DECIDE has racked up many comments on this site, as well as in a related SJG forum thread.

http://www.gamesdiner.com/decide
http://forums.sjgames.com/showthread.php?p=458795

While the idea behind DECIDE is solid, any specific implementation is a throwaway item: there are plenty of possible implementations, and you can easily drop one you don't like and replace with another, all without tossing out the core idea. So I'm not too interested in debating implementations; any will generally fall under "yeah, whatever works for you".

But in the spirit of designer's notes, here are thoughts on why I use the implementation I do, with commentary on objections to it:

On melee

My implementation allows any defender a choice: commit to defense against an apparent threat without first confirming hit/miss ("immediate defense"), or commit to defense only after first confirming hit/miss ("last-second defense"). To set a difference between the two, I used a variable defense penalty for the latter.

There's little (if any) objection to both the article's idea and implementation where bullets are concerned. In melee, meanwhile, there's general agreement that a novice fighter could certainly "waste" a defense as DECIDE allows (though no one, including me, is pushing it as something we necessarily want to worry about).

There is objection to allowing a veteran fighter to "waste" a defense, under the assertion that the vet would never do so (at least, not for game purposes).

That may be. I don't see a problem: by employing free choice in how to defend, the implementation specifically allows the vet to never make the mistake; it never forces him to do so.

In fact, not only does the implementation leave vets free to never make the "wasted defense" mistake, it also – arguably much too generously! – allows novices to never make the mistake! True, the implementation, with its "last-second defense" penalty, can make life hard (and short...) for overconfident novices who proclaim, "I'm not going to defend unless I know I have to!" But that said, it leaves them free to try.

Good luck, little novice.

On discrete breaks

Still, there's the claim that there shouldn't even be an option for the vet to "waste" a defense. If that's what you want, you can certainly change the implementation to that effect, as follows: Let defense versus bullets require action without hit/miss knowledge, for everyone. Let defense versus melee attacks require the same for novices, but not for vets: the latter use RAW defense, always knowing melee hit/miss before committing to defense, which they then perform at no penalty. Done.

What's wrong with that? Nothing so far. You do have to complete it, though, before it's ready to play. You've got to set the discrete "break" between novice and vet, in order to apply the rule. Do you draw the line using "highest combat skill level is X or higher"? Possibly with Combat Reflexes worked in too? Or some other definition of "veteran"?

Actually, you have to set breaks in two dimensions: the break between novice and vet, and the break between melee attacks and bullets. No, the latter isn't obvious: there's an infinite range of speeds in between. Somewhere along that speed continuum is a break where the vet goes from 100% RAW ability to know hit/miss in advance, to no ability to do so. Does the break occur when the missile is as fast as an arrow? A crossbow bolt? Where? Does range of the missile, as some have suggested, further come into the picture? Or is it just a matter of whether or not you can visually track the missile (GM call)?

The breaks don't need to be anything tricky; I'm not setting out to make them look difficult. I only note that they need to be set.

One thing I like about my implementation is no need to set discrete breaks. Both dimensions are already covered, if somewhat abstractly, by penalties. The penalty that the vet can easily afford is one that the novice can't; in practical play, judging that hair-splitting miss becomes a feat for the vet only. Likewise, the variable penalties mean that even the vet will find it prudent to seek cover as missiles become faster, rather than retain full hit/miss predictive ability regardless of missile speed (until that break point where, poof, he loses the ability completely).

That nicety doesn't kick other possible implementations into the dust, but I do like the simplicity of not having to set breaks. Yet it's only a mechanical nicety. There's still a realism objection to tacking any penalty onto the vet for "last-second defense". Is the penalty unrealistic?

On the defense penalty

To correct one apparent misconception, my implementation's "immediate defense" doesn't have a vet (or novice, for that matter) make a defense until it's clear that a legitimate threat exists. It injects no funny business of making defenses way too early to be effective (though with or without DECIDE, you could certainly interpret a really bad defense roll as precisely that, if you like).

Ideally, a defender commits to defend (or not defend) after he's certain that the blow will hit or miss. My implementation only allows for commitment at the split-second prior to that, where the blow could be on target, but the defender hasn't confirmed whether it will be.

If there's anything creating confusion there, it's probably poor naming on my part. My names "immediate defense" and "last-second defense" are perhaps still very poor. I'm only looking to name "Standard defense versus a seemingly legitimate threat" versus "The same, but with an extra split-second to make 100% certain that the attack's not going to fall short by a hair". If you've got better labels, let me know!

Names aside, here's a simple example of a vet taking the penalty in question, which feels plenty realistic to me:

An Orc swings his sword at our hero, Sir Vet, making one normal Attack maneuver per turn – a standard, one-on-one combat situation.

Sir Vet has no reason to hold back his defenses and "confirm" hit/miss. There's no need for him to say "I'm not going to take action unless I know that'll hit!" Against any Orc blow that looks to be a reasonable threat (i.e., is not a crit miss or other lousy TH roll), he's free to say "I defend!" and actively use, not "reserve", his parry. He has no reason to risk doing otherwise.

But now another Orc joins the fight, and Sir Vet is the target of two Attack maneuvers each turn. He's no longer free to parry every blow; if he parries the first blow in a turn, he'll take a hefty -4 to parry the second. If at all possible, he'll want to reserve his parry when that first blow comes, using the defense only if it's absolutely necessary. When the first blow comes, Sir Vet holds back for a tiny instant, committing to a parry only when he's certain that the blow won't miss by a hair. He's fighting differently from how he was a moment ago.

My implementation makes this cautious, selective style of defense a (slightly) more difficult task than less-harried, freer defense, and I find that works fine for novices and vets alike. Defending with intent to "reserve" defenses whenever possible is harder than defending with no such restrictions.

That's the key: the "last-second defense" (or whatever it should be called) is a more difficult task. The way to handle more difficult tasks in GURPS is via a penalty, which is applied equally to both high skill and low. Letting high skill have its usual effect (in this case, a better defense score), while also removing a penalty because skill is high, is "double-dipping" and generally isn't done in the game.

In summary, I like the penalty-based implementation for its non-discrete application of effects in both the novice-vs-veteran dimension and the slow-vs-fast-attack dimension, more than I would like arbitrary, binary break points. I also like it for its ease of play, for its lack of any patches, extra steps, extra rolls, special cases, and other futziness, for its use of a simple, standard penalty mechanism for task difficulty, and, yes, for its realism.

If anyone still disagrees on the realism issue, then let's happily disagree!

But a penalty??

There have been objections to the implementation (and even the DECIDE idea) for making defenses "harder". What can I say? Yes, removing unnatural predictive powers from defenders does make their job harder. I call that an improvement, and will note that it applies equally to everyone; there's no special picking on PCs.

Fortunately, for the main application of DECIDE – knowledge of bullet hit/miss – most every commenter agrees that the change is indeed an improvement. For those who don't like its application in melee, the solution is easy: don't use it.

But what if you do happen to like my DECIDE implementation in melee, but just don't like any penalty involved (even a small one applied to PC and foe alike)?

You're in luck: there's a dead-simple adjustment you can make. Just add +1 to all the suggested modifiers. The result in melee:

You make a normal melee defense at +0 (instead of the -1 I used) when you confirm TH before committing to defense. This is exactly per RAW.

Or, you can defend at +1 (instead of the +0 I used) if you commit to a defense before confirming TH.

Hmm. Although I've always been happy to play my "old" implementation, I rather like this little change! There's just that +1 difference from the old, but it's perhaps much better design. It leaves everything unchanged for a RAW melee defense, in which you first confirm hit/miss. Meanwhile, the +1 to defense for not confirming hit/miss better emphasizes that this is a new defense option: a defense that commits to countering any legitimate-appearing threat, but which carries the risk of "wasting" the defense against an attack that would have missed by a bit.

In other words: No change for those choosing RAW; a change only for those choosing a new option. That's a design approach I try to use through all my hacking, but have neglected to apply here. Maybe I should have been playing DECIDE this way from the start!

Closing

Would that simple "+1" tweak make for a better implementation than the current one? Let me know, readers.

And whether you think that change is better or not, if you've got some totally different, shinier, preferred way to handle the whole fershluggin' deal, then that's the implementation for you. Happy defending!