In an earlier post I lamented how trivial picking a lock felt during a session, and then dug into the question of how long lock picking should take.
I promised to suggest an alternative that felt more appropriate.
That post took a tour of 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 a couple of variant rules, and investigated how long it takes in real life, but didn’t settle on a proposal.
The next post was another preparatory one. It investigated how probabilities change under various conditions, particularly given multiple attempts. It also discussed the 3E concepts of “take 10” and “take 20”. There’s a lot of detail in it, but the key takeaways are that the chance of success rises rapidly if you are allowed multiple attempts, and adding a chance of failure means that you will probably come a resolution one way or the other within a few rolls.
With all that preparation, it’s time to try to come up with something which works the way I feel it should.
Additions
But a little bit more preparation first. Since writing the previous post, I have come across a couple of other proposals:
Groundbreaking-Law-3 suggests on Reddit making lockpicking into a puzzle, where the lock has pins which need to be picked in a particular order, and the GM provides hints as to whether a particular target pin is next in line, adjacent to the next in line, or not. The character skill level determines how many “guesses” they get.
This reduces the effect of character skill, making it more about player skill. Which may be what you want. It also makes lock-picking take longer at the table.
Harry Menear, on Black Citadel, has a suggestion that lockpicking involves rolling a series of dice to get within a target range – a wide range for easy locks, a single number for the hardest locks. Overshooting has a 3-in-4 chance of consequences, and progressive skill leads to rolling smaller and smaller dice (thereby making it harder to overshoot).
Harry’s suggestion rewards proficiency and has consequences for failure in quite a neat way.
Both of these convert lock-picking to a new game-within-the-game.
Existing Sources – summary
Here’s an overview of the different systems I looked at.
| Source (with printing consulted in brackets) | Time for open locks check | Chance of success normal lock | Failure? |
| BECMI (1983 Menzter/Rules Cyclopaedia) | Unspecified. Tentatively 1 turn (10 minutes) by convention? | 15% at L1 45% at L7 75% at L15 | One chance |
| AD&D (1978) | Unspecified. Tentatively 1 turn (10 minutes) by convention? | 25% at L1 50% at L7 99% at L17 | One chance |
| D&D v3 (v3.5, first printing 2003, reprinted 2012) | 1 full action (6 seconds) | DC25 <= +4+Dex at L1 <= +10+Dex at L7 <= +18+Dex at L15 | Repeat at will |
| D&D 5e (2014) Tales of the Valiant (2024) Free5e (17-Nov-2025) | Unspecified, possibly 1 action by convention | DC15 +2+Dex at L1 +3+Dex at L7 +5+Dex at L15 | Repeat at will |
| D&D v4 (2008) D&D 5e (2024) Level Up! Advanced 5th Edition | 1 action | Repeat at will | |
| Socratic Dungeon (OD&D) | 10 minutes, reduced by skill, increased by conditions | Automatic | N/A |
| Essential Masterminds of Crime (2025) | 1 action per pin, typically 3-6 pins | DC10 x 3-6 | 1-4 failures allowed |
| Groundbreaking-Law-3 | Puzzle | Allowed prof bonus “guesses” | |
| Harry Menear | Multiple rolls until hit target window | 5-number window with sequential d8s if proficient Rogue | 75% chance of failure if go over target |
| Real life | 6-90s for an expert – 1-15 rounds | Yes |
Melestrua’s Variant Lockpicking
Let’s put it all together. What do I want?
In reviewing alternatives, I want what I pick to “feel” right, as in, for a character of a given level of proficiency, does their probability of success and the time it takes match what I would expect.
I also want it to feel like “normal” D&D 5E, and not some special game-within-a-game.
That means most of the options considered just disappear. The way lock-picking works in 5E is a DEX skill check with Thieves’ Tools, so that’s the baseline. Ideally, I’d like to still use the guideline DCs as well.
And, while I want it to be easier if you have time, I don’t want the character to be able to try forever.
So, what knobs do I still have under my control?
- I can introduce a chance of failure
- I can introduce a chance of failure with consequences
- I can impose advantage or disadvantage
- I can rule on the impact of proficiency
- I can change how long an attempt takes
And what am I looking for out of this?
- Something which may take longer than a single action or round
- Something which rewards proficiency
- A way of making it harder under pressure
- Something which allows for failure
Test cases
In mentally testing these rules, I will consider:
- Jinnie, 7th level Rogue who has Dexterity 18 (+4) and Proficiency with Thieves Tools (which at level 7 is +3), for a total modifier of +7
- Ramon, 1st level Rogue with Dexterity 14 (+2) and Proficiency with Thieves Tools (+2 at level 1), for a total modifier of +4
- Lluisa, 13th level Rogue with Dexterity 18 (+4) and Expertise with Thieves Tools (+10 at level 13), for a total modifier of +14
- Gareth, 7th level Fighter, who has Dexterity 12 (+1) and no Proficiency, for a total modifier of +1
- Carlos, Commoner, level 0, no proficiency, all scores at 10 (+0), for a total modifier of +0
It feels like Lluisa should have an excellent chance of picking a standard lock, Jinnie should have a good chance of picking a standard lock, Ramon a moderate chance, and Gareth and Carlos should struggle., Ji
In fact, I think I will say up front that Gareth and Carlos should not be able to pick a lock – you need to know what you’re doing to have any chance (i.e. you must have Proficiency with Thieves’ Tools to have any chance).
We also have different levels of lock to consider:
- Simple lock, such as on a backwater farm in a calm area – DC10
- Ramon should have a pretty good chance at this, and it should give Jinnie and Lluisa very little trouble
- Average lock, the sort you encounter everywhere – DC15
- Ramon should have a fair chance, Jinnie a pretty good chance, and it should give Lluisa little trouble
- Good lock, like those used by rich merchants and dwarf moneylenders – DC20
- Ramon might have a bit more trouble, Jinnie should have a fair chance, and Lluisa should find it relatively straightforward
- Exceptional lock, the door to the ancient dwarven vault – DC25
- Probably beyond Ramon, Jinnie may struggle, Lluisa should still have a fair chance
- Master lock, such as the lock Dwarven god Kagyar created to protect his ancient shrine – DC30
- Probably beyond both Ramon and Jinnie, requires an expert like Lluisa to even have a chance
Standard rules
How do the standard rules fit this?
| Ramon (+4) | Jinnie (+7) | LLuisa (+15) | |||||||
| Number of rolls | 1 | 3 | 6 | 1 | 3 | 6 | 1 | 3 | 6 |
| Farm door lock (DC10): | 75% | 98% | 90% | 99% | 100% | ||||
| Standard lock (DC15) | 50% | 88% | 98% | 65% | 96% | 99% | 100% | ||
| Merchant’s lock (DC20) | 25% | 58% | 82% | 40% | 78% | 95% | 80% | 99% | |
| Ancient dwarven vault lock (DC25) | 0% | 15% | 39% | 62% | 55% | 91% | 99% | ||
| Kagyar’s master inner sanctum (DC30) | 0% | 0% | 30% | 66% | 88% | ||||
I have to say that, looking at these chances, the initial chances look fairly reasonable if there is no pressure, but the overall chances of success feel like they ramp too fast and too far.
Chance of failure
What happens if a fail by 5 or more is terminal? Remember that has a 99.2% chance of resolving one way or the other within 3 rolls, since there is only a re-roll if the roll misses the target by between 1 and 4.
| Ramon (+4) | Jinnie (+7) | LLuisa (+15) | |||||||
| Number of rolls | 1 | 3 | 6 | 1 | 3 | 6 | 1 | 3 | 6 |
| Farm door lock (DC10): | 75% | 93% | 93% | 90% | 99% | 99% | 100% | ||
| Standard lock (DC15) | 50% | 62% | 62% | 65% | 81% | 81% | 100% | ||
| Merchant’s lock (DC20) | 25% | 31% | 31% | 40% | 50% | 50% | 80% | 99% | |
| Ancient dwarven vault lock (DC25) | 0% | 15% | 19% | 19% | 55% | 68% | 69% | ||
| Kagyar’s master inner sanctum (DC30) | 0% | 0% | 30% | 37% | 37% | ||||
That feels like the ultimate chance of success is too low.
What about if fail by 10 is terminal? That has a 91% chance of resolving one way or the other within 3 rolls, 99% chance of resolving within 6 rolls.
| Ramon (+4) | Jinnie (+7) | LLuisa (+15) | |||||||
| Number of rolls | 1 | 3 | 6 | 1 | 3 | 6 | 1 | 3 | 6 |
| Farm door lock (DC10): | 75% | 98% | 90% | 99% | 100% | ||||
| Standard lock (DC15) | 50% | 66% | 67% | 65% | 85% | 87% | 100% | ||
| Merchant’s lock (DC20) | 25% | 33% | 33% | 40% | 52% | 53% | 80% | 99% | |
| Ancient dwarven vault lock (DC25) | 0% | 15% | 20% | 20% | 55% | 72% | 73% | ||
| Kagyar’s master inner sanctum (DC30) | 0% | 0% | 30% | 39% | 40% | ||||
Hmmm. Slightly more chance ultimately, but not enough to be significant.
Critical failure
What if, instead of a “fail by”, we introduce critical failures – a roll of 1 is terminal. That adds a guaranteed 5% chance of failure on every roll – 5% on the first roll, 9.8% within two rolls, 14.3% within three rolls, 18.5% within four rolls, 22.6% within five rolls and 26.5% within 6 rolls. I will assume that there’s still no roll for Lluisa when her bonus means she cannot fail.
| Ramon (+4) | Jinnie (+7) | LLuisa (+15) | |||||||
| Number of rolls | 1 | 2 | 3 | 1 | 2 | 3 | 1 | 2 | 3 |
| Farm door lock (DC10): | 75% | 90% | 93% | 90% | 94% | 95% | 100% | ||
| Standard lock (DC15) | 50% | 72% | 82% | 65% | 84% | 90% | 100% | ||
| Merchant’s lock (DC20) | 25% | 42% | 55% | 40% | 62% | 74% | 80% | 92% | 94% |
| Ancient dwarven vault lock (DC25) | 0% | 15% | 27% | 37% | 55% | 77% | 86% | ||
| Kagyar’s master inner sanctum (DC30) | 0% | 0% | 30% | 49% | 62% | ||||
That feels like too much of a penalty for easy locks, and not enough for the very hard locks.
What if instant failure is on (DC/10)?
| Ramon (+4) | Jinnie (+7) | LLuisa (+15) | ||||||||
| Number of rolls | Fail | 1 | 2 | 3 | 1 | 2 | 3 | 1 | 2 | 3 |
| Farm door lock (DC10): | 1 | 75% | 90% | 93% | 90% | 94% | 95% | 100% | ||
| Standard lock (DC15) | 1 | 50% | 72% | 82% | 65% | 84% | 90% | 100% | ||
| Merchant’s lock (DC20) | 2 | 25% | 41% | 52% | 40% | 60% | 70% | 80% | 88% | 89% |
| Ancient dwarven vault lock (DC25) | 2 | 0% | 15% | 26% | 35% | 55% | 74% | 81% | ||
| Kagyar’s master inner sanctum (DC30) | 3 | 0% | 0% | 30% | 46% | 56% | ||||
Similarly too much of a penalty for easy locks, and not enough for the very hard locks.
What if instant failure is on (DC/5) – 2?
| Ramon (+4) | Jinnie (+7) | LLuisa (+15) | ||||||||
| Number of rolls | Fail | 1 | 2 | 3 | 1 | 2 | 3 | 1 | 2 | 3 |
| Farm door lock (DC10): | 75% | 93% | 93% | 90% | 99% | 99% | 100% | |||
| Standard lock (DC15) | 1 | 50% | 72% | 82% | 65% | 84% | 90% | 100% | ||
| Merchant’s lock (DC20) | 2 | 25% | 41% | 52% | 40% | 60% | 70% | 80% | 88% | 89% |
| Ancient dwarven vault lock (DC25) | 3 | 0% | 15% | 25% | 33% | 55% | 71% | 76% | ||
| Kagyar’s master inner sanctum (DC30) | 4 | 0% | 0% | 30% | 45% | 52% | ||||
That’s starting to look more reasonable. But it’s also getting quite mathematical…
How long does it take
Maybe I’m approaching this the wrong way. Rather than imposing potential failure conditions, maybe I should look at how long it takes.
In the various sources it varies between one round (6s) and one old-school turn (10 minutes). From watching the lock-picking lawyer, it seems to typically take him between 15 and 90 seconds.
One option would be to have an exponential back-off. Maybe the first attempt takes 1 round (6s). The second attempt takes 1 minute. The third attempt takes a turn (10 minutes). Beyond that it’s not happening.
That still leaves it with the “too quick” situation on the first check – six seconds – which prompted this whole series of posts.
Maybe roll a die to see how many rounds it takes – say d4+1.That means the first check takes between two and five rounds (12-30 seconds) – if you’re in a pressure situation, that’s going to add to the pressure. It also makes it potentially very costly during combat, and certainly you’re only going to get one, or at best two attempts.
However, if you’re not in a pressure situation (i.e. counting in rounds), the time doesn’t matter on the first check – even 30s is far less than the nominal 10-minute adventuring turn. And if each attempt takes 12-30s, you can still fit nearly 30 attempts into a 10-minute turn. There needs to be some additional limit.
Maybe failure on the first attempt indicates that it’s taking more working out.
If the first one failed, does the second one also take d4+1 rounds? Or does it take longer? Does the die get larger – maybe d6+1 rounds? Or double dice – 2d4+2?
Realistically we’re now into territory where it’s not going to happen during combat, but it is still a small part of a full 10-minute turn, so it feels like we’re arguing about angels and pinheads. I don’t think this is a good way to go.
How about we add a scaling factor? Maybe the second attempt takes d4+1 minutes. The first two attempts still both fit within a turn. But the second attempt still isn’t going to happen during a combat situation, unless for some reason it feels appropriate to count time at a granularity between rounds and turns. Which might be appropriate during some chases, where picking a lock may still be relevant.
Do we allow a third attempt? From looking at the way the probabilities with failure on miss-by-5 converged on 3 attempts, it feels like maybe if there is no pressure, a third attempt is allowed, and takes a full turn, but no more.
Under Pressure
What about if the player wants to speed things up? Attempt it in a single round? Do we allow that?
If it normally takes longer than a round, maybe you can move one step faster, but at Disadvantage. First attempt in one round, second attempt in 2-5 rounds, final attempt in 2-5 minutes.
That would have no effect for Lluisa on a normal lock – since she can’t fail, she still can’t fail with disadvantage, so she can happily pick any normal lock in a single round. What impact does it have across the different locks (on the basic check before we start adding complications)?
| Disadvantage | Ramon (+4) | Jinnie (+7) | LLuisa (+15) | ||||||
| Number of rolls | 1 | 2 | 3 | 1 | 2 | 3 | 1 | 2 | 3 |
| Farm door lock (DC10): | 64% | 81% | 92% | 81% | 96% | 99% | 100% | ||
| Standard lock (DC15) | 25% | 44% | 58% | 42% | 67% | 81% | 100% | ||
| Merchant’s lock (DC20) | 6% | 12% | 18% | 16% | 29% | 41% | 64% | 87% | 95% |
| Ancient dwarven vault lock (DC25) | 0% | 2% | 4% | 7% | 30% | 51% | 66% | ||
| Kagyar’s master inner sanctum (DC30) | 0% | 0% | 9% | 17% | 25% | ||||
As a reminder, here’s the baseline with neither advantage nor disadvantage:
| Straight | Ramon (+4) | Jinnie (+7) | LLuisa (+15) | ||||||
| Number of rolls | 1 | 2 | 3 | 1 | 2 | 3 | 1 | 2 | 3 |
| Farm door lock (DC10): | 75% | 94% | 98% | 90% | 99% | 100% | |||
| Standard lock (DC15) | 50% | 75% | 88% | 65% | 88% | 96% | 100% | ||
| Merchant’s lock (DC20) | 25% | 44% | 58% | 40% | 64% | 78% | 80% | 96% | 99% |
| Ancient dwarven vault lock (DC25) | 0% | 15% | 28% | 39% | 55% | 80% | 91% | ||
| Kagyar’s master inner sanctum (DC30) | 0% | 0% | 30% | 51% | 66% | ||||
So here’s how much difference it makes:
| Difference with Disadvantage | Ramon (+4) | Jinnie (+7) | LLuisa (+15) | ||||||
| Number of rolls | 1 | 2 | 3 | 1 | 2 | 3 | 1 | 2 | 3 |
| Farm door lock (DC10): | -11% | -13% | -6% | -9% | -3% | – | |||
| Standard lock (DC15) | -25% | -31% | -30% | -23% | -21% | -15% | – | ||
| Merchant’s lock (DC20) | -19% | -32% | -40% | -24% | -35% | -37% | -16% | -9% | -4% |
| Ancient dwarven vault lock (DC25) | 0% | -13% | -24% | -32% | -25% | -29% | -25% | ||
| Kagyar’s master inner sanctum (DC30) | 0% | 0% | -21% | -34% | -41% | ||||
Well, that makes a difference! And makes more of a difference as the lock gets harder. That’s a nice tactical choice – do I try to open it in a hurry, and accept disadvantage, or do I spend the time? That’s the sort of thing which lets a player really engage with the situation.
I like it.
Delay based on roll
Another approach would be to let the difference between the roll and the DC determine how long it takes – with higher scores indicating a quicker pick.
I could work out some numbers…but to be honest, it feels like there’s too much maths already to really work at the table, and this is definitely veering into the “multiple interlocked calculations.” Even if this gives a realistic sense, I think it’s too impractical to be fun.
Conclusions
Time to decide. What do I feel is the right balance between cost and playability?
I like the d4+1 to determine how long it takes, with the option of speeding up at disadvantage.
I also like the thought of each attempt taking longer – d4+1 rounds for the first one, d4+1 minutes for the second, an additional 10-minute turn for the third attempt. If no success in the first three attempts, it’s beyond the character, and they need more training (i.e. go up a level) before they can try again.
What is the roll? Do I introduce an instant failure?
Looking at the various probabilities, my feeling is that the difference isn’t enough to be worth the extra complication at the table, given that the number of attempts is now capped at three, and each one is more costly in time than the previous one.
So, after all that experimentation with different failure options, I’ll not use any of them, but just let the time be the cost, and add a cap on the number of attempts.
Which, given that this all started because it felt like a lock-pick was too quick and easy, is probably appropriate – it directly addresses my concern.
So, my final proposal:
- A character needs proficiency with Thieves Tools to be able to pick a lock
- The first attempt takes d4+1 rounds
- The second attempt takes d4+1 minutes
- The third attempt takes a full 10-minute turn
- The character can make a faster attempt – one round for the first attempt, d4+1 rounds for the second attempt, etc – with Disadvantage on the roll
- If they still haven’t succeeded after three attempts, it’s beyond the character, and they need more training (i.e. to go up a level) before they can try again
Final questions…
What about trapped locks – when do they trigger?
That one is easy – it’s almost always listed with the description of the lock.
What about attempting to bash the door down?
No skill required – the standard rules and consequences can apply. An attempt takes an action, causes a lot of noise, and can hurt the person attempting. But they can keep trying, so unless it’s a really tough door and/or they’re very weak, it’s likely they’ll eventually get through…with every other character and creature within a fairly large radius aware the party is there and bashing down the door.
Now to try it at the table and see how it plays out for real.
Thank you to my awesome supporters on Patreon who help fund the research which goes into posts like this. If you like what I’m doing, please consider supporting me, too, or buy me a Kofi. I only charge on months when I put out a detailed post or two smaller posts.
