I have a space for computers secured with a simple deadbolt. Someone keeps coming to pick the lock. While working there, I have scared them away three times.

There are cameras, but not in useful places or all exits and the building manager won't let me run wires for more. I contacted the police, but maybe it is a low priority for them.

With each subsequent visit, does a lock picker gain further progress in undermining the door's ability to keep out? Could they be doing something to the door each time that is getting them closer to being able to open the lock really quickly? Is there anything I can do to stop their ability to pick the lock?

  • 14
    For what it's worth: Cameras can also be placed without wires. Commented Nov 25, 2014 at 9:44
  • 36
    @MikePennington Just from reading that he failed three times and is still coming back, I think there is a chance that we are not exactly dealing with a genius. Commented Nov 25, 2014 at 11:31
  • 14
    Are you sure, they did not already entered that door in an unobserved time? And now come back to read out the HW-Keylogger for example?
    – Marcel
    Commented Nov 25, 2014 at 13:15
  • 7
    @Village I'm confused. How have you scared them away but not been able to identify them? Do you mean you hear a noise on the other side of the door but when you open it, they are gone? Please describe how you scare them away. This might be all in your head...
    – OpenCoderX
    Commented Nov 25, 2014 at 21:07
  • 11
    Has anyone mentioned the importance of logging (a) the attempts to pick the lock and (b) the efforts to get the building manager to add cameras? That way, if the picker eventually gets through, you can prove who's really at fault, which will at least protect your job. Commented Nov 26, 2014 at 17:08

6 Answers 6


Yes, there's a classic attack that involves incremental access.

The attacker starts out with a blank key that fits into the lock in question.

The attacker approaches the door, puts the key in, jiggles the key a bit, grumbles something about how the office numbers changing, and leaves.

Then in private he examines the impression pattern on the key. Where there's evidence that the pins were bound, he files the key down a bit. Every day he visits the door with his increasingly-filed-down key, and every day he progressively files it down a bit more, using the impression pattern in the key as his guide.

Then, one day, he'll have filed the key to match all of the pins, and the door will open.

This attack has the advantage that it doesn't look like an attack. It just looks like a lost tenant who briefly visits the wrong door, and then leaves once he's realized his mistake. And when he's done, he'll have a working key.

  • 8
    Most door locks in the US are pin tumbler locks. A pin tumbler lock uses a set of pins that are split at a differing places along their length. When the correct key is inserted, the split points all align at the shear line of the rotating plug and the lock can be opened. The problem with the 'classic' attack described is that ALL the pins typically apply EQUAL pressure on a key blank as all the pins are being driving into the key with similar springs. See en.wikipedia.org/wiki/Pin_tumbler_lock I'd like to see a reference or link, if you have one, of this attack being used. Commented Nov 26, 2014 at 19:28
  • 4
    @MrWonderful not sure whether this is the sort of thing you are after? youtube.com/watch?v=Bj9KEmLWRek
    – R15
    Commented Nov 26, 2014 at 20:23
  • Will any of the soot be left visible on the lock?
    – Village
    Commented Nov 27, 2014 at 23:24
  • 3
    @Village while the technique I learned used soot to enhance the markings made by the pins, watching the videos you can see that today this is done using a normal clean key, with magnification to see the scratches in the metal. So no sign of tampering needs to be present.
    – tylerl
    Commented Nov 28, 2014 at 2:20
  • 6
    Why would anyone do this when a simple bump-key opens any tumbler lock in 5 seconds flat? Commented Nov 28, 2014 at 3:37

The answer to your question is yes, though whether this will ever actually help them is dependent on the lock and their 'skill'.

With a typical (cylinder?) deadbolt repeated attempts can advance an impressioning attack with a key blank (see tylerl's answer for more detail), if using picking tools the extra visits will improve the feel for the lock and in theory will make the time to pick it quicker, though it should be noted that someone proficient at picking locks is unlikely to need multiple attempts against a simple deadbolt lock.

Lever tumbler locks, as used in some mortise locks, are more time consuming to pick, so fitting one of these (at least 5 levers is generally accepted as necessary) might be a good option. Multiple attempts against this type of lock is likely necessary to develop a successful attack against it so it will give you more opportunities to catch someone.

Restricting access to the lock so that it is more difficult to use lock-picking tools may be effective depending on the lock and installation, generally speaking a key needs less space to operate than lock-picking tools.

But if someone is determined it does not matter what lock you fit, with time and patience any lock can be defeated (eventually)...

Depending on the building layout the door could be bolted from the inside out of hours, so that the only entrances available to the attacker are ones that are covered by the existing camera installation.

It a mortise lock does not discourage them and you cannot bolt from the inside, the refusal to install an additional camera is an implied acceptance of the risk by the building manager - provided that person is authorised to accept the risk there is little more that can be done.

And to add, since this is an information security site, the computers that are behind this potentially insecure door, which is subject to a an observed threat should:

  • Have strong passwords configured (including BIOS) and never be left logged in;
  • Disable USB interfaces etc to prevent introduction of malware;
  • Be configured with full disk encryption so that if the attacker gets in and steals any machines they are of hardware value only;
  • Store all your backups and removable media somewhere else;
  • Conduct daily checks to make sure unauthorised hardware has not been installed (packet sniffers, key loggers etc).
  • 6
    Forget being defeated through lockpicking, if I considered the information valuable enough I'd just buy (make?) a batch of explosives. Or, depending on the door, a chainsaw. It'd be really obvious (and likely to alert somebody), but ideally I could just grab the machine and run within a minute. Commented Nov 25, 2014 at 9:12
  • 11
    @Clockwork-Muse a well placed foot will get you into many buildings, for everything else there is the rubber hose to the security guard method Commented Nov 25, 2014 at 11:24
  • 44
    @Clockwork-Muse This isn't the movies. Using explosives would seriously increase the penalties for getting caught. It would also seriously increase the chance of getting caught by increasing the level of law enforcement interest from "Yeah, whatever" to "Possible terrorist." Commented Nov 25, 2014 at 15:43
  • I forgot about impressioning in the original answer...now added. I have not included much detail because tylerl posted the detail before I had something drafted up.
    – R15
    Commented Nov 26, 2014 at 10:25
  • When you say Mortise lock, do you really mean something more like a lever tumbler lock? I could be wrong, but it seems Mortise is more about the form factor than the type of actual lock and key.
    – user
    Commented Nov 27, 2014 at 14:15

You choice of lock matters a surprising amount.

There exist locks which have not been defeated through "covert" mechanisms (picking, pick guns, etc.) in the open literature. Abloy's disc detainer locks, and one other type (I think it was a plastic lock from a subsidiary of Kaba?) are two such locks. Replacing the lock may be a suitable recourse, if you're willing to splash out a fair bit per-lock.

Some locks which have not yet been picked in the open literature are simple to defeat -- but it's extremely noticeable. It sounds to me that this isn't suitable for you. Good disc detainer locks are known for being hard to compromise in a short time.

To even think about picking a decent disc detainer lock, you need something beyond the bog-standard torsion wrench & hook.

Other than that, there's a huge range of questions to ask.

Why is the person picking the door lock? Is this an outer door on your premises, or is this the lock to the data centre? If so, how did they get so far in? Why isn't the attacker smashing the door in? Could they have already breached the premises, and be attempting to recover something (e.g. keylogger, packet sniffer, etc.)? How good's your alarm system inside the door? Is it set when you leave? How fast do your security team react when this alarm goes? Can the attacker be in and out prior to apprehension? Have you tested this? Are there any other entry points which could be a risk? Since your threat model clearly involves people willing to pick locks, you need to consider things like lifting suspended ceiling tiles and going over doors and the like. Can you set up to get a positive identification of the attacker, assuming it's the same one each time? Can you set an alarm on the place where the attacker is trying to attack from -- that is, detect them before they reach the door? If it's a data centre, is the hardware in locked cages, or "open"? Can you put multiple locks on the door (e.g. a lever-tumbler with relocking mechanism along side a pin-tumbler lock, or other more "difficult" lock -- sidebars seem to be popular, but be careful, some are known as regional sidebars, and can be considered as "public knowledge" to some degree)

For the most part, your security system seems to be working. You know how long your response takes, you've notified local law enforcement, the lock is holding up long enough for you to respond adequately.

If you repeatedly pick a lock, it often does get easier, as you begin to get a feel for that particular lock (e.g. for a pin-tumbler, you might note that pin 1 sets first, then pin 4. Pin 3 is a security pin) etc. so you can get faster at a particular lock, but it is heavily dependent on the lock type & skill level of the attacker. As such, you will want to start getting a positive identification of this attack and denying them access to the lock, or getting them thrown in jail. If this is not an option, you may be stuck changing or re-pinning the lock periodically (e.g. after X detected attempts, select X wisely), which is costly.

A slightly more expensive lock with a re-locking mechanism might be good deterrent. Chubb-style lever locks are cheap in comparison to good disc detainers, and often alert you to the fact someone tried to pick it when you next try to unlock it, as their re-locker mechanism will have triggered, and the lock won't open without a special key.

If you do go down the line of (something like) a disc detainer lock, do you research. Many of the knock-off, cheap versions of the disc detainer lock are susceptible to attacks which are both quick and simple. Further, they don't hold up well to physical attacks (drilling is the primary mechanism for door-mounted locks).

Speak to a qualified locksmith, a good one should be quite knowledgeable about this, and should be able to help you make a suitable choice of lock.

  • 2
    "...the lock is holding up long enough for you to respond adequately." This is the only reason the change the lock, if there's a need at all. If you don't have an alert when the door is opened, and you only check on it twice a day, then you need to use a lock that will keep a determined picker out for 12 hours. If you can't change the lock, then you have to change the response time - and a wireless door chime might reduce response time enough if cameras aren't an option.
    – Adam Davis
    Commented Nov 25, 2014 at 15:29
  • 1
    That's not entirely true, as with certain mechanisms, knowledge about the construction of each particular lock lock is very useful (e.g. working out where security pins are). This would reduce the attackers time required to compromise the lock. Ergo, your response times may become inadequate at some point in the future -- unless you have entry alarms, and a good response time for an unauthorised entry. Commented Nov 25, 2014 at 16:36
  • See here for a video for picking Abloy Protec 2 (which is the best disk detainer lock there is): youtu.be/6zVSJ_wauwg – those can be picked open but there are not yet off-the-self tools for the pick tip. I know that Tarehjerne and Huxleypig can pick Protec 2 but they haven't show their tip design publicly because Abloy doesn't have more recent core design available. For an off-the-shelf tool that can pick many disk detainer locks, see this: youtu.be/DuZWhRaLzhk Commented Jun 19, 2021 at 21:25

As a former locksmith, I can provide a qualified "no" for you. It is not possible to incrementally pick a lock, especially if it is a standard pin-tumbler variety deadbolt (Kwikset, Schlage, Baldwin, etc.).

One thing you can do to assure that the lock is 'reset' is to stick the key in it, then removed it - turning it isn't necessary. This will force any stuck pins to be reset to their default position. They would only become stuck if the lock is worn or is particularly shoddy anyway. Either way, it's very unlikely.

To address another person's comment, mortise locks tend to use the same type of pin-tumbler mechanics, and while their higher precision manufacturing and mushroom shaped top-pins can make it more difficult to pick, they operate using the same principle as the lower-end deadbolts you can buy and any hardware store, and can be bypassed by a well-practiced lock picker.

It's entirely possible to buy a higher-end lock that will thwart most picking attempts. Medeco, Assa, and some other manufacturers make them. If you install one of those, your main concern becomes thwarting a forced entry, not lock picking.

  • 1
    What about stuff like this though?
    – Seth
    Commented Nov 27, 2014 at 23:52
  • 4
    They call that impressioning a key, and it is very, very difficult to do. I've tried it many times, was taught by someone who could do it, and have rarely been successful at it. In other words, it's very unlikely to work.
    – willc
    Commented Nov 29, 2014 at 4:14
  • Unfortunately, being a locksmith doesn't make you expert on picking locks. The covert attack to any pin thumbler lock that you can frequently access e.g. 5 seconds a time is "key impressioning" attack. See here for an example against a high quality lock: youtu.be/p0euhqTq_9o – also note that some people claim to be able to do this attack against Abloy disk detainer locks, too, but I haven't found any reasonable demonstration about that. Commented Jun 19, 2021 at 21:31

Your deadbolt's security does not seem nearly as important as the lack of monitoring on the door. Even the most braindead criminal will eventually realise the value of a bolt cutter, at which point the whole discussion becomes moot.

If a particular layer of your security (usually the outer-most) is vulnerable when you can't afford for it to be you have two options: deter an attacker from attempting an attack on it or add another layer.


Yes. Continuing attempts at a lock do increase a lockpicker's chances of actually picking it. With each try they get a better feel for for how to place the anchors and rakes.

If you're looking to deter the intruder you can set up a current with maybe 3 or more D cell batteries and (be careful about how many you use, you don't want to seriously injure anyone) run it through the lock, you can add plastic or fiberglass to the end of your key to keep it from affecting you and a switch inside the room for ease of disarming.

  • 3
    1) edited your second answer into this. 2) I doubt that the current from the batteries would do anything other than run through the lock itself. The path of least resistance would not be through a person at the end of a stick.
    – Jeff Ferland
    Commented Nov 26, 2014 at 2:13
  • 4
    You need more than the 4.5V you get from 3 D cells to deliver a shock that could be felt. Electric fences and cattle prods operate at a much higher voltage, but are carefully designed to not deliver a dangerous amount of energy to the victim. That said, electrifying the knob is not a sane way to go, and is almost certain to create liability if it is not outright illegal.
    – RBerteig
    Commented Nov 27, 2014 at 0:34

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .