Who decide which check to make on a recall knowledge action?

In Pathfinder 2, a character can attempt a Recall Knowledge check to get some information about a monster. There are several skills that can be used for such checks, but only some of them work for any monster.

Let’s assume a player decides to use Recall Knowledge on an animated armor (as a Construct it can be identified with either Arcana or Crafting). How does it work? Here are two guesses:

  • The player chooses one skill she wants to use (either Arcana, Crafting, or maybe Religion if she thinks it might be an armor possessed by a ghost). The player rolls and the DM tells her about what she learns regarding that specific skill (for example if the player chose the skill Nature and rolled well the DM would say “this is definitely not an animal, a plant, a fey, or some other natural thing”)
  • The DM tells the player about which skill to roll (depending on what the monster is and which skills the character is trained in). The player rolls this skill and eventually get more information.

The first hypothesis seems more natural to me as a DM but my players told me it wasn’t supposed to work like that. Who is right?

As a GM, how do you decide when a creature has discerned an illusion for what it is?

Many illusion spells, such as minor illusion and silent image have a clause that reads:

If a creature discerns the illusion for what it is, the illusion becomes faint to the creature.

The illusions often have a method or two in the spell description explaining how a creature can determine there’s an illusion, but I’m wondering what other methods, if any, would allow a creature to discern an illusion.

This answer states (and I concur) that there’s nothing in the text that would indicate that the methods given are the only methods to discern the illusion and many people seem to agree that illusionists are not subject to their own illusions, even though that isn’t explicitly mentioned in the spell description; another potential method for discerning illusions is by being told by a trusted ally that the object/creature is an illusion, although I don’t see people talk about that as much. Thus, it seems likely there are more potential methods people may be familiar with as well.

What methods do you allow characters to use to discern illusions outside of those that are mentioned in the spell description, why, and how do you effectively communicate your decision on the issue to players? I suspect there has to be a consideration of game balance, ease of adjudication, and believability, among other things, so I’m hoping to hear about how decisions actually play out when used at the table.

Decide if a language has a word of a given size

Suppose that $ L$ is some language over the alphabet $ \Sigma$ . I was asked to show that the following languages is decidable:

$ $ L’ = \{w \in \Sigma^* | \text{ there exists a word } w’\in L \text{ such that } |w’| \leq |w| \}$ $

I.e., $ w \in L’$ if $ L$ has some word with length smaller than $ |w|$ .

The way I was thinking to show that is observing that $ L \cap\Sigma^{|w|}$ is finite, and $ (L \cap \Sigma) \cup (L \cap \Sigma^2) \cup \ldots\cup (L\cap \Sigma^{|w|})$ is finite too, hence decidable. But the main thing I am struggling with is how can any algorithm for $ L’$ know if some $ u \in L$ ? this is undecidable, so it’s unclear to me how any algorithm for $ L’$ can verify that indeed some word is in $ L$

If I have Extra Attacks, when can I decide which one is a sneak attack?

So, if a character multiclasses to a point where they have multiple attacks in a round, I assume you can decide once the hit has been confirmed by the DM, but before you roll damage, whether you make the hit a sneak attack?

I ask because you may know an enemy is weak, and don’t want to waste the damage to finish it. I also want to confirm that you cannot declare sneak attack damage after you have already rolled your base damage, to finish off a still standing opponent for example. The “Is it dead? No? Oh! Then sneak attack!” seems dumb and nonsensical to me. I expect it to be like smite, decided on hit, but before damage is rolled.

Decide whether two strings $x, y$ can be split into substrings $a,b,c$ such that $x=abc$ and $y=cba$

What is the fastest algorithm for the following problem?

Given two strings $ x, y \in \Sigma^*$ as input, decide whether there exists strings $ a, b, c \in \Sigma^*$ , such that $ x=abc$ and $ y=cba$ .

By calculating all the length of the longest common prefix of $ s = x$ y$ and all suffixes of $ s$ , we can compute all candidates for $ a$ in $ O(n)$ time.

For each candidate of $ a$ with length $ k$ , we now just need to check whether $ x_k, x_{k + 1}, \ldots, x_{n}$ is a rotation of $ y_0, y_1, \ldots, y_{n – k}$ . This can of course be done in $ O(n)$ .

However checking the rotations naively results in a worst-case $ O(n^2)$ algorithm.

It seems that using the result from either https://arxiv.org/pdf/1601.08051.pdf or https://arxiv.org/pdf/1311.6235.pdf would make the algorithm run in expected $ O(n)$ time.

Is there a simpler way of speeding up the rotation checking, where it is still faster than $ O(n^2)$ ? Is there a way of making it deterministic, so that it still runs in $ O(n)$ time?

How to decide for a database structure for a financial accounting app (keeping in mind scaling)?

We are building a financial accounting application for users to manage single and multiple companies under them. The user can be an accountant with n number of companies under it or a single company itself. We are trying to understand how the database for such an application needs to be designed.


  1. The ability of an accountant to see all open invoices across all the companies he is handling.

  2. The ability to archive datasets of companies when they leave us.

  3. The ability to fetch data from multiple companies under one accountant to generate reports.

Database structure:

There are three possible database structures but we need to know which one best suits us:

  1. Have a parent database that holds all accounts and company information. Every company getis its own database to handle and store all transactions.

  2. have a single DB to hold all users and company profile data and every individual company gets its own set of tables to store transactions.

  3. Have a single DB that holds all the transaction data of all companies in a single table called transactions.

We are trying to understand which DB architecture suits us the best. I have MySQL/MariaDB in mind(solely because data is all relational) but if you think other databases would be better, i would definitely like to know more about it.

What does a caster get to decide about the effects of a spell?

When a character casts a spell, a number of decisions need to be made. Some choices are explicitly made by the player. For example, a character can choose the target of a spell (emphasis mine):

A typical spell requires you to pick one or more targets to be affected by the spell’s magic. A spell’s description tells you whether the spell targets creatures, objects, or a point of origin for an area of effect. (PHB)

Certain spells, like Control Water, explicitly allow a character to make additional choices.

Until the spell ends, you control any freestanding water inside an area you choose that is a cube up to 100 feet on a side. You can choose from any of the following effects when you cast this spell. As an action on your turn, you can repeat the same effect or choose a different one…

Other spell, like Earthquake, let the GM make certain decisions:

Fissures. Fissures open throughout the spell’s area at the start of your next turn after you cast the spell. A total of 1d6 such fissures open in locations chosen by the GM. Each is 1d10 × 10 feet deep, 10 feet wide, and extends from one edge of the spell’s area to the opposite side. A creature standing on a spot where a fissure opens must succeed on a Dexterity saving throw or fall in. A creature that successfully saves moves with the fissure’s edge as it opens.

Still others, like Conjure Animals (who decides the exact creatures), Major Image (who decides the exact appearance of the image), or Misty Step (who decides the exact destination), leave the decider ambiguous.

In the case of summoning creatures, the Sage Advice Compendium gives an official interpretation, under the question “When you cast a spell like conjure woodland beings, does the spellcaster or the DM choose the creatures that are conjured?”:

Other spells of this sort (summoning spells) let the spellcaster choose from among several broad options. For example, conjure minor elementals offers four options. Here are the first two:

• One elemental of challenge rating 2 or lower

• Two elementals of challenge rating 1 or lower

The design intent for options like these is that the spellcaster chooses one of them, and then the DM decides what creatures appear that fit the chosen option

A spellcaster can certainly express a preference for what creatures shows up, but it’s up to the DM to determine if they do. The DM will often choose creatures that are appropriate for the campaign and that will be fun to introduce in a scene.

Does this interpretation extend to other ambiguous spells like Major Image or even Misty Step? When a spell requires a choice, and the decider is left ambiguous, is that choice ultimately made by the DM? If true, I realize that in most situations, a reasonable DM would enact the preferences of the spellcaster, barring certain abuses. However, I’m curious about what the Rules as Written have to say about this.

In Eberron: Rising from the Last War, can Warforged become exhausted if they decide not to take a long rest?

In “Eberron: Rising from the Last War” under “Constructed Resilience”, it states the following as a bonus:

You don’t need to sleep, and magic can’t put you to sleep.

and it goes on to say

Sentry’s Rest. When you take a Long Rest, you must spend at least six hours in an inactive, motionless state, rather than sleeping. In this state, you appear inert, but it doesn’t render you Unconscious, and you can see and hear as normal.

Now, what I’ve noticed that has changed from “Wayfinder’s Guide to Eberron” it read

Warforged Resilience … You don’t need to sleep and don’t suffer the effects of exhaustion due to lack of rest, and magic can’t put you to sleep.”

So, in this approved version of warforged, if this living construct refused to take a long rest, would it suffer exhaustion that a normal humanoid of the flesh would likely take, or would it be safe from this kind of treatment?

First time DM, how do I decide whether PC’s can tie up an enemy and other creative ideas

I know these are subjective questions, I just want advice from more experienced player.

For example: My PC tried to trip up a zombie which ran past him while in stealth as an opportunity attack. He rolled a high d20 but 0 for damage (-1 STR) so I decided this would be fair. He trips the zombie up but it takes no damage.

More complex one: He tries to tie up a prone but otherwise perfectly healthy zombie. I thought a die roll isn’t even worth it, because his strength is 9 but the zombie is 16 so following logic he wouldn’t have the strength to grapple the zombie into place in order to tie him up. Should I make it a hard strength check (20) or just narrate it as “You try but the zombies greater strength pushes you back”.

I’ve never DM’d before so I had a quick practice session with one PCs and I’m glad I did, there is a lot to consider!