Try The Window.

Before you fix a weakness you have to know it exists. Below is the full map of the entry points that routinely get ignored, rated by how easily they're defeated without tools, noise, or skill — because that is the realistic threat, not a professional with a pick set. If you want the profile of who actually does this and how they choose a house, that's a separate read: how burglars choose targets. Five pips is worst.

This is the chapter that separates this guide from the other four hundred window-security articles online, and it's because of a bias in who writes them. General security writers evaluate window hardware by what it's called. Somebody who spends their evenings taking locks apart evaluates it by what's inside. Those produce very different answers.

A latch is not a lock, and the industry knows it

Start with vocabulary, because vocabulary is doing a lot of quiet work here. A latch holds a door or sash in position. A lock resists a person who is actively trying to defeat it. The crescent cam on your double-hung window is a latch. The hook on your slider is a latch. Neither is sold or rated as a security device, and if you read the packaging closely you'll notice the manufacturers are careful never to say they are. Manufacturer

The distinction isn't pedantic. It tells you where to spend. Reinforcing a latch is pointless — you can't make a stay into a barrier. What you can do is add a second, independent mechanical stop that has to be defeated separately. That's why every good fix in this article is an addition rather than an upgrade.

Wafer cylinders: the upgrade that isn't

Here's the one most homeowners get wrong, and it costs them money. You notice the flip latch is flimsy, so you buy a keyed window lock. It has a key. Keys mean security. Problem solved.

Except most low-cost keyed window locks, cam locks, and auxiliary track locks don't contain a pin-tumbler cylinder. They contain a wafer cylinder — a simpler mechanism using flat wafers instead of stacked pin pairs, usually four or five of them, with tolerances loose enough that a poorly cut key still turns. Community That tolerance is the entire point of the design: it makes the lock cheap to manufacture and forgiving to use. It also makes it trivially raked, jiggled, or opened with a tensioned blank. Wafer cylinders are what you find in filing cabinets, desk drawers, and vending machines — furniture-grade hardware. They exist to establish that a compartment is closed, not to resist attack.

So the keyed window lock you bought may genuinely be an improvement over the crescent cam — but not because of the key. It's an improvement because it adds a second physical stop. The cylinder is close to decorative. If you understand that, you'll stop paying extra for keyed versions and start buying whatever produces the strongest mechanical interference for the money, which is usually a plain pin or bolt.

Spring latches and the credit card problem

Any latch with an angled face — the sloped bolt that lets a door click shut without turning the handle — can potentially be pushed back by sliding something thin against that slope. The technique is called loiding, it's older than most of the houses it works on, and it's the reason a spring latch alone is not a lock. It applies to casement window latches, interior doors, and above all to the garage passage door we'll get to in Chapter 4.

A deadbolt is immune to this — the bolt is square-faced and doesn't retract under side pressure. That single mechanical difference is the whole argument for putting a deadbolt anywhere a spring latch is currently doing security work. Padlocks have their own version of this problem, which is shimming, and it's worth knowing about if you're using one on a gate or shed.

The screws are the lock

The most expensive lock in the world is fastened to something. On a typical American house, the strike plate is fastened to a jamb of finger-jointed pine roughly three-quarters of an inch thick, using the screws that came in the box — also three-quarters of an inch. Those screws terminate in the jamb. They never reach the stud behind it.

So when a door is kicked, the lock does not fail. The bolt does not shear. A quarter-inch strip of softwood splinters, and the whole assembly leaves with it. You can watch this in any bodycam footage of a forced entry: the hardware is usually still intact, dangling from a door that's now in the hallway. Documented

Three-inch screws through the strike plate and into the framing convert that failure mode into something entirely different, and the parts cost about four dollars. This is the single highest-leverage security purchase available to a homeowner, and almost nobody makes it. It's covered properly in the door frame reinforcement guide.

If your house has one, this is your weakest opening, and the dowel you put in the track has probably given you false confidence for years. Three things fail here and they fail independently, which means fixing one leaves the other two untouched.

Failure mode 1 — the hook latch

Factory sliders use a hook that swings into a slot in the frame. It engages perhaps a quarter inch into thin extruded aluminum. Leverage at the meeting stile deforms the aluminum before the hook itself gives, and the hook clears the slot. On doors more than twenty years old the frame is often soft enough that this needs no tools at all — just body weight applied in the right direction.

Failure mode 2 — lift-off

This is the one people miss. Sliding panels are installed by lifting them into the upper channel and dropping them into the lower track. That process is reversible from outside. Lift the panel, swing the bottom clear, and set it aside — the latch is still perfectly engaged and completely irrelevant, because the door is no longer in the frame.

A dowel in the track does absolutely nothing about this. It prevents horizontal travel. Lifting is vertical. If the bar is the only thing you've done, you've addressed one of three failure modes and left the quietest one wide open.

The fix is embarrassingly cheap: drive two or three screws partway into the upper track above the sliding panel, leaving just enough clearance for the door to run. The panel can no longer be lifted high enough to clear the bottom track. Cost, roughly nothing. Time, ten minutes.

Failure mode 3 — the track bar itself

Cut-to-length track bars are usually sized generously so they'll fit a range of doors. That generosity leaves play — often an inch or two — and an inch of travel at the latch edge is frequently enough to work the hook free or get a tool into the gap. A bar should be cut so the door moves essentially not at all. If it rattles, it isn't finished.

Better than a bar: a foot bolt or auxiliary pin lock drilled through the sliding panel into the fixed panel or frame. That's a steel pin passing through both members. It defeats sliding and lifting simultaneously, and it doesn't care how the latch is doing. Deeper treatment in the dedicated sliding door guide.

Everything about the garage inverts normal security logic. Elsewhere in the house, being seen is the deterrent. In a garage with the door down, an intruder has walls, a roof, no sightlines, and as much time as they want. The garage isn't an entry point so much as a staging area.

The emergency release cord

Every automatic opener has a manual release — the red cord with the plastic handle, required so you can get your car out during a power failure. The bypass is not new and it is not obscure. A straightened coat hanger goes through the gap at the top of the door, hooks the release lever or its cord, and pulls. The trolley disengages and the door lifts by hand. It was demonstrated at an industry trade show in 2002, and the door-industry association itself has published on it. DASMA There is at least one granted US patent whose entire purpose is defeating this attack, which tells you the industry considers it real. Patent

Mitigations, in order of how much they inconvenience you:

• Zip-tie the release handle to the carriage. A hard pull from inside snaps a light-duty tie; a fished wire can't generate that force. Reversible, costs pennies. Read this part: opener manufacturers discourage modifying the release, and UL requires the cord to stay reachable from the floor — it exists so you can get out during a power failure. Manufacturer Use a tie light enough to break by hand, never wire, and test it before you trust it. If anyone in the house couldn't snap it, don't fit it.
• Shorten the cord so it can't be reached from the top of the door.
• Fit a shield over the top of the door where the wire goes in.
• If the garage is detached or rarely used, a slide bolt or a good padlock through the track defeats the whole category — at the cost of not being able to open it remotely.

The passage door is the actual problem

Once someone is in the garage, only the interior passage door stands between them and the house. Builders classify that door as interior, so it typically gets: a hollow-core slab, a spring latch with no deadbolt, short screws, and sometimes a bright window at head height.

Every weakness in this article converges on that one door. It's loidable because it's a spring latch. It's kickable because the screws don't reach framing. It's punchable because the core is cardboard honeycomb. And the attacker has complete privacy while working on it.

Treat it as an exterior door, because functionally that's what it is: solid core, a real deadbolt, three-inch screws, reinforced strike. Under a hundred dollars if you hang it yourself, and it is the best money in this entire guide.

Opener codes

Openers made before the late 1990s often use fixed codes, which can be captured and replayed. Documented Modern units use rolling codes that change every activation and are not vulnerable to simple replay. If your opener predates rolling code, replacement is the fix — and it's a good moment to note the same principle governs smart locks: the mechanical side and the electronic side are two separate attack surfaces, and a product can be excellent at one and negligent at the other.

And the ordinary version of this, which matters more than any of it: the remote clipped to your visor is a house key sitting in an unlocked car. Bring it inside, or use a keypad.

Ordered by cost, not by importance — because the useful question is never "what's the best possible setup," it's "what's the most security I can buy with what I have right now." Start at the top. Most homes get the majority of their available improvement before spending a dollar.

Half of window security advice online is repeated because it sounds sensible, not because anyone checked. Here's what doesn't survive contact with the mechanism.

Worth saying plainly: breaking glass is the option people worry about most and use least. It's loud, it's slow, it leaves an obvious signature, and it cuts you. Given a latch that opens quietly in four seconds, nobody picks the noise. That's why the bulk of this guide is about latches.

But it happens, and the glass you have determines what happens next.

Standard float glass

Older single-pane windows. Breaks into large shards, leaves a clear opening. No meaningful resistance.

Tempered glass

Required by code in doors and near floors. Roughly four to five times the strength of annealed glass — and this is where people go wrong, because that strength is real but it is strength against accidental impact. Once the surface compression is breached the entire pane dices at once into small blunt granules, leaving a completely clear opening. Documented Tempering is a safety specification written to prevent lacerations. It is not a security specification, and glazing-industry sources are explicit that it offers essentially no forced-entry resistance once broken. Stronger before it breaks, more open after.

Laminated glass

Two panes bonded to an interlayer — the windshield principle. It cracks but holds together, and clearing an opening requires sustained, loud, visible work. This is the only glass that meaningfully resists entry. Expensive to retrofit, which is why it makes sense on two or three openings rather than all of them.

Security film on existing glass

The retrofit approximation of laminated glass. Quality matters enormously and the market is full of tint marketed as security film. Look for a stated thickness — generally 8 mil and up for security grade — and, critically, a mounting system that attaches the film to the frame, not just the pane. Film bonded only to the glass can be pushed out as one flexible sheet, which is a well-known limitation of DIY installs.

Glass break sensors

Two kinds. Acoustic sensors listen for the specific frequency signature of breaking glass and cover a room; shock sensors mount to the pane and detect impact. Acoustic is more common and less prone to false alarms from a slammed door. Neither prevents anything — they shorten the time between the break and someone knowing about it, which is a different and still useful job.

Nothing in this guide is exotic. There's no product here that will surprise a locksmith and no technique that isn't decades old. What's unusual is only the accounting: looking at the whole perimeter at once instead of the one opening that gets all the marketing.

The front door gets the attention because it's the door you use. It's the one you see every day, the one guests arrive at, the one the camera points at. The bedroom slider is background. The basement hopper is invisible. The garage passage door isn't even thought of as an exterior door — but it is one, and it's the flimsiest one you own.

An intruder isn't choosing between your openings on aesthetics. They're choosing on time, noise, and visibility, and by those three criteria your front door is usually the worst option available to them. Which means the effort you've put into it has, in a sense, worked perfectly — it moved the problem somewhere else. The whole job now is noticing where.

So go outside tonight, walk the perimeter, and look at your house as a stranger would. Then start at the free rung and work down. You'll get most of the available improvement before you spend anything at all.

Try the window. Then fix it.