[Drubbels]

Well, yes, I strongly prefer actions and reactions coming from the same pool, too.

Hmm... we could have this:

You can use an achieved action as the 'Ready' action to prepare a specific reaction for later use under specific circumstances. You can have several of these (at a penalty to their execution, as outlined in the document), and you can chain several of them together and execute them in sequence.

Alternatively, you can, when your turn comes, choose to roll fewer action dice then you would normally do. Say, Inquisitor Shyloque, who has Speed 5, wants to run to the edge of the roof, then wait for further developments. The player first rolls 3d6 and achieves 2 actions, sufficient to reach the edge. What he uses his remaining actions for depends entirely on what happens later in the turn. For example, any of the following might occur:

*Sgt. Henchman, standing in the street, lines his heavy stubber up to open fire on the rooftop. Deciding he doesn't want to get shot, Inquisitor Shyloque declares his two remaining actions to run backwards for a few yards and throw himself prone, then use 'Ready' to prepare fire at the door to the roof. He rolls his 2d6 and achieves 1 action, running back and throwing himself prone. Sgt. Henchman may now Pause for Breath and re-assign any further actions, as his intended target is his lost.

*An Arco-Flagellant bursts through the hastily barricaded door and on to the roof. Because he doesn't want to risk having to fight the thing at close quarters, Inquisitor Shyloque wants to drop himself over the edge, and will spend his next action crawling into cover on street level. He rolls 2d6 and achieves 2 actions, dropping down and hiding. The Arco-Flagellant may now Pause for Breath.

*Inquisitor Shyloque spots an injured Inquisitor Tyrus running away from Shyloque's allies. Deciding to take a lucky shot, he wants to spend his two actions aiming and firing his plasma pistol at Tyrus. Rolling 2d6 and achieving both actions, Shyloque hits Tyrus.

*Nothing of note happens within Shyloque's line of sight during the remainder of the turn, and once the Speed 1 step rolls around, Shyloque decides he isn't going to wait around on the roof all day and carefully drops himself to street level.

This way, what you're essentially doing is splitting your turn up into two halves. You can take the remainder of your turn at any time during the round, even interrupting the turn of a lower-speed character. The penalty to this approach comes from splitting your action dice up into two groups without prior knowledge of which ones will succeed and which ones will fail, meaning you may very well have insufficient actions in one half of your turn and more than you know what to do with in the other. The advantage is that you're guaranteed a chance to use the actions (aside from the action dice failing) - even if nothing at all happens to trigger your reaction, you can still use them for whatever you want to be doing (most likely movement) near the end of the round.

[Alyster Wick]

There's a lot going on here in terms of reaction discussions, so I'm just going to drop in my two cents on a few things:

- It makes sense for characters not to announce what they are saving their reactions for. While I prefer the actually triggering mechanic you mention Noob, the impact (as Marco points out) on actual gameplay is too much. Most players who know that their character will be shot if they take a specific action will take that into account (consciously or not) and it'll impact the way folks act. Which is a shame, I like the idea of declaring reactions like that.

- I'm in favor of allowing players to store a certain number of reactions that cross turns. The main reason for this is that characters with a low speed are pretty much relegated to NPC status in these scenarios. The higher speed characters will have shifted back and forth between each other in dramatic duels and reactions by the time it comes down to the lower speed folks who never get the chance to operate that way. Which brings me to my next point:

- Allow one reaction at a time. The whole "splitting the turn" dynamic sounds interesting, but I think allowing one reaction at a time helps allows characters to have some agency outside their turn while preventing another character's turn from being completely highjacked. It also makes it less appealing for a character to store up an inordinate amount of reactions.

- One thought on limitations: A character can only take a number of actions in their given turn equal to their speed. Meaning that if a character has a particularly lucky roll (5 of 5 actions, for example) then they are unable to take an reactions that turn.

Those are really random thoughts, feel free to use or discard as you see fit. I think I'll abandon my own CC thread as it seems more productive to just follow along here and add to this conversation.

Also, Marco, if you have a working draft of your rules in a single document I'd love to check it out and take them for a test drive. 

[Drubbels]

I do partly share the concerns about influencing how you play when you know about certain reactions, but consider the following two scenarios:

* I'm playing Inquisitor Shyloque. I could either round the corner and walk past the building towards my objective, or take a long detour. Around the corner, I know as a player, is the opposing Sgt. Henchman. He is holding his lasgun and has prepared a reaction to shoot me once I enter line of sight. I take the detour.

* I'm playing Inquisitor Shyloque. I could either round the corner and walk past the building towards my objective, or take a long detour. Around the corner, I know as a player, is the opposing Sgt. Henchman. He is holding his lasgun and has reserved a reaction to use as he sees fit. What do I do now?

I'm inclined to say that in the second scenario, just like in the first, my style of play will be influenced (or I might consciously will myself into taking the bad decision, just to prove that I'm not metagaming - which is not a highly desirable outcome either). I wonder if perhaps the difference in actual gameplay might not be as big as we think... but then again, there are probably lots of scenarios where it does matter, too. I suppose, for when it really needs to be secret, you could always just use notes or whisper into the GM's ear, but that's not practical as a standard mechanic. I do think that the option to do it like this should be there though, as it's really just a very extensive generalization of the existing overwatch mechanic.

To end this rant, I've found my opinion shifting somewhat towards your position while writing this. So yeah, there should be a solid system for non-pre-declared reactions.



As to your second point, I'm not 100% sure what you're for and what you're against. Could you re-phrase?



I agree that perhaps splitting the turn isn't optimal - while (in my mind at least) it's the most natural way to extend the existing action mechanic to reactions, it would perhaps be a total bookkeeping nightmare in large games. I do think there should be some possible way of taking multi-action reactions, though - perhaps only as prepared reactions?



However reactions are handled in the details, I'm firmly of the opinion that they should be drawn from the same pool as regular actions, so that'll provide a limitation in and of itself.

============================================================================================================================

So, new suggestion:

*Prepared Reactions: You can prepare reactions using the 'Ready' Action, set triggers, and make sequences, as discussed in my original document. Perhaps, to reduce bookkeeping, a character can't have more than one multi-action sequence of reactions.

*Reserved Reactions: You can set aside action dice before rolling for actions, and store them (visually indicated by leaving the appropriate number of d6's next to the character's model) for later use. You can interrupt another character's turn, but you can declare only one action at a time to do so. You roll however many dice you want, but any passes after the first are disregarded - rolling more than one die serves only to increase your chances of passing the action, not granting a chance at multiple actions. Failing all dice means the dice are wasted, and you cannot execute the reaction. Should you have lots of action dice left, you can attempt several reactions, but you can only make one attempt to react to each of the acting character's actions - so you can try again, but not until he performs another action.

*As discussed earlier, Action rolls will pass on 3+. Also, when rolling action dice for Reserved Reactions, there is no 'free' action if you fail all rolls.

[MarcoSkoll]

You'll have to forgive a rather short response, I'm in the midst of insomnia at the moment.

I'm inclined to say that in the second scenario, just like in the first, my style of play will be influenced.

It was a rather simplistic example, and I have to admit that both would have an effect. Still, what's rather important is that the two scenarios can have different lead-ins.

In the first scenario that Sgt Henchman would probably have to aware of hostile characters - in the same way I'd normally forbid a the declaration of actions against characters that you're unaware of, I'd also want to know exactly why the Sgt is aiming at that corner "just in case" someone came around it.

In the second case, the undeclared reaction could reasonably be held on the excuse of "general paranoia".

*Prepared Reactions: You can prepare reactions using the 'Ready' Action, set triggers, and make sequences, as discussed in my original document. Perhaps, to reduce bookkeeping, a character can't have more than one multi-action sequence of reactions.

*Reserved Reactions: You can set aside action dice before rolling for actions, and store them (visually indicated by leaving the appropriate number of d6's next to the character's model) for later use.

Largely, this is similar to my current drafts (although the idea of using multiple stored reactions at once to improve your chances is an interesting one), except for omitting the "Wary" action - which I think is worth keeping.

It's only my current draft that explains the design intent of Wary, so it might seem a little out of place otherwise in the earlier versions. The short version is it's there mostly because of how often I see characters end their turns with actions left over and players express laments that they can't be saved for later.

Also, Marco, if you have a working draft of your rules in a single document I'd love to check it out and take them for a test drive.

There is a very out of date link, but the current version is a mixture of very messy notes and in my head.

I do need to try and produce a more up-to-date version though.

[Drubbels]

Yes, it seems that I've found my opinions drifting closer and closer towards what you'd already written. If you could put together another draft, I'd very much like to read it (testing will probably have to wait 'till next month, because of upcoming exams).

I suppose reactions are partly settled now, so perhaps we could introduce another important topic of discussion: close combat. I've brainstormed a bit on the previous page, and most of it probably isn't usable in its current form, but I hope it at least provides some useful ideas.

[MarcoSkoll]

If you could put together another draft, I'd very much like to read it

I'll do what I can, although it may take a little while to get it into a vaguely polished format.

I suppose reactions are partly settled now, so perhaps we could introduce another important topic of discussion: close combat.

I actually quite like your idea that close combat attacks automatically manoeuvre the opponent, as giving the movement for free means players will try and actually use it.
I'd still want an option where the intent is actually to move the fight, as that's quite reasonable to want to do, but I suspect it will be considered quite situational.

As far as Hazardous Parrying... I think that's one to veto. I find parrying and counter attacking considerably more interesting, but rarer, than attempting to dodge in close combat, and don't want to do anything to dissuade it. Keeping it appealing may already be fairly tough, because once you have to beat your opponent's margin of success (effectively penalising your defence), the higher chances of sidestepping will already be a significant advantage.

[Drubbels]

As far as Hazardous Parrying... I think that's one to veto. I find parrying and counter attacking considerably more interesting, but rarer, than attempting to dodge in close combat, and don't want to do anything to dissuade it. Keeping it appealing may already be fairly tough, because once you have to beat your opponent's margin of success (effectively penalising your defence), the higher chances of sidestepping will already be a significant advantage.

Fair enough. I wasn't really convinced either. I was thinking of melee fights in movies, and one thing that appealed to me was how these typically involve grabbing your opponent and pulling them to the ground and/or repeatedly ramming them into solid obstacles. This was just the first thing that came to mind to try and emulate that. And if we want to incentivize sidestepping over parrying, we'll probably have to drop Critical sidestepping, too.

(Come to think of it, nearly all of these movie fights are unarmed, so maybe modelling Inquisitor's swordfighting on them isn't a great idea.)

[MarcoSkoll]

Sorry, it's taking longer than originally planned to get my latest thoughts down in words as, predictably, thinking about it keeps changing what I have in mind.

What I'm currently contemplating is if I can stop close combat sitting apart from the rest of the rules, by removing or minimising any specific "melee" state. Instead characters would get automatic reactions if an action is performed within melee range of them (Within certain limitations - no running past your own characters to give them reactions!). Making melee something simply defined by a vicinity could also allow freer movement. No specific "step back" action to get to arm's length, it's now just standard movement.

Hopefully the whole thing will allow more fluid engagements where characters are able to constantly manoeuvre in and out of melee range.

I'm going to have to work on ironing out a few issues (like not making it too easy to just naff off out of a melee), but I think it could be interesting.

[Drubbels]

I just got to play an unexpected game yesterday, intending to test out the reaction mechanic and some close combat ideas.

Not a single character was involved in a close combat (one unconscious npc was executed by pumping an autopistol into his neck at point blank, but that hardly counts), so not much news on that front.

We used reactions exactly as outlined in my earlier post from January 3, which I think went reasonably well - it sure made for some interesting scenarios which wouldn't have been possible under the normal rules. We had several characters standing on rooftops and bridges, jumping on a truck as it sped by. There was also one instance of an assassin preparing to drop a plasma grenade on that same truck as it attempted to make its escape under a walkway he was standing on, only for the telekinetic psyker who was riding shotgun to use his reserved reactions to fling it right back.

[MarcoSkoll]

Ugh... sorry, very late response to this.

it sure made for some interesting scenarios which wouldn't have been possible under the normal rules.

Sounds good - that's largely what I'm hoping for, that a reaction system allows a better flow of the action.

Have you any thoughts on how it affected the speed of gameplay?

[Drubbels]

Given that this was our first ever time using the mechanic, I felt it went pretty smoothly. I suppose next time I could play two different games of a similar size, one with reactions and one without them, and have someone time all the turns with a stopwatch to see the actual results. But by gut feeling, I don't believe there was a significant slowdown, save the few times we'd go "so I'm going to do this - oh no wait, there's reactions, so I guess instead I'll do this..." (because it was our first time).

[MarcoSkoll]

Don't worry about timing it - I'm not too fussed about the individual seconds, more the overall impression of the rate the game was progressing.

While I've played a few games with reactions and felt the speed was good, I'm pretty good at juggling rules in my head, so it helps to get feedback from others too.
A more reactive game is a desirable goal, but if it actually felt sluggish at the same time then some drastic rethinking would have been needed.

[MarcoSkoll]

Right, I've been in a thinky place for the last few hours, and it's been about re-engineering semi and full auto fire more.

I don't want to move onto Dark Heresy's method where the number of hits is just a direct function of the degrees of success - whenever autoguns get used in our DH sessions, they often score four or five hits (fine for shooting NPCs, but not so great in a PvP system) and almost never score a full ten hits (in most circumstances, a DH character would benefit from autoguns being Full 5 or 6, as they're very unlikely to score 7 or more hits anyway but would get more bursts from a magazine)

However, while this is a very early draft that I've not rigorously run the numbers on yet and for which I will probably want to tweak the modifiers (a lot), here's a possible proposal for semi-auto fire:

Basic BS, double range penalty, +5 per shot in the burst, +20 for Semi-auto.
If a shot beats the margin of success of the last shot in the burst, roll for another shot (up to the maximum used in the burst). If not, all remaining shots automatically miss.

This obviously abstracts things more, such as the sequence of the shots (the shots that hit might not actually be the first in the burst), but an "exploding dice" mechanic does considerably reduce the amount of rolling needed, as you have to roll a maximum of one miss per burst (and quite possibly none, as a shot can be a hit, but not successful enough to allow the next shot).

Semi (6) therefore wouldn't be particularly more time to resolve than Semi (3) - and a variant of this for full auto fire would make a massive difference to the time it took to resolve a burst, as bursts of many tens of shots would normally be resolved in just a few rolls.

This is weighted IMHO somewhat better than DH's system, as the need to chain successes means that one hit is more likely than four (in DH, both are equally likely if your target number is high enough), but there is actually a reason you might want to fire more shots than you're likely to hit with (as it affects your target number) and no shots that are absolutely doomed to miss (although admittedly are still very unlikely to hit!)

[Drubbels]

That's... actually a very interesting idea. Very unorthodox (I'd never have thought to do it like it that), but it sounds cool. I particularly like the fact that it takes no real effort - apart from remembering the result of the d100 you just rolled - to see whether you score another hit.

I would like to run the numbers, but that actually looks fiendishly complicated in this case.



Having scribbled around a bit, I think this is how to do it:

(I'll be ignoring automatic passes and failures, because frankly that'd be too much work. Plus, to-hit chances below 5% almost never happen anyway, outside of the RAW version of full-auto fire)

Call x your final to-hit chance in %. This includes everything - BS, range, +5's and +20's... Call n the number of shots you fire (i.e., you're firing on semi(n) or full(n) ).

The odds of scoring 0 hits are then (100 - x)% . The odds of hitting your first shot are x%. The expectation value of the margin of success obtained on the first roll, given that it passed at all, is [(x+1)/2] , which will be your target number for shot nr 2.

The odds of scoring exactly 1 hit (hitting nr 1, missing nr 2) are then x% * (100 - [(x+1)/2])% . The expectation value of the margin of success obtained on the second roll, given that it passed at all, is [([(x+1)/2] +1)/2] = [(x+3)/4] , which will be your target number for shot nr 3.

The odds of scoring exactly 2 hits are then x% * [(x+1)/2]% * (100 - [(x+3)/4])% ...

...and so on, until you reach the nth roll, where the final factor is omitted because no matter how good you roll, there won't be an (n+1)th shot.

To summarise, given a to-hit roll of x%, the chance of getting exactly k hits is:

k < n: P(k) = (100 - [(x + (2^k) - 1)/(2^k)])% * Product of ( [(x + (2^j) - 1)/(2^j)] % ) for j going from j = 0 through j = k-1

k = n: P(k) = Product of ( [(x + (2^j) - 1)/(2^j)] % ) for j going from j = 0 through j = k-1

k > n: P(k) = 0

And then the expectation value of the number of hits is Sum of ( k * P(k) ) for k going from k = 0 through k = n (or k = any larger number, for that matter).



One last thought is that this might cause issues when a character louds multiple different kinds of ammo in his magazine and fires more than one kind during an action (which shots hit), but that's such a rare event that I really don't care all that much.

[MarcoSkoll]

It is a slightly odd mechanic, but I've been trying to figure out an escape from the normal one hit roll for every shot - while it's fine as a system when each hit roll is one die (as in WH40K), it gets messy when you're using two dice for a D100 roll (even if you've got multiple D100 colours, as I have - they tend to get jumbled) and rolling it one shot at a time gets tedious past about Semi (4).

I don't like the DH method much (for the reasons in the last post), so exploding dice occurred recently.
Having a flat target number was a problem (as the expected number of hits is an exponential function that ramps up scarily fast with initial hit chance). I was originally playing with an idea to have cumulative penalties that built up per shot, but it was getting rather maths heavy.

The margin of success thing on average halves the likelihood from the last shot, just with practically zero maths.

One last thought is that this might cause issues when a character louds multiple different kinds of ammo in his magazine and fires more than one kind during an action (which shots hit), but that's such a rare event that I really don't care all that much.

I'd suggest either continuing the abstraction (and having the first hit be the first round and so on), or just simply randomising it.

It's also messy with Hazards and weapon jams, but I think that can probably have some kind of fudge factor applied so that the first roll has a higher risk and then it's ignored for the rest of the burst.