if that would be a good thing considering it then becomes difficult for players to know how threat is calculated.
07 Dec, 2009, Mudder wrote in the 7th comment:
Votes: 0
Should players know exactly how it is calculated? If you're going for anything that you want to seem realistic, unpredictability is your best friend in that aspect.

There should be some faily intuitive constants to provide a similar framework but let other things vary.
07 Dec, 2009, Runter wrote in the 8th comment:
Votes: 0
Mudder said:
Should players know exactly how it is calculated? If you're going for anything that you want to seem realistic, unpredictability is your best friend in that aspect.

There should be some faily intuitive constants to provide a similar framework but let other things vary.


I'd rather have actual numbers that people can study to know why they're constantly being the person in the party getting smashed.
07 Dec, 2009, KaVir wrote in the 9th comment:
Votes: 0
Runter said:
I'm developing a threat system to determine who is "tanking."

As this is something that's clearly useful to the players, have you considered simply allowing them to specify their own threat levels? You could allow them to set their position in the group, eg "front line" gets hit the most, "rear" gets hit the least (but perhaps can only attack with reach/ranged weapons), "flank" is able to ignore shield bonuses, and so on.
07 Dec, 2009, Runter wrote in the 10th comment:
Votes: 0
KaVir said:
Runter said:
I'm developing a threat system to determine who is "tanking."

As this is something that's clearly useful to the players, have you considered simply allowing them to specify their own threat levels? You could allow them to set their position in the group, eg "front line" gets hit the most, "rear" gets hit the least (but perhaps can only attack with reach/ranged weapons), "flank" is able to ignore shield bonuses, and so on.


I've done that before in "formation" code. While it worked in the design, I'm trying something different this time around. A tank not on his toes in my proposed system may could let his threat sink too low—there's a lot of dynamics and strategy around quantified threat that I find interesting.
07 Dec, 2009, Runter wrote in the 11th comment:
Votes: 0
Mudder said:
Should players know exactly how it is calculated? If you're going for anything that you want to seem realistic, unpredictability is your best friend in that aspect.

There should be some faily intuitive constants to provide a similar framework but let other things vary.


Thinking about the subject a little more, I would be more likely inclined to have mobiles that perhaps do not follow the standard formula for threat. But a large majority of my mobiles likely will. Specifically there may be encounters where the mechanics of the fight over-ride the general threat mechanics.
07 Dec, 2009, Lyanic wrote in the 12th comment:
Votes: 0
Runter said:
Lyanic said:
I like the system I've seen in Final Fantasy games where there are multiple tactics that NPCs use for assessing threat. One might be to go for the party member that looks most intimidating - the archetypal "tank". One might be to go for the healer to decrease overall survivability of your party. One might be to go for the party member dealing the most damage to it. It would be cool to set it up such that you can assign these different tactics to NPCs appropriately. For instance, a dumb NPC like an Ogre might attack the heavily armored tank, whilst allowing a row of Magi in the back to burn him to a crisp, whereas a Demon would go rip the Magi's hearts out before coming back around to the tank.


It would be easy enough to give different mobiles different threat assessment algorithms. However, I'm curious if that would be a good thing considering it then becomes difficult for players to know how threat is calculated.

Mudder said:
Should players know exactly how it is calculated? If you're going for anything that you want to seem realistic, unpredictability is your best friend in that aspect.

There should be some faily intuitive constants to provide a similar framework but let other things vary.


I have to agree with Mudder on this one. I'm not really in favor of using completely static behaviors that enable players to learn the inner-workings of the system, thus successfully applying the same strategy in every encounter ad nauseam. It might be a good idea to limit the number of threat assessment algorithms to a certain number, though - perhaps 6-8? Then, players just have to be careful and observant the first time they face a new NPC - learn its strategy, and remember it for the next time.
07 Dec, 2009, Runter wrote in the 13th comment:
Votes: 0
Quote
I have to agree with Mudder on this one. I'm not really in favor of using completely static behaviors that enable players to learn the inner-workings of the system, thus successfully applying the same strategy in every encounter ad nauseam. It might be a good idea to limit the number of threat assessment algorithms to a certain number, though - perhaps 6-8? Then, players just have to be careful and observant the first time they face a new NPC - learn its strategy, and remember it for the next time.


The threat system is hardly the strategy for a fight. It's once piece. But I intend to make it an important piece with skill/spells/character customization choices to be made involving the threat system. People will want to know they get assigned threat values. They should be able to know fairly consistently why they are pulling threat. At least this is my opinion, and my game will be this way.


And in my opinion if it's random it's as good as no system. Maybe even worse since depending on the fight the same actors will always get targeted. And they won't even know why it's happening.


Also, this particularly goes to how I plan to separate design of boss encounters from trash encounters. Boss encounters *will* feature unique mechanics to the threat system per encounter.
07 Dec, 2009, jurdendurden wrote in the 14th comment:
Votes: 0
Perhaps instead of every player having a percentage of threat, you could base it purely off of the damage they do/have done, and just reset it after each battle.

Example: (just throwing out a formula like 1 threat per 2 damage points for sake of simplicity.)

Round 1:
Runter the tank gets 4 hits in this battle doing 20 points of damage per hit. His threat is now 40.
Mudder the black wizard casts fireball, doing 50 points of damage, giving him a threat of 25.

Round 2:
Runter misses twice this round, still doing 20 dmg per hit, bringing his total threat to 60.
Mudder casts acid blast this time, the mob is vuln to this and takes 90 damage, giving Mudder an additional 45 threat, bringing him up to 70 total. He has now successfully pulled aggro.

Round 3:
Runter rescues Mudder, does some damage, repulls aggro, and Mudder doesn't cast this round to allow Runter to build up a bit of threat.

Etc..etc..

From there you could possibly make spells/attacks that target vulns add 1 threat per 1.5 damage, spells/attacks that target resistances 1 threat per 3-4 damage, and so forth, in order to make things more interesting.
07 Dec, 2009, KaVir wrote in the 15th comment:
Votes: 0
jurdendurden said:
Perhaps instead of every player having a percentage of threat, you could base it purely off of the damage they do/have done, and just reset it after each battle.

The problem then is that you lose the concept of a tank. Usually the tank is capable of withstanding lots of damage, while others focus of inflicting damage - but if you were hit based on how much damage you'd inflicted, the tank would end up being mostly ignored while all the backstabbing thieves and fireballing mages were torn apart.
07 Dec, 2009, jurdendurden wrote in the 16th comment:
Votes: 0
Of course, this is why you wait to cast that fireball for a round or so, and allow the tank to 'taunt' (hate to pull that straight from WoW) or something to add to threat. Or maybe even better, allow paladins/fighters/rangers to accrue threat at a faster rate. Or maybe you could even exclude backstabbing from threat, since it auto-draws aggro anyhow. Any number of ways around that. :)

Edit: backstab auto-draws aggro if not already engaged in combat. So you could always wait to backstab until the tank starts tanking.
07 Dec, 2009, KaVir wrote in the 17th comment:
Votes: 0
jurdendurden said:
Of course, this is why you wait to cast that fireball for a round or so, and allow the tank to 'taunt' (hate to pull that straight from WoW) or something to add to threat. Or maybe even better, allow paladins/fighters/rangers to accrue threat at a faster rate. Or maybe you could even exclude backstabbing from threat, since it auto-draws aggro anyhow. Any number of ways around that. :)

Sure, but I was responding to your suggestion that "instead of every player having a percentage of threat, you could base it purely off of the damage they do/have done" (emphasis mine). Your "ways around that" seem to involve not basing it purely off of the damage they do/have done :P
07 Dec, 2009, shasarak wrote in the 18th comment:
Votes: 0
KaVir said:
The problem then is that you lose the concept of a tank. Usually the tank is capable of withstanding lots of damage, while others focus of inflicting damage - but if you were hit based on how much damage you'd inflicted, the tank would end up being mostly ignored while all the backstabbing thieves and fireballing mages were torn apart.

It seems to me that tanking and damage avoidance should be the responsibility of the player characters rather than the responsibility of the NPC doing the attacking. If a spell-caster needs to stay out of range of an NPC's attacks, then the spell-casting character should have skills (either character-based or player-based or both) which aid him in staying clear of the melee (to take a Diku-like example, a thief's hide and sneak skills, or a mage's invisibility spell); and a tank should have skills that help distract the attention of the enemy, or aid in manoeuvering in such a way as to keep the tank physically interposed between the monster and the spell-caster. If the spell-caster's combat-avoidance skills and the tank's combat-redirection and interposition skills aren't sufficient (because the NPC has been practising his foot-work) then the spell-caster shouldget hit.

Any intelligent NPC should, if only for reasons of realism, attack whichever opponent he regards as the greatest threat; and that determination should be dynamic and based on observation - if someone who is apparently not a threat unexpectedly hits you for major damage, then the logical thing to do is to direct your attacks towards them, or at least to try to - it should then be up to the targets to prevent that from happening if they can.
07 Dec, 2009, jurdendurden wrote in the 19th comment:
Votes: 0
Quote
Sure, but I was responding to your suggestion that "instead of every player having a percentage of threat, you could base it purely off of the damage they do/have done" (emphasis mine). Your "ways around that" seem to involve not basing it purely off of the damage they do/have done :P


Well disregard the word purely ;) I should have worded that differently but I simply meant that instead of having straight percentages (50% tank, 20% mage), just use a building point system that resets each time combat is ended.
07 Dec, 2009, quixadhal wrote in the 20th comment:
Votes: 0
As much as people may groan at this… look at the balance that World of Warcraft has with respect to threat/aggro.

A "tank" has various abilities and stances which are designed specifically to generate threat. The infamous "taunt" skill does no damage, but it infuriates the target, making it very likely it will attack the tank for a time.

A DPS class generates threat by doing damage, and THEY sometimes have abilities to reduce the amount of threat they are generating. A hunter gets misdirection to make the next few shots build up aggro on misdirection's target, and feign death to clear their aggro table temporarily.

A healing class generates threat by healing and casting buffs. For them, certain skills lower the aggro generated by these actions. Using a heal-over-time ability typically generates less threat than a big flash heal, but it also takes longer.

So, while the choice of initial target (if the mob is the one starting combat) may be based on perceived threat or perceived weakness, once the party engages, their actions quickly overwhelm this choice and the mob will switch targets as it deems one is more or less troublesome. It is worth reading up on some of the strategy guides for the end-game instance dungeons… you may not like the game, but the strategies the players have to figure out to complete the encounters is well worth examining.
0.0/36