A falsewrk cohort · for structural engineers, 0–5 years in
Be the engineer who can tell when the drawings are wrong.
There's no shortcut. This is the long way, organised.
Software is going to get faster than you. It's going to become more accurate than you, too. What's left? An independent check, one you can do in about 2 minutes, that tells you whether to believe the output on the drawing. This course is 12 weeks of the fundamentals, with a deadline and a group of peers doing it beside you.
Nobody got good at engineering by reading about it. They got there by attempting something difficult, failing, finding out where they were wrong, then repeating it 50 times. The right projects (if you were lucky to get them) used to give us this exposure over a decade. There is going to be less and less opportunity to get this through projects. It's time to hit the gym and start training.
Part 01 · Who it's for
Is this course right for you?
- 01 — The gut check
You can’t tell whether something looks right without running a detailed calculation.
Which means you have no preliminary assessment to offer in the room — only “I’ll check and come back to you”.
- 02 — The decay
You’re worried you might have forgotten the fundamentals you learnt years ago.
You feel like you should know this stuff, but it’s been so long since you actually used it.
- 03 — The skipped reps
You know the content, but you’ve never had the chance to practise it.
The job threw you straight into project management, generating finite element models, and skipped the reps where you actually build foundational understanding.
This is not for you if
- You’d rather watch than practise. There is supporting material every week, but most of what comes back is feedback on work you did yourself. The practice is the course.
- You’re starting from scratch. This assumes you’ve met most of these ideas before and let them go cold — rusty is who it is for. If structural mechanics itself is new, start somewhere more introductory.
- You want advanced analysis and complex structures. This is deliberately the foundation the rest of it sits on. More advanced topics may follow; they are not in these twelve weeks.
- You need code compliance or software training. This is about how structures behave and how you decide, not a clause or a package.
Part 02 · How it runs
A gym, not a lecture theatre
The shape of a week
- 01
Drillthe week’s problems, 10 minutes a day.
Look at a problem, commit to a number in about 30 seconds — listen to your gut, or do a quick hand calc — then find out how close you were. Do this a few dozen times to start building that intuition.
- 02
One challengetoo messy for one clean shortcut.
Takes about 10 minutes, and you answer it in one sitting. The problem is more complex than you can crunch in that time, so you use your intuition and hand checks to get as close as you can. Include your reasoning and say how you got there. Week 12 is the exception: allow four to five hours for the final challenge.
- 03
Commitby Sunday — your answer and the reasoning under it.
A deadline forces a decision. There is never enough information to answer it perfectly and you’ll need to make some assumptions.
- 04
Markedagainst named criteria, privately, before the next one lands.
Three things have to hold: your assumptions, your reasoning, and a number in a defensible range. You won’t get a score out of 10, but you’ll get feedback on each criterion. If a criterion is not met, more drills unlock the challenge again.
Four beats. The drills build your intuition, the challenge tests it, and the marking is the part you can’t do by yourself.
The drills are already in the wild
falsewrkPractise is the free engineering estimation app. It is also the foundation of the daily drills in this course. Here is what engineers who have used it have said.
My confidence in my wrong answer made me first question falsewrk. Then, after thinking about it, I questioned my confidence in the problem type and adjusted my belief in my own initial assumptions.
Joshua Peters
I’m sure some engineers don’t need it. Maybe they have their own way of studying, or actually get great mentoring. That wasn’t my experience.
Palisa Huoth
I was using it to confirm member capacities while doing a review for a temporary works package. Being able to get section capacity straight out of the steel member designation surprised me.
Jack Lin
Part 03 · The twelve weeks
What the course covers
Twelve weeks, each one a topic, a week of drills and a challenge. Open any week for what it covers and what you’ll do.
A week before week 1 — 25 January 2027 — the app, the Discord space and the material open, with nothing due. You get a short onboarding quiz then too: a few questions about what you want out of the twelve weeks, so I know what the room came for before week 1 lands.
Week 01Load paths and idealisation
The analysis starts with a decision nobody writes down.
Released Mon 1 Febdue Sun 7 Feb
Pinned or fixed? Beam or plate? One-way or two? How you idealise the structure does more to the result than any input after it. Load goes where the stiffness is, and the idealisation is where you told it to go.
You'll doDrill the idealisation on a dozen structures, a couple of minutes each, then mark the load path through a structure you have not seen before.
Week 02Estimating loads
Total a floor in your head, to within twenty per cent.
Released Mon 8 Febdue Sun 14 Feb
The numbers a senior engineer never writes down but has in their head: material unit weights, what the finishes add, what the live load honestly is. Get the load wrong by a factor and nothing downstream can save you.
You'll doDrill unit weights and area loads until you stop looking them up, then estimate the load for a whole structure.
Week 03Equilibrium and free bodies
Almost every error in practice is a free body drawn wrong.
Released Mon 15 Febdue Sun 21 Feb
If the forces do not sum to zero, you have a dynamics problem, and structural engineers do not like dynamics. Three equations that work on anything, including the structure nobody has modelled — and a force left off, a reaction assumed, a cut that was not closed is where the mistakes actually live, not in the arithmetic.
You'll doCut free bodies out of real structures until it is reflex, then find the force in a member with nothing but equilibrium.
Week 04Determinacy and stability
Count before you calculate.
Released Mon 22 Febdue Sun 28 Feb
Before any analysis runs, count the restraints and the degrees of freedom. It is the first test of whether the structure can stand up at all — and a structure can pass the count and still be a mechanism.
You'll doClassify determinate or indeterminate on sight, then find the instability in a scheme and the redundancy you would add to fix it.
Week 05Shear and moment diagrams
Draw the diagram before the software does.
Released Mon 1 Mardue Sun 7 Mar
wL²/8 and its half-dozen cousins, and the shapes that belong to them. Sketch the diagram first and the output becomes something you check; sketch it after and the output becomes something you receive.
You'll doDrill the standard cases until the shapes arrive without thinking, then sketch a complex frame and compare it against the analysis.
Week 06Sections and bending stress
Where the lever arm actually comes from.
Released Mon 8 Mardue Sun 14 Mar
Second moment of area, section modulus, and why depth buys so much more than width. This is the step where a moment becomes a stress and the member either works or does not.
You'll doDrill section properties by hand and by approximation, then determine stresses in members with non-trivial shapes and check them against the real values.
Week 07Deflection
Sketch the shape first. The number comes second.
Released Mon 15 Mardue Sun 21 Mar
The deflected shape is a drawing, not a calculation — and if you can draw it you can see immediately when the model’s version is wrong. The magnitude then follows from a few coefficients you can carry in your head.
You'll doSketch deflected shapes for frames and continuous beams, estimate the magnitudes, then check both.
Week 08Indeterminate structures
Stiffness decides who carries it.
Released Mon 22 Mardue Sun 28 Mar
Where load goes when it has more than one route — and why the member you stiffened to fix a deflection now has moment you did not design it for. Redistribution by feel, not by matrix.
You'll doMove moments around continuous structures by hand, then predict what a change will do before you make it in the model.
Week 09Influence lines and moving loads
The worst case is not standing where you left it.
Released Mon 29 Mardue Sun 4 Apr
Where to put the load to make it hurt. An influence line finds the critical case; running the model a hundred times and hoping one of them was it does not.
You'll doBuild influence lines by feel, then place a moving load on a continuous structure for the worst effect and verify it.
Week 10Stability and buckling
Slenderness is a stability problem in a strength problem’s clothes.
Released Mon 5 Aprdue Sun 11 Apr
Euler, effective length, and the fact that the restraint you assumed moves the answer further than the material you specified. Buckling is a failure that gives no warning.
You'll doDrill effective lengths against real restraint conditions, then catch the bracing problem in a scheme while it is still cheap to change.
Week 11Limit states and reliability
The factors are somebody else’s reliability calculation. Yours to understand.
Released Mon 12 Aprdue Sun 18 Apr
A factor is not a safety blanket. It is part of a reliability calculation most engineers never see. Change the factor and you have changed the safety margin — but by how much, and what does that mean for the probability of failure?
You'll doDrill what each factor is buying, then take a real capacity check and work out what moving one factor does to the reliability behind it.
Week 12Checking computer output
Eleven weeks of rules of thumb, against one model you did not build.
Released Mon 19 Aprdue Mon 26 Apr
The output will often be right — that is not the point, because you cannot sign it on the grounds that it looked plausible. This is the week the other eleven were for: a complex structure, its analysis, and no tool but the approximations you have been drilling for three months.
You'll doTake an unfamiliar structural analysis apart using hand calculations. Work out what you believe, what you do not, and whether the results deserve your confidence.
Part 04 · The cohort
Your own work. Not on your own.
Up to 20 engineers. The same problems. The same deadlines.
You’ll do your drills, submit your challenges and receive feedback privately.
Up to 20 engineers will tackle the same problems, often with different assumptions and different answers. A dedicated Discord space gives you somewhere to compare approaches, challenge each other’s reasoning and ask questions.
Nobody is required to participate in the chat. The work happens in falsewrk. The conversation happens with the cohort.
Part 05 · Pricing
What it costs
One payment, and not until the cohort is confirmed. No subscription, no membership, nothing to cancel. The twelve weeks and your first annual renewal are included. The renewal is the twelve weeks again, with fresh challenges. Renewals after that are optional.
Most firms have a training budget and an approver who needs a single page. Here is a letter you can fill in and send — it states the cost, the time commitment, and the fact that nothing clashes with project work.
The App is where you practise. This is where you train. Twelve weeks of purpose-built drills and challenges, with marking, feedback against named criteria, and a deadline. None of this course’s problem bank is in falsewrkPractise today.
Nothing is charged today, and nothing is charged if it doesn't run
- When you take a seat
- You give card details and the card is not charged. That holds your seat — one of 20 — and that is the whole of what happens on the day.
- When the card is charged
- The cohort goes ahead once 10 seats are taken. That is the moment the saved card is charged $450, once, with a receipt and an invoice you can expense. If 10 are already taken when you reserve, the charge follows within a day or two rather than at the checkout page — same price, same terms, and I will have emailed you before it lands.
- If it doesn't fill
- Enrolment closes 18 December 2026. If 10 seats are not taken by then the cohort does not run, your card is never charged and your details are deleted. There is nothing to claim and nothing to watch for — the money never left.
- If you do the work and it wasn't worth it
- Once you have paid: do the drills and submit all twelve challenges, and if you still do not think it was worth $450, email me within a month of the cohort ending and I will refund you in full. No argument. The condition is the work, not passing every challenge.
Change your mind before the charge? Email me and the card is deleted — there is nothing to cancel. The full terms are on the terms page.
Still weighing it up? Try one exercise — same marking, smaller problem, and it will settle the question faster than another paragraph from me.
A certificate that expires
Twelve months, then it expires. To keep it current you earn it again — twelve weeks, twelve fresh challenges, the same criteria — and your first renewal is included in what you pay now. The renewal opens twelve weeks before expiry, so you have the chance to keep it current without a gap.
Passing a challenge once proves you could do it then. It doesn’t prove you can do it today. You cannot get a feel for how a continuous frame redistributes in March and still expect to have it in your hands two years later without touching it again. Judgement is something you maintain, not something you collect.
Skills you don't use are skills you used to have.
- What it says
- That you met the criteria on all twelve challenges. Not that you submitted twelve — that you passed twelve. It is performance against twelve defined sets of criteria, which is a narrower and more useful claim than “competent”.
- What it does not say
- That you would pass the same problems cold, with no warm-up and no week to think. That is a different claim, and this does not make it.
- How you keep it
- Twelve weeks, twelve fresh challenges, the same criteria — renewal opens twelve weeks before your certificate expires. The record that you passed stays yours permanently either way: it is the currency that lapses, not the achievement.
Structural Fundamentals
Specimen · Cohort 01 · 2027
12 / 12 challenges passedValid 12 months
I got the reps. That shouldn't be luck
Early in my career I left engineering to run a martial arts gym full time. I watched students nod along and ask intelligent questions while I demonstrated a technique — and then fumble that same technique the first fifty times they tried it. You can follow a thing completely and still not be able to do it. You have to feel your shoulder turn over and protect your chin as you punch.
Capability is not explained into people.
It is drilled in.
Nobody in a gym believes otherwise. Yet in engineering we spend an enormous amount of time explaining skills and remarkably little deliberately practising them. A slide deck and a worked example will not do what fifty repetitions do, and I think the profession is quietly paying for the belief.Why there are no lectures in this course.
The rest of it: almost fifteen years on infrastructure, most of that designing bridges — from mega projects down to small condition assessments and design reviews. I got lucky. The right projects, the right people, complex structures early enough to learn on. That luck is the only reason I can write this course, which is precisely what bothers me about it.Fifteen years. Bridges, mostly.
I’m building it for the engineer I nearly was — the one who spent two years on condition assessments in a flat market and came out the other side wondering whether they had fallen behind. You hadn't. You just weren't given the problems. This is twelve weeks of being given them.
I write this cohort myself — every challenge, every rubric, every criterion you will be held to — and I am in the discussion every week while you work it. Ahead of February the material goes to experienced practitioners across specialisations to be reviewed for technical accuracy, for the assumptions it rests on, and for whether the assessment criteria are the right ones — because this subject is too important to be only my view of it.Not only my view of it.
Analysis gets cheaper every year. Judgement still takes years to develop, and nobody is being given the reps. I'm building falsewrk so that in ten years' time there are still engineers who can tell when the answer is wrong.
Questions
The ones that actually get asked. If yours isn’t here, email me — I answer these myself and I would rather talk you out of it now than refund you in March.A refund in March costs us both more than an honest no in January.
What happens after I enrol?
Taking a seat means giving card details, and the card is not charged then — it is charged once 10 seats are taken and the cohort is confirmed, as one payment of $450 with a receipt and an invoice. If the cohort is already confirmed when you reserve, that charge follows within a day or two instead, and I email you first either way. After that nothing is asked of you until the cohort opens: enrolment closes 18 December 2026; a week before the start, on 25 January 2027, the app, the discussion and the material open with nothing due and you get a short onboarding quiz — a few questions about what you want out of the twelve weeks; and week 1 lands 1 February 2027. If 10 seats are not taken by 18 December 2026 the cohort does not run, your card is never charged and your details are deleted — there is nothing to claim and nothing to watch for. This is the first cohort and I am still writing the later weeks as the earlier ones run, which is the honest position; what is fixed is the dates, the twelve topics, and the way you are assessed — three criteria per challenge, met or not met, no score. The criteria themselves get sharper as practitioners review them before release; the framework they sit in does not move.
What if I miss a week?
Missing the Sunday deadline does not lock you out. You can submit late, and it is still marked and you still get the feedback — the challenges stay open for one month after the cohort ends, then close. What you lose is the rest of the cohort working the same problem in the same week, which is most of the value. Miss two in a row and I will email you, because that is usually the point where people quietly stop.
What if I do not pass a challenge?
You sit it again — but first, more drills on the criteria you missed. The point is to close the gap before you attempt it again, rather than to have another go at it the same afternoon. The resit counts exactly the same as a first pass, and your certificate does not record how many attempts it took. All twelve have to be met for the certificate, and the challenges stay open for one month after the cohort ends.
How much time does it actually take?
One to one and a half hours in a normal week: the drills in whatever gaps you have, ten minutes at a time, and the challenge in one sitting. The challenge is short on purpose — it is harder than anything you drilled, and answering it in one sitting is the whole test. Week 12 is the exception: it pulls the eleven weeks together against one real analysis and takes four to five hours, so give it a weekend morning. There is nothing scheduled, so you choose when.
What do I need to know already?
You should have studied structural mechanics and be comfortable with basic beam analysis — equilibrium, bending, deflection. You do not need to remember every equation; that is rather the point. If you can work through a simple beam design with a spreadsheet open and an hour to spend, you have enough to start. It is not a first introduction to structural mechanics, and it is not advanced — it is the middle, done properly.
Who is it not for?
Anyone after a shortcut, anyone who wants the certificate without passing all twelve, and anyone who needs code compliance or software training. It is also not an introduction to structural mechanics, and it is not an advanced analysis course — it is about getting fluent in the fundamentals you have already met.
How much access do I get to you?
I write every challenge and every rubric, and I am in the cohort discussion every week — arguing about the problems alongside everybody else. What I do not do is mark every submission by hand, one-to-one mentoring, or scheduled calls. I do read the marking you dispute, and I am the one who settles it. What you get from me is the problems, the criteria, the weekly discussion, and a human decision when the marking is disputed.
What does renewing actually involve?
The twelve weeks again, with twelve fresh challenges and the same criteria. Not a formality and not a refresher video — the whole thing, because a certificate that renews on one easy problem is not evidence you can still do the other eleven. Renewal opens twelve weeks before your certificate expires, so you have the chance to keep it current without a gap. It is effectively the course again rather than a refresher, and your first one is included in what you pay now.
Does it count toward CPD?
It is structured learning with assessed output, which is the part that usually matters — but whether it counts, and under which category, is your registration board's call rather than mine, so I am not going to claim it on their behalf. What I provide is the record: your participation and the hours you actually logged, which is around 15 to 20 for somebody who does the drilling. Keep your marked submissions too; they are the evidence a board asks for.
How do I get my employer to pay for it?
Most firms have a training budget and an approver who needs one page. There is a letter below you can fill in and send — it states the cost, the time commitment, and that there are no scheduled sessions or attendance requirements, and leaves you a line to say what the firm gets back.
Could I not just do this myself from textbooks?
Yes, in principle, and some people do. The textbook on your shelf with a bookmark a third of the way in is the usual result — not for lack of ability, but because nothing was asking for the work. The information is already out there. What you are buying is a structured progression of problems, feedback on your reasoning against named criteria, deadlines that make you do the work, and other engineers doing it alongside you.
Is this an AI course?
No, and I am not going to be coy about where a model does appear. The engineering is yours: you draw the shape, commit to the number, make the assumptions and write the reasoning. A model then marks that submission against the written rubric I set for that week, which is what lets criterion-by-criterion feedback come back in a minute rather than a day. If you think it applied a criterion wrongly, challenge it and I read that one myself. Your submissions are never published and are never visible to the rest of the cohort. The challenges are mine, the criteria are mine, and the engineering is entirely yours — a model assesses your work, it does not do it.
Be wrong here.
Not in the meeting.
The engineers everyone brings the hard problem to were not born knowing the answer. They were wrong enough times, early enough, to know when they might be wrong again.
- Starts
- 1 February 2027
- Price
- $450, once
- Enrolment closes
- 18 December 2026
20 places, and the cohort runs with at least 10. Your card is saved when you take a seat and charged only once it is confirmed — if it never fills, it is never charged. And if you do all the work and decide it wasn’t worth $450, I refund you.