Read
the thread on fixing/improving Recall Knowledge and really let the arguments provided there soak. Skip my posts if you have to, if you instead read those made by others.
I read them through (all of them!), thanks for the pointer. I also apologize for overstating how much information you give your players. It wasn't clear from your posts and I wrongly assumed it was more than it is. I'm now thinking that for the werewolf example you'd give: creatures starting HP, DR level, but not necessarily what counters it, nothing about attacks or reactions. Please correct me if that's in error.
Since we both like Celebrim's approach, let me quote him here:
The same basic problem applies to something like "Recall Knowledge". You are dealing with something that is information dense. If you tried to have some sort of highly granular rules for "Recall Knowledge" you'd quickly run into the problem that those rules would be larger than all the rest of your rules combined. Since any limited set of rules would be insufficient to cover even a default campaign setting set in a specific location in that world, much of any "Recall Knowledge" resolution process has to be left to fiat both in its DC and the information that results.
Combine that information with the RAW that on a success the GM is required to provide a "
useful clue".
These are very much aligned and very much my experience. RK is a deliberately loosely defined ability that requires the GM to choose the information that results. The only requirement is that it be useful (which, by definition means that RK cannot be a waste of an action, otherwise it would not be useful!).
My contention is that, in your style of gaming, since the players know (because they handle the hp for monsters) that the monster has DR 5, and they know (from previous experience) that silver works against werewolves, that the information that "this werewolf has DR 5 silver" is not useful, and the GM should tell you something that is.
Essentially, because the information is required to be useful, the play style where some monster info is already known by players takes that off the table as a possible RK success result. That puts a bit more burden on the GM, as they now have to move on to less obvious stuff, but from your comments, it seems like you are thinking they would still give the same info, even though it's not useful?
Of course, there is then a problem for the GM if there is no useful information to be gained. Suppose you RK on a generic bandit with no special abilities whatsoever. Your players just fought 20 of them and are pretty clear on their stats, so even giving exact defenses and so on isn't really useful. If I were the GM I might just refund the action as I, as a GM, am unable to comply with the usefulness requirement. I might, if I've run Fate recently, throw in a made-up detail "this bandit favors his right side, you can get a +4 bonus with this knowledge" or the like to make sure I satisfy the usefulness criterion.
The main point is: By RAW, the information must be useful. So if your play style has already disclosed some information, re-disclosing it is not useful, so it should not be considered as a possible RK success.
---------------------
The reason I'm spending some time on this is not to be argumentative, but because our group has genuinely found Recall Knowledge to be not just a bit useful, but a highly valuable thing to do, and I'm trying to work out why this isn't your experience -- because I run at cons and I don't want to TPK parties because they think RK is useless and I expect it to be commonly used. It might be because the opponents we have fought have not been obvious and so RK has been significantly advantageous to us. I'll give some examples, but I'll change some details significantly to protect those who haven't played Ashes yet:
We fought a monster with a reaction that, if you missed them, attacked and on a hit pulled you into a tree and tied you to the tree, from which you would then have to escape. A warpriest went first, did the usual Move-Swing-Swing, missing and getting hit by the reaction oil the second swing and was then taken out of the fight. If they had waited for the magician to go first and do his usual thing (recall knowledge, 2-action spell) they may have changed tactics and used ranged spells, or even just used a single attack and then shielded.
Later we came across an invisible slime. The fighter/cleric had learned his lesson and waited until RK had been rolled. The rogue managed it (RK-bow-bow) and so we found out that the creature could paralyze on a hit. Everyone else immediately changed plans away from melee and we sent in the half-orc warpriest to tank as they had the best combined defense versus the attack on AC followed by fortitude. Full-disclosure: despite shielding he was still hit, failed his save even with a hero point and spent that fight paralyzed also. But his sacrifice helped.
Later again, and late in the evening, we came across some undead and with player knowledge were pretty sure they were boring undead. We expected the skeleton to have DR 5 unless bludgeoning, but were lazy and didn't both with RK. The two big fighters moved to engage some zombies at the far end of the room, leaving a skeleton for the others to handle.
It turned out the skeleton had low AC, DR 10 and regeneration, but very few hits. Our fighter and warpriest were each doing about 2d12+6 and were likely to one-shot the skeleton, but the less powerful hitters were only just beating the DR and never did anything serious to him. Meanwhile his close burst powers were killing them. At the other end, the zombies had a melee reaction to being hit that rolled a critical on the raging no-shield attacker and dropped the main fighter. We came close to a TPK. If we had rolled RK, we would have simply switched targets. The fighter and warpriest on the skeleton would have killed him in a round, probably not needing both their full turns, and the others would have range-attacked the zombie tuning the fight from a near TPK into a cake-walk.
In another session, we all sneaked up to sleeping monster, ready to launch initiative and murderize him. The first player made the RK check and we found it wasn't actually a monster but a trap with a burst attack. So we instead sneaked away, saving us from being fireballed.
---------------------
It is genuinely hard for me to reconcile these examples with your experience. All I can think of is that you generally use pretty well-known monsters for which it's hard to provide useful information for. I might be biased because I tend to play systems where monster variety is high (PF2, D&D 4E, 13th Age) and not so much ones where it relatively low (AD&D, 5E).
But PF2 does seem, at least in published materials, to have a high degree of unusual monsters, so it strongly seem to me that when a GM is required to give some useful information as a result of a RK check, it seems common that there is something useful to give.My experience of the utility of Recall Knowledge is vastly different from yours, so if you have any thoughts on why we get very different value form it, it would help my GMing. Thanks!