But I don’t understand something…
You are saying when the adversary acticates, rolls for detect a hidden model (obscured/clear, the nearest), because he “targets” at the nearest one, taking that model into account for the next 2 actions (even if he does not have any possible action to actually target the hidden enemy).
If it succeeds, removes the token and performs his 2 actions as usual.
But if he fails that roll, he would need to look for the 2nd nearest model to move or do something, according to you. Then, a 2nd hidden model would be targeted from the priority list, so they would try again to remove the token… until he removes a token or fail all the INT test (probably when they have INT 1 
), in wich case we reach the same case as before… what do they do then? Moving to where? Does he move now to the nearest hidden model without removing any token??
At this point, I understand the @Modiphius-Gavin 's position, I think he should “go to the nearest” (regardless of their position or tokens), but not “go to the nearest in LoS” (In FWW, for example, nearest does not need any LoS, unless specified). If you don’t need LoS, you don’t need to roll for detecting a model, he moves to the hidden model anyway, because he is the nearest, even without LoS or anything
Blockquote But according to @Modiphius-Dom , they don’t take into account the hidden models with blocked LoS when you look at the target priority list on the faction card, then it could bring an odd situation against the Dark Brotherhood in the 1st round, so you would need to know where they go in that case. But according to you, they would move to the nearest, for example, without rolling, wich seems reasonable for game purposes (not from a RPG point of view, they go straight to the hidden models at 30"), and with that info, all this problem would be solved, because you would not need to roll for detect them when moving, just when engaging to consider them engaged
@Modiphius-Dom is correct. The target priority only works if they have a chance to see someone. If all your DB models are all hidden and there isn’t any chance of seeing anyone then there’s no valid targets anyway. That’s the point when the head towards the enemy. It’s very noticable in the Hold The Gates scenario where there is Blocking Terrain over half the table - the wall. If the AI models don’t have an assumption that they know where the enemy is, they can just lurk around on the far side of the wall on an A response, as they can’t see the player’s models. There’s a clause in the Ranged Attack Branch of the Attack action that says if there is no LoS to an enemy, the model will move to gain LoS. This is an example of the implicit ability of AI models to know roughly where player models are - they want to get within that 6" where they can try and make a INT check to detect Hidden models.
The AI system isn’t perfect. There will be times when you have to make a judgement call because we can’t foresee every eventuality. As ever, you should make a choice based on whatever makes narrative sense and will move your story along in the most satisfying way. We also know that the Adversaries chapter needs work - it’s a big focus of the rules revision - so this feedback is really helpful.
Why would the 6" matter except for getting the green die bonus on the INT test (and for reactions)?
Maybe now is a good moment to put a minimun range on the detection roll, that would solve some situations too, like getting into 10" of the adversary to test your Sneak or they test their INT
That would probably make the issue come up less frequently but, for me, it wouldn’t actually solve it (or at least it wouldn’t make it any clearer).
Is this missing the rule about rolling with any responses? Or does it mean they roll just with the A response?
I will try to do an overview later explaining everything step by step, trying to understand this, and please @Modiphius-Dom let me know what is right and what is wrong, just to play fairly/balanced against the AI
Because within 6" you can’t hide from them, or re-hide once spotted, and the enemy get the Green dice making spotting more likely.
Let’s see the steps in Delve Mode:
- I begin my game with 4 Dark Brotherhood models sneaking because they have LoS Blocked and they are at 30"
- The AI is Draugr faction with 4 models (skeleton archers and common draugr)
- I have priority so I move a model leaving it with Clear LoS
- Now a common Draugr activates and I roll for the matrix, rolling M response
- At this point I should roll an INT test because the common draugr targets the hidden model I moved because it is the nearest
- If that draugr fails the that first roll it continues trying to detect every single hidden model (next nearest) wich has obscured/clear LoS
6.1. After failing all the possible tests the draugr moves anyway towards the nearest hidden model (the one I moved) and blocks anyway - If that draugr fails the first roll and rolls for the second hidden model (because it is the 2nd nearest model and it has Obscured LoS)
and it succeeds it moves towards that second nearest model and blocks
This is my understanding after reading @Modiphius-Dom 's last post. Are this steps right?
The important thing here is that the AI has always the chance for detecting you with every single model even at 30" and with the M/A response.
That seems pretty odd because the rulebook says they detect models exactly the same way as player models, but they get the chance when they “search for a target to Attack”, so does it mean the M is considered “searching for a target to Attack”? Or “searching for a
target to Attack” means the search of the target following the target priority list?
In the case of a player’s model in the same situation he could not try to detect my models until they are at range to attack them by a ranged weapon, so it does not matter any distance or move action in that case, they cannot roll until they announce a movement to engage me or an attack, that would be the only way they could roll to detect me, wouldn’t?
Ah that’s ok then; the way it was worded I thought you meant the adversaries had to be within 6" to test at all.
Hey @DK-dark
Your step by step is correct.
There are no differences between how AI and Player Models Detect Enemies Models, other than AI Models following the Response Matrixes when Player Models get an active choice in how they act on the tabletop.
As per ‘Detecting Hidden Models’ on page 27, Detection only happens when the hiding model is about to become a target or is going to have an enemy move into base contact with it. So if you theoretically had an attack with thirty inch range or could move the same distance in a single action to get you into base contact with an enemy thirty inches away, yes you could attempt a Detection. But that isn’t normally going to be the case - ranges in the game tend to be a lot shorter than that.
I think the key issue (exclusive to the AI) is the distinction between a target for a response and the target for an attack/action and it’s still not entirely clear how that operates; the wording of becoming ‘a target’ is too vague in that sense as there’s an inherent difference.
In @DK-dark’s example above, (step 4./5.) the Draugr is rolling to detect the Hidden model because it’s the nearest as a result of the Response, even though it doesn’t have a means of actually targeting it with an action or an attack. I’m not sure that is correct (and your last response seems to imply that it isn’t, despite also saying the step example is likewise correct).
In order to have that consistency with player models, I think the intent is more:
- An Adversary will target an enemy for a Response even if that target is hidden.
- An Adversary will however only roll to detect the hidden model if, as part of its Response, it would target the hidden model with an action or attack.
- If the detection roll is failed and there is another potential target for an action or attack within range, then the adversary will now attempt to carry it out against that next target.
- If an Adversary’s Response would mean it would attempt to Move to engage in melee with a hidden model and subsequently carry out an Attack action, then the INT test for detection would take place prior to the Move; if it is failed then the Adversary will instead move to the next model available according to its Target Priority.
That clear separation of having a target for a Response and for an Action means that you don’t have the scenario @DK-dark describes above where, regardless of range, every adversary would try and detect every hidden enemy model that had clear or obscured LOS.
Hey,
I must have misunderstood the example given by DK-Dark then.
I don’t think there is any need for matching AI Responses to Player Model Actions - both of them use Actions, whereas only one uses AI Responses.
In terms of interacting with Hidden Models, Responses and Actions function the same - you need to be aware of them. As Gavin said, you need to be able to see them - whilst they are Hidden, Hidden Models cannot be spotted. They can only be Detected during Actions.
Hi,
I agree on AI Responses and Player Actions not needing to match; that was partially sort of my point; they don’t function the same.
I’m not sure this is the case though; an Adversary can, for example, Move towards a hidden model even if they can’t see them at all (and will have no chance of doing so over the course of their Activation); Gavin even gave the example above of Hold the Gates and Adversaries having an innate ‘sense’ for where their enemies are, allowing them to respond to them, even if they can’t (and won’t) actually be able to detect them.
What they DO need to detect them for is to actually target that enemy with an Action (such as an attack); ergo a Response and an Action not being the same thing. Moving towards a ‘target’ as part of a Response being a separate thing to having a ‘target’ for an attack.
so my point 5 is wrong when I rolled for detecting the model at 30" due the M response?? The Draugr would go anyway towards the first nearest model without rolling, is that right?
Hey both,
Apologies both. It would appear then that Responses effectively do ignore the Hidden Markers, whilst the Actions of an Adversary would not.
I think (but feel free to correct me on this one) in @DK-dark 's example, point five would have the Draugr using the Hidden Model at 30 inches away for the M Response, but the regular targeting rules would apply to the Action itself.
So the Adversary doesn’t roll INT test after rolling the Response? It would move towards the nearest hidden model with clear/obscured LoS
In that example, it would move towards the nearest model full stop, even if it were Hidden and LOS was blocked.
The difference is that it couldn’t then target that hidden model with an Action (such as an Attack) without detecting it. If it’s selected action weren’t in range then it wouldn’t even attempt to detect the Hidden model and would go for the next nearest enemy (although logically nothing would be in range at that point most of the time and it would Block etc.).
Thanks for the Update, it’s definitely an area I feel could benefit from a clearer wording in the updated ruleset because the timing of when the detection test is made etc. can be quite specific and have a tangible impact on its potential success (i.e. if an Adversary moves towards an enemy before taking the test then there’s more chance they’ll get the bonus green die etc.). I know the Adversary rules in general have had some (very good) amendments but this particular area hasn’t been touched as far as I can see.
The one snag I’ve got remaining is detection tests and engaging in melee (and this goes for player models, not just adversaries). Moving into base contact with an enemy model isn’t in of itself, an Action that targets that enemy. So when attempting to engage a Hidden model, when is the detection test made? Prior to even Moving (which we’ve hopefully established now, would be a change to the normal procedure)? Or would the model actually move into base contact but not attempt the detection test unless they were then attempting a melee attack (as the hidden model wouldn’t technically be ‘engaged’ still at that point)? Or would they stop 1/2" away and attempt to detect before carrying on to engage if successful?
Well, this backs to the beginning When I asked, I had played this way, not targeting with the response, the thing was the LoS, because if the hidden model has LoS blocked… well, is hard to imagine how the enemy could know where you are, going towards you even hidden and with LoS blocked, so we need to accept this just for rules purposes.
Still, following the rulebook, with the image I uploaded before, I don’t know if it is a typo, but it says ‘adversaries do the same for detect models, however, they try to detect hidden models when “searching” for a target to Attack’. It seems an additional way of rolling INT test, and it seems the A Response (or maybe M if it is a potential attack), because if it means an attack action that would be redundant (it is the same way for player models, why write the same thing again?)
Following the rulebook, it says you should roll when you enter in contact (not only targeting an enemy), so you would roll after moving and before attacking, and if you fail, you can continue the movement (if you have more movement), engage another enemy (if you have enough movement to reach him), or stay there at 1/2".