Obviously, I’m talking D&D here. I’m not planning on breaking in anywhere.
A couple of recent things have got me asking this question.
The first thing – a lock pick
First, the party were exploring the depths of a dungeon (actually a ruined version of Ragnar’s Keep), and came across a locked door. They were also under attack by a load of “pig-leeches” (My players’ term based on my description and the illustration taken from D&D Beyond).
Jinnie, the Rogue, rubbed her hands enthusiastically about the locked door, and immediately set about picking it. In combat. She succeeded (as she often does, drat her!).
It felt too easy, and particularly in retrospect, felt too quick. A single action – six seconds max – while under attack.
The second thing – a blog post
A few days later, The Angry GM published a post ranting about doors. Specifically how they are now treated as push-button obstacles.
“It’s a door”
“I <skill-check> the door – 17”
“It opens”
(For <skill-check> read “pick-the-lock” or “batter-it-down”.)
As he points out, this rather trivialises it, both as an obstacle and as an interaction with the world. There is no description here of the door – but it almost certainly isn’t like today’s flimsy frame-and-panel doors.
And the lock – what is the lock like? What does it take to pick? Is there a key? As he also points out, because locked doors are nowadays viewed as road-bumps rather than real obstacles, there’s often not even a key.
Lock-picking in early editions
Let’s start with the red box basic set. I have the 1983 Mentzer edition. We immediately have a description of a typical door:
A normal door is made of wood, and often has metal reinforcing strips across it. Some have metal hinges, a ring or doorknob, and possibly a lock, either a padlock hanging on a latch of a lock inset into the door. Normal doors can often be forced or broken open, but a strong bar of wood mounted on the other side will prevent this.
(Basic Players Handbook p57)
A first-level Thief (the BECMI equivalent to the modern-day Rogue) has a 15% chance of opening any given lock, increasing to 20% at level two, 25% at level three, 45% at level seven (where Jinnie is). Even at level 15 (of the maximum 36) they still only have a 75% chance. And that’s a normal lock. “A stuck or exceptionally difficult lock … may cause a penalty to be applied to the normal chances (-5%, -10%, -20%, etc).” (Master Players handbook, p11).
If a character fails to pick a lock, obviously they don’t understand that type well enough, so they can’t try again until they learn more (gain a level).
So locked doors are significant.
I can’t find anything about how long lock-picking actually takes.
How about AD&D? What did Gary say? Surely he has a multitude of rules for this. (Opens Players Handbook – a 1978 1st edition).
Actually, no more than BECMI. A thief’s chance at Open Locks starts at 25% at level 1, 29% at level 2, 33% at level 3, 50% at level 7, up to 99% at level 17 (the maximum). Dwarf thieves have a +10% chance, gnome, halfling and half-orc thieves have a +5% chance, elf thieves have a -5% chance.
Again, it’s a once-per-level chance – if they fail, they cannot succeed until they gain a level.
I have found suggestions that the general rule of thumb at the table was that anything that required careful work was turn-based (10 minutes) rather than the round-based (10 seconds), and so lock-picking fell into that. Which fits my memories. But I haven’t been able to find explicit sources for that.
21st century D&D
Third edition is the first to explicitly specify a time for open locks, and here it says “Opening a lock is a full-round action”. A full-round action is one which consumes the whole (now 6-second) round, preventing any other actions or movement apart from free actions.
So we have our first timescale rule, and it’s six seconds. It also says “You cannot pick locks untrained.” Which makes sense.
In Fourth Edition, it now becomes “Open Lock: Standard action in combat or part of a skill challenge”. Even faster – and matching what we played at the table. Here it also says “The DM might decide that some uses of this skill are so specialized that you are required to be trained in it to have a chance of succeeding.”
In Fifth Edition (2014), pick a lock is listed as an example of a Dexterity Check, but there is no indication of time. The fact that it is a check can’t be taken as a guide, since “march or labour for hours without rest” is listed as an example Constitution check, and that clearly takes hours… However, the Cunning Hands of the Thief archtype of the Rogue allows it as a bonus action which does lean towards it being an action otherwise.
Tales of the Valiant just says “For tasks that require tools, your GM might call for an ability check that uses a tool instead of a skill” and “Thieves tools can be used to pick locks.” No mention there of how long it takes.
Free5e says “proficiency with [Thieves Tools] lets you add your proficiency bonus to an ability checks you make to disarm traps or open locks.” Again, no timescale.
In the latest version of Fifth Edition (2024), under Thieves Tools, it says “Utilize: pick a lock…”, where Utilize “lists the things you can do with the tool when you take the Utilize action.” Now it clearly is just an action.
Level Up! Advanced Fifth Edition is explicit. “You can use an action to open a lock by making a thieves’ tools check…”
Variant Lockpicking 1 – Socratic Dungeon
I found a post by Socratic Dungeon on Rethinking Lockpicking. Note this is based on the OD&D Thief, similar to the BECMI above.
As a starting point, he assumes the picker will succeed – it’s a matter of when rather than if.
He starts with a base time of ten minutes – which maybe confirms the “tentatively 1 turn” above, or maybe contradicts it, given that’s the time for automatic success.
Every 10% chance a Thief has of succeeding on their roll reduces that time by 1 minute. So the 7th-level Thief above with a 45% chance reduces this baseline by four minutes to six minutes. Improvised tools, poor lighting and good locks increase the time by a multiplication factor, and poor quality locks take half the time.
If the Thief is interrupted, they need to make an Open Locks skill check to be able to continue successfully, otherwise they need to restart.
Interesting angle. But I still think there should be a level of skill as to whether the lock can be picked at all, and randomisation in a pressure situation to determine progress.
Variant Lockpicking 2 – Essential Masterminds of Crime
With very serendipitous timing, Essential Masterminds of Crime (affiliate link) dropped through my letterbox yesterday, and they have a section on locks. They have some interesting mechanics.
As a starting point, a lock has a number of pins, and you need a separate check for each pin. So for a 3-pin lock, that’s a minimum of 3 rounds (18s). For a 5-pin, a minimum of 5 rounds (30s). Maybe that interacts with the Thief’s Cunning Hands to allow two checks per round.
They start their DCs lower, possibly to compensate for the number of checks – an Easy lock is DC5, a Moderate lock is DC10, a Challenging lock is DC15, a Very Challenging lock is DC25 and an “Impossible” lock is DC30.
They also have a lock endurance score, from 1 to 4, indicating how many failed checks to pick it can take before it jams.
Building on the added pin mechanism, they introduce some interesting additional options. A counter lock resets as soon as you fail any pin, a timed lock has a number of rounds in which it can picked, after which it jams, and hardened locks have an increasing DC for each pin picked. They also have rusty or old locks increasing the DC by 10, and sticky locks which require both the picker’s action and bonus action.
Again some interesting ideas here. Having to pick each pin individually brings in a longer duration and a random time element based on lockpicking skill, and probably matches the LockPickingLawyer pretty well, and I like most of the variations. But I feel it could goes too far the other way. It now makes lockpicking much more difficult, and also I can see it could get tedious at the table.
Lockpicking in real life
I know D&D is not trying to be a simulation of reality, but rather an approximation that is reasonable and plays well. But it got me wondering – how long actually does it take an expert to pick a lock?
Which then led into what is involved in picking a lock.
A question on Angry’s post got a suggestion of visiting the LockPickingLawyer YouTube channel. From watching a few of these videos and eyeballing the timestamps, we get: about 30s, about 90s, about 90s (which has a comment about it being impressive that it took him more than 15s). This interesting one which shows what’s happening internally with three different tools takes 30s, 7s, 45s.
Which suggests a full round is best case, and multiple rounds is much more realistic.
Except… How did mediaeval locks compare to modern ones?
First source – this YouTube video suggests they just had key guards in the shape of the key, and hence could easily be defeated with a bent strip of metal. None of the multiple pins of nowadays. Maybe the 3s pick is easily plausible.
Or maybe that’s just a simple lock.
According to LowRateLocksmith’s page on ancient lock mechanisms, even the ancient Egyptians had pin tumbler locks, although it looks like there wasn’t a key, so presumably the pins had to be raised by eye. The Romans had simpler locks, but even they had combination locks which required a series of rotations to unlock. And by mediaeval times, the locksmiths had reintroduced the pin tumblers and also added keys.
So we’re effectively back to the same setup and situation as the modern lock. It feels like picking in a single round is only for the simplest of locks, and most locks would take longer.
Conclusions
We have looked at lock picking rules in D&D from BECMI and AD&D, through 3e, and 4e to 5e (both 2014 and 2024), Tales of the Valiant, A5E, Free5e, and then a couple of variant rules.
What have we got from this review of the editions?
| Source (with printing consulted in brackets) | Time for open locks check |
| BECMI (1983 Menzter/Rules Cyclopaedia) AD&D (1978) | Unspecified. Tentatively 1 turn (10 minutes) by convention? |
| D&D v3 (v3.5, first printing 2003, reprinted 2012) | 1 full action (6 seconds) |
| D&D 5e (2014) Tales of the Valiant (2024) Free5e (17-Nov-2025) | Unspecified, possibly 1 action by convention |
| D&D v4 (2008) D&D 5e (2024) Level Up! Advanced 5th Edition | 1 action |
| Socratic Dungeon (OD&D) | 10 minutes, reduced by skill, increased by conditions |
| Essential Masterminds of Crime (2025) | 1 action per pin, typically 3-6 pins |
| Real life | 6-90s for an expert – 1-15 rounds |
At my table we play D&D 5e, mainly 2014 edition with various imports from 2024, ToV and A5E. Based on that, in answer to my question “how long does it take” I suppose I could claim it’s unspecified, although with a possible implicit assumption of one action.
But is that what I want? As Angry says, it feels a bit too much “I <pushbutton> the door”, and it feels like pressure should have some sort of impact.
In a follow-up post, I will explore options for change, try to evaluate the impact, and then propose a new set of rules that capture more what I want without adding too much friction where it doesn’t help the story.
