Technical justification of stubbornness
I have a younger daughter. If she wants something, she goes for it without seeing obstacles. Let me phrase that carefully: not overcoming obstacles — literally not seeing them. Those are two different engineering modes. Overcoming implies you detected the obstacle, assessed it, and budgeted resources for it. Sanya works differently: an obstacle isn’t an object to her, it’s noise on the channel.
I spent a long time wondering where this comes from. My working theory: video games. A normal person, having played a level four times and understood its logic, goes looking for cheat codes or turns the whole thing off. Sanya is only just warming up at that point. She can run the same sequence thirty times in a row when the level’s logic has been obvious for ages and the only thing missing is motor precision. And on the thirty-first attempt — she clears it.
In the virtual world the price of this trait is time. In the real world the price is different, and it’s usually paid by somebody else.
Sanya channelled her stubbornness into sport. You’ll ask which one. The most stubborn one available, obviously. People usually object here: every sport is about stubbornness. No. In most sports you run into a boring physical ceiling that training slowly pushes back, or you need a worthy opponent to drag you upward. Climbing is more discrete than that. A wall, holds, a route — that’s it. Your opponent is you. More precisely, three entities: you, your fingertips worn down to blood, and your dopamine system, which after every successfully sent route says sweetly, “how about that black one over there?”
Climbing doesn’t let you improve gradually. It gives you either a zero or a one. It’s a sport for people who don’t understand why anyone would skip a level.
From which it followed that sooner or later there would be a pull-up bar in this house.
Three lines of procurement
It started beautifully. I’d even say exemplary.
— Dad, let’s buy a pull-up bar. — Okay.
— Dad, look, do you like these bars? — Yes.
— Dad, I’ve picked a bar, can you give me the money and I’ll buy it? — Stop. Show me the bar.
That “stop” is the only thing in this entire story I managed to do genuinely on time. After that I was purely reacting.
— Fine, here’s the same bar but cheaper. Seventy euros. Ordered. Wait for it to arrive.
The word “wait,” in a conversation with Sanya, isn’t an instruction. It’s a sound. Waiting is not a form of activity for her. If a process requires curing time, she uses that time to improve something in parallel.
Three circumstances you need to know
First. At this point I’m not at home. I’m playing Mission Control: dispatching the children’s activity by phone and video. Commands go out by voice, telemetry arrives as a shaking camera, and the object’s orientation in space has to be reconstructed from indirect evidence.
Second. I’m on antidepressants. Which makes me calm as a boa constrictor. Occasionally tedious. That combination, incidentally, is exactly what Mission Control requires: don’t panic, and repeat the same key message as many times as necessary without changing your intonation.
Third, and most dangerous. Sanya had, remotely, just drilled the first holes of her life into a wall. Inserted her first wall plug. Screwed something to a wall. This inspired her enormously. A boundless world of new possibilities had opened up in front of her.
That third circumstance is like a flight clearance issued after one successful take-off. Formally everything checks out: holes drilled, plug holding, object mounted. Functionally: a human being has just discovered that a wall is not a barrier but a medium.
Finding the spot
— Dad, how do I hang the bar? — First, find a place to hang it. — Found it! Great spot. In the storage room, above the door. — Yeah, okay.
Some time later I call back.
— No. Not okay.
Above a door there’s most likely a concrete lintel. And our hammer drill is from Lidl, the cheapest one in existence. It will overheat before it gets through fifty-year-old concrete. We have stubbornness, we have stubbornness in abundance, but monitoring the thermal state of a hammer drill remotely is unreliable. The temperature sensor in this scheme is me, and the data link is video with symptoms of seasickness.
— “Most likely”? So it might not be there?
And that’s where I understood that our problem was not the lintel.
“Most likely” is my standard engineering hedge. A person who has seen a certain number of walls in their life doesn’t say “there’s concrete there.” They say “most likely concrete,” because a lintel can be timber, can be steel, and sometimes doesn’t exist at all — somebody laid rebar across the brickwork and forgot about it. To me, “most likely” is an honest declaration of residual uncertainty.
To Sanya, “most likely” is hope.
She heard my hedge as an open door. Not “there’s probably concrete there,” but “there’s a chance there’s nothing there, in which case this works.” That, broadly, is the difference between an engineer and a person with very strong motivation: the same sentence gets read in opposite directions.
I explain “most likely.” I advise: if you’re hanging it above the door at all, go 10–15 cm above where the lintel should be.
— Oh, that’s too high. And too close to the ceiling. Is the lintel definitely there?
Hope plus stubbornness. A combination against which probability theory is powerless.
I tell her: the house is old. In principle, after settlement and thermal cycling, the lintel should have separated from the brickwork. If you find a thin hairline crack right above the door, that’s the boundary of the lintel.
She found the crack.
I felt no triumph whatsoever. Triumph would require having predicted something non-trivial. All I’d said was that a residential doorway has a lintel over it. That’s roughly on the level of predicting that a chair has four legs.
— Too high. Let’s keep looking. Can I use the garage? — You can. Just watch the mounting position: it must not obstruct the circulation of people, objects, or my head. Which sits at 186 centimetres.
The garage turned out too low. Then came several more spots, each of which was “perfect!” and, ten minutes later, “no, can’t do it there.”
Anyway. We found a spot. A good one.
— How do I mount it?
The top row
— Show me the bar.
She does. Two mounting plates, four holes each, two per row.
Good manufacturer. Honest bar. And here begins the part that makes this whole story worth telling.
It never, ever occurs to anyone that the bottom row of holes carries no load at all.
None. It’s there for geometry, for stabilisation, for peace of mind. All the load goes into pulling the top row out of the wall: a person hangs, the moment rotates the plate away from the wall, the upper fixings work in pure tension, the lower ones get pressed into the wall and essentially rest.
And when the top row lets go — and it will go first, if anything goes — the bottom row inherits not a static load but a dynamic one. A human being falling. It fails in one second.
Top row first. Then, immediately, the bottom.
Which means all attention goes to an ABSOLUTELY RELIABLE fixing of the top row specifically. No compromises. Those four upper fixings aren’t “preferably good,” they’re “either yes, or we don’t hang it.”
I say this slowly and in full, because it is the one line in the spec that is not open to negotiation.
Marking out
— Hold the bar at the height you actually want it. Mark the first hole with a pencil. Then take the spirit level and draw a guide line for the top row — not across the whole wall, just enough to cover the region where those holes need to be. Bar-width. There’s a ruler on the level. Then do exactly the same for the bottom row, offset vertically by the hole spacing.
The marking came out perfect. All eight holes in the bar sat honestly on parallel and perpendicular lines — which, by the way, is not always a given: I’ve seen products where the bottom row was drilled by eye and matched the layout about as well as a horoscope matches a biography.
The calipers that don’t exist
— What diameter do I drill the wall holes?
The largest plug diameter the bar itself will accept. So first we need to know what the bar accepts.
— Take the calipers— — What are calipers? — Little pincer things with an LED display, black carbon fibre. — Which box? — The one that came out of the boiler room.
Pause. Rustling. The shaking camera shows me the floor, the ceiling, a box, the ceiling again.
— Nope, can’t find them. — Fine. Wait, I’ll send you plugs and you can measure against those.
The anchors
— Dad, there are some spare parts in the box!
She shows me eight attractive metal sleeve anchors. Genuinely attractive — shiny, hefty, with a nut. They look more reliable than anything you have ever held in your hands.
Absolutely not.
Anchors like these are excellent for fixing static things. A shelf. A bracket. Something that hangs and doesn’t move. A pull-up bar is dynamics: shock loading, swing, hundreds of cycles of alternating stress. A metal anchor transmits all of that straight into the metal of the expansion sleeve, the metal chews out the wall of the hole, and the anchor comes out. Top row first. Then, immediately, the bottom.
Also, tightening the nut on these things is a quest in its own right, one I wouldn’t wish on an enemy: either it spins the whole anchor or it doesn’t turn at all.
I explained the anchors. I mentioned that professionals on site often operate a blanket ban on them in scenarios like this — not because the anchor is bad, but because it’s extremely easy to use one in the wrong place.
— Right, take those eight anchors and put them in the bin… actually, hold on. What diameter are they? — How do I find out? — Take the calipe— ah, right. There should be a number stamped on the side. — Oh! Ten.
Excellent.
We never did find the instrument, but we solved the problem. That’s normal engineering practice: if a measurement can be replaced by reading a marking, read the marking. You just have to think of that before you send someone to look for calipers in a box from the boiler room.
Twenty euros
— I sent you two 10×80 plugs along with the drill. Check whether their screw passes through the hole in the bar. — It does! — Great. Those are what we’re using. — But I’ve only got two. — So you drive to Brico and buy more. — Could we use metal ones? — No.
And I explain again. About materials. About how a nylon sleeve isn’t a cheap substitute for steel — it’s a damper. It takes the shock, deforms slightly, distributes the load along the wall of the hole, and returns. Steel can’t do that: it’s rigid, it doesn’t absorb, it passes the impact straight into the wall material, and the wall material is the first thing to give up.
— Fine, off to Brico.
Call from Brico: — These plugs? — Yes. — Why are they so expensive?
Because three or four euros of that price is simply the fact that the thing is hanging on a shelf in a shop. Logistics, rent, nice packaging.
Total: twenty euros.
— Same as yours? — Yes. The plug is the part that matters. I don’t know what our walls are made of — clay blocks, brick, or sand-lime blocks. We need a plug that holds properly in both hollow and solid material. Which is why we’re not buying off the bottom shelf.
Houston
— Dad, I have a problem.
There is one sentence that injects adrenaline into a father’s bloodstream instantly, regardless of his age, temperament, or medication. This is it.
“Dad, I have a problem” is a construction with a lower bound and no upper bound. The lower bound is money. That’s the easiest, most merciful scenario, and you grab for it like a handrail. There is no upper bound at all, and a father’s brain sprints the entire range upward in a fraction of a second, and up there at the top it’s better not to look.
Even on Apollo 13 it was phrased more precisely. Swigert said: “Okay, Houston, we’ve had a problem here.” Houston asked him to repeat. And then Lovell gave the full version: “Houston, we’ve had a problem. We’ve had a main B bus undervolt.”
That second sentence has everything I was missing. First, the tense: we’ve had. The event has occurred, it is complete, what’s being assessed is a result — not “this is happening right now and nobody knows where it stops.” Second, a parameter: undervoltage on main bus B. Not “a problem,” but “this one, here, involving this.”
“Dad, I have a problem” is present tense and zero parameters. Pure alarm with no data attached. A sledgehammer to the nerves.
I asked her to phrase it differently. “Dad, question about a plug.” “Dad, off-nominal on a screw.” Anything where the first word carries information.
— I couldn’t get the screw all the way in and I stripped the head. Look. Now it won’t go either way.
I look. I see a stripped head. I see a screw. I see a wall.
Stop.
— Hey. Where’s the bar? — Oh, I’m screwing the screws in and out. So it’s easier to screw them in later.
This is where I learned something new about the world.
It’s a perfectly logical model — in a world where the thread is cut into metal and where “working a joint in” genuinely means something. The problem is that in a nylon plug the screw cuts its thread into a soft material once. Every subsequent insertion follows a track that’s already been chewed. “Working it in,” here, isn’t preparation. It’s wear.
— Who taught you to do that? — Nobody. I’ve always done it. You never told me not to.
Which is entirely true. I never did. There was no clause in the spec saying “do not exercise the thread for practice,” because it hadn’t occurred to me that it would occur to anyone. Every list of prohibitions is a museum of other people’s accidents, and it never contains an exhibit for something nobody has done yet.
Back to the imperative: the four upper holes must be certain, with no compromises.
And here the sunk cost effect kicks in — the same one that collapses bridges and blows through deadlines. A day already invested. Twenty euros already spent. It’s already almost up. And the brain starts negotiating: but what if it’s fine?
It isn’t fine. Pull the plugs, drive to Brico, buy new ones.
Yes, another twenty euros.
Torx
And then I stop.
— Hang on. Why a screwdriver head at all? The screw has a hex head for a spanner; the cross recess in the middle is just for starting it. Hey. What plugs did you actually— — You said those would be fine.
Yes, if they’re exactly those. A screwdriver recess on a 10×80 is frankly comic: the tightening torque is such that the driver will cam out long before the screw goes home. And if you are going with a driver bit, the recess should be Torx.
— What’s Torx? — A star, not a cross.
Offended, because someone has questioned a professional’s qualifications:
— Dad, why are you being so condescending? Torx is exactly what I bought.
— Okay. Well done.
Meanwhile I’m thinking: right, and what state is that bit in after all that sadism?
The score isn’t in my favour here. My hypothesis was elegant, tidy and wrong: she’d bought the correct plugs.
Hammer in reverse
— So: take pliers and slowly, mournfully, back the screw out. And next time you drive it home in one go.
Call: — It won’t come.
— Fine. Take a needle file. Grind a little off two opposite sides of the head — make two flats. Then pliers again. More contact area, less slipping.
Worked.
— Dad, I can’t get the plugs out. Listen, what if I use the bar? It’s basically a crowbar.
— No. Categorically no.
The bar is a lever over a metre long. The force at the far end is such that the plug will come out together with a cone of masonry, and instead of a clean hole we’ll have a crater. And we need that hole intact: those are our top four.
But the idea was right. Just applied from the wrong end.
Not leverage — inertia. The bar is sitting on the screw, and the joint has play. You accelerate the bar through that play and strike — not against the wall, but against the underside of the screw head, from the inside. A normal hammer drives things in; this one pulls them out. A slide hammer, an inertia puller, is a completely standard tool — it’s how you pull bearings and work out dents.
Worked.
And here’s the funny part: the exact dynamic loading we’d spent two days treating as the project’s principal enemy finally worked on our side. The impulse travels along the axis of the screw, the wall barely feels it, and the plug pops out.
A person who runs into a wall for the third time in one day starts seeing objects not by their function but by their properties. The bar stops being a pull-up bar and becomes a steel mass with holes in it. You cannot teach that in a lecture.
The spanner
Back to Brico. This time on the road bike, which was itself the product of a separate, self-contained fit of stubbornness — and that, believe me, is another story.
Call: — Dad, I have a problem.
Adrenaline. Zero parameters. Main bus B, Sanya. Main bus B.
— The screws won’t go all the way in. And I’ve bent the spanner.
— The hole in the wall is almost certainly too shallow. — It can’t be. I measured against the plug. — Hold the plug up against the screw.
Pause.
— Oh. The screw’s a bit longer.
Right. Measure twice, cut once.
— But I measured everything! I measured everything so carefully!! — Did you measure the depth of the holes? — …I didn’t know I had to. — Meten is weten. Measuring is knowing. You of all people know that one. — Dad, why are you being so condescending?
Here I don’t defend myself, because on the substance she’s right: the tone isn’t helping.
— Sanya, it happens. It happened to me too. It’s normal.
And that isn’t a pedagogical trick. The error “I measured the part but not the mating feature” is part of the standard kit; everyone makes it. The plug isn’t the final element — the screw goes deeper than the sleeve and needs clearance. You always drill deeper than the plug, not least because dust from drilling collects at the bottom.
Knowing this is not a sign of intelligence. It’s a sign that you have already bent one spanner.
— So what do I do now? — Take the bar down and drive to Brico for new plugs. Pull the old ones. — But it feels solid, and it’s only a couple of plugs that didn’t go all the way.
And I go again, in the same flat voice: top row, sequential failure, dynamic loading, required confidence in the upper four. Top first, then, one second later, the bottom.
— Okay. Fine. I’m going… Dad, how much did the bar cost? — Seventy. — And I’ve spent sixty on plugs. — Yes.
Silence on the channel.
— So I’ve basically bought a second bar in hardware. — You haven’t bought hardware. There’s twenty euros of hardware. The other forty is three round trips, one stripped head and one bent spanner.
And at that moment a person with a physics degree independently derived, without a textbook, the concept of transaction costs. In economics that’s what you pay not for the good itself but for the process of acquiring it: search, travel, negotiation, mistakes, repeat attempts. A quantity nobody ever puts in a budget, and which routinely catches up with the price of the object — and sometimes overtakes it.
Bar: seventy euros. Installing the bar: sixty. Of which twenty ends up actually in the wall. Forty went on tuition, which is, incidentally, still cheaper than any course.
But first — the tirade. About despair. About disappointment. About not being sure she picked the right degree at university (physics). About how if a person can’t hang a pull-up bar, what can they do.
I answer what I actually think: technical intuition doesn’t descend from the sky. It is earned through mistakes. That’s the only known currency. Right now you’re paying sixty euros and two days for something that will stay with you for forty years. Frankly, the exchange rate is excellent.
— Also my arms hurt. I’ve been at this bar for two days. Everything aches. — Well, there you go. That’s what the bar is for. You’ll build muscle.
The plug that spins
— Dad, I have a problem.
I’m sitting down. By this point I take a seated position in advance, before the comms session begins — not because of the adrenaline, but so the camera doesn’t make me motion-sick. That is not a figure of speech; it was an actual preparation procedure: sit, brace, put the phone on the table, do not hold it.
— The plug spins freely in the hole and I can’t drive the screw.
Right. After all that in-and-out, the hole has opened up a little.
— Not a problem. There’s acrylic-gypsum filler in there. Coat the plugs with it, push them into the holes, wait a couple of hours.
Then the standard protocol: where’s the filler, how do I apply it, how much, is this it, is this one it? All by video. All with the shaking camera.
Exactly two hours later:
— Dad, it didn’t work. — Then it hasn’t cured yet. — But you said two hours. — You need to keep waiting. — … — I’m hoping you haven’t touched the other screws? — No. — Then just wait.
She waited. Call:
— I’ve got it all screwed up, the bar’s holding, but that first plug is still spinning freely.
The post-flight review I conducted internally and did not broadcast: she rushed it. You don’t hang the bar until you’re sure every plug is holding. She should have reapplied filler to the first one and waited again.
But that isn’t us. We’re stubborn. We see the goal and we don’t see obstacles.
I didn’t say that out loud — not out of nobility but out of pure calculation. A spoken “I told you so” solves zero problems and adds one: an offended operator at the other end of the link. All right. Let’s solve it.
The collar
I call back with a solution:
— Push with the driver so the bottom of the plug bottoms out in the hole. It’ll jam on friction and you can drive the screw. — No. The plug has a collar. It won’t go any deeper.
Right. The collar. The flange.
And here I went full Rube Goldberg: how do you get rid of a collar when the bar is already on the wall and the collar sits between the bar and the wall? So, approach from the side. So, a blade. So, hold the plug still with the blade to stop it rotating while simultaneously operating the driver. So, three hands. So—
And then: oh.
The collar.
— Sanya. Just tighten the other bolts. The collar gets clamped between the bar and the wall. Friction does the rest.
Five minutes of conversation about why glue won’t do it: nylon doesn’t take adhesive, its surface energy is such that nothing sticks to it — that’s not a defect, that’s its physics. It has to be clamped by friction. And if clamping doesn’t do it, call me and we’ll keep solving.
The beauty of this solution isn’t that it’s clever. It’s actually stupid. The beauty is that the collar — the single component that had been in the way the entire time — turned out to be the exact feature you could grab the system by. I spent half an hour working out how to remove it. It didn’t need removing. It needed clamping.
That, if you want one, is the only moral I’m willing to defend: most of the time the element that’s in the way is an element you haven’t used yet.
The finish
— It’s up! It’s all screwed in! Look at the bar!
Joy. Celebration. The camera is shaking now not from uncertainty but from delight — an entirely different frequency, I’ve learned to tell them apart.
It’s up. It holds. Top row: four plugs, each in a hole of the correct depth, each on its own bed of acrylic-gypsum filler, each tightened once and all the way. Bottom row: just in case, for the geometry and for peace of mind.
Project totals: two days, three trips to Brico, sixty euros of plugs against a seventy-euro bar, one bent spanner, one stripped head, one needle file, one pull-up bar that moonlighted as a slide hammer, one road bike in the role of transport, a certain quantity of tirades about choice of degree, and one case of aching arms — for the treatment of which, by happy coincidence, there is now a suitable apparatus on the wall.
And somewhere back there, in the storage room above the door, a thin hairline crack remains — the boundary of the lintel that we very much hoped might not have been there.
The whole world in the palm of your hand — you’re happy and speechless, and only a little envious of those others whose summit is still ahead.
Hers is ahead. That black one over there, for instance.