Best Practices: Reading the Field Well

This chapter distils the habits of infrastructure literacy: which numbers to trust, which sources to follow, and how to maintain calibration as the field changes.

The Core Discipline

Infrastructure news is full of numbers. Most readers skim past them. A literate reader pauses.

Three reflexive questions:

1. What's the unit? Watts or watt-hours? Tonnes of CO2 or tonnes of carbon? Dollars per MWh or dollars per year?
2. What's the scale? Is this small, medium, or large for this category?
3. What's the context? Is this per household, per year, per capita? Compared to what baseline?

If a piece of writing doesn't answer these, suspect it. Not of malice; probably of laziness. Writers often repeat numbers without understanding them.

The Most Important Checks

Capacity vs energy

Chapter 02 emphasised this. It matters enough to repeat. A 1 GW plant and 1 GWh of battery storage are very different things. Confusing these is the single most common infrastructure mistake in coverage.

Rule: if the unit has "hours" or "years" in it, it's energy. If not, it's power. A "battery of 100 MW and 400 MWh" is a battery that can deliver 100 MW for 4 hours (400 ÷ 100).

Capacity factor

When a press release says "will power 300,000 homes", check the math. Convert the capacity to energy using a plausible capacity factor. Solar at ~20%, wind at ~35%, nuclear at ~90%. Then divide by the energy an average home uses.

Often the PR math assumes 100% capacity factor and dramatically overstates output.

Total system cost vs LCOE

LCOE is useful for comparing sources, but it's not the total cost of reliability. A grid that's 80% cheap solar plus 20% expensive firming may have lower total cost than one that's 100% medium-priced natural gas. The headline number isn't enough.

Primary vs final energy

Electric cars are ~4x as efficient as gas cars at using primary energy. Comparisons of "energy used" should specify which. Headlines sometimes don't.

Lifecycle vs operational emissions

A solar panel has some CO2 emissions in manufacturing. Over its life, it produces much less than a gas plant. Fair comparisons use full lifecycle emissions, not just operating emissions. Watch for which a source is using.

Sources Worth Reading

Infrastructure has better writing and data than most people realise. A partial list:

Data sources

• IEA (International Energy Agency): annual World Energy Outlook, plus many shorter reports
• Our World in Data's energy section: excellent clean visualisations
• EIA (US Energy Information Administration): US-focused but much globally applicable
• Ember: clean energy and grid analysis, free data
• NREL (US National Renewable Energy Laboratory): technical reports, cost estimates
• Energy Institute's Statistical Review of World Energy: annual, comprehensive

Commentary and analysis

• David MacKay's Without the Hot Air: free online, still foundational
• Vaclav Smil's books: dense, quantitative, essential for perspective
• The MIT Technology Review's energy coverage
• Saul Griffith's Electrify
• Michael Liebreich (formerly BloombergNEF): incisive commentary
• Utility Dive, Energy Institute newsletters: trade-press style

Specific topics

• Carbon Brief: climate and energy explained; good on UK
• The Pragma: policy-focused
• Heatmap News: climate and energy journalism
• Volts podcast (David Roberts): long-form interviews
• The Energy Gang podcast

Skepticism

Read widely enough to notice when sources are biased. Fossil-funded research, renewable industry PR, and NGO campaigns all have perspectives. The best readers triangulate across sources with different incentives.

Habits That Help

Keep orders of magnitude

You don't need to memorise specific numbers. You do benefit from knowing orders of magnitude for the things you care about. "A large data centre is hundreds of MW" or "a US house uses about 10,000 kWh/year" are useful anchors.

If you see a number that's an order of magnitude off from your expectation, pause and check. Usually one of the two of you is wrong.

Learn one topic deeply

Infrastructure is wide. Expertise in one slice (your local grid, a specific technology, one country's water system) produces transfer: you understand the shape of the problem better, and you can navigate adjacent topics faster.

Notice geographic variation

Energy and infrastructure policies are local. What works in Iceland doesn't transfer to India. Always ask: where? Generic "what country" advice is usually wrong for your specific context.

Follow a few specific projects

Pick one big project (a transmission line, a new nuclear plant, an offshore wind farm) and follow it over years. You'll learn how long these things take, where they get stuck, how costs evolve. Pattern-matches well to other projects.

Do the math occasionally

When a number feels fishy, check it. A 15-minute calculation beats hours of opinion. Infrastructure is a quantitative field; literacy includes the comfort to do basic arithmetic.

Common Failure Modes

The single-silver-bullet

"X will solve climate." Solar, nuclear, hydrogen, efficiency, carbon capture. Each is part of a solution; none is the solution. Watch for sources that present a single answer as obvious.

The "it's impossible"

"Deep decarbonisation is impossible." Sometimes true for specific paths; rarely true globally. Many things declared impossible turned out possible with sustained effort. Watch for sources that rule out progress a priori.

The timescale confusion

"By 2030 we'll be..." Most big infrastructure changes take decades. Headlines often compress timescales to feel urgent. Real timescales are long; meaningful change comes from sustained trajectory, not instant transformation.

The cost comparison without context

"Solar is cheapest." Per MWh on average sunny days, yes. At 3 AM, no. Full system cost matters. Headlines often pull the cherry-picked comparison.

The climate narrowly framed

"This plant is green." Green on operating emissions? Green on lifecycle? Green including the land it displaced? Framing determines answers.

The expert dismissal

"The experts say..." Which experts? With what track record? Experts have been very wrong before (nuclear cost projections, fossil-fuel demand forecasts, solar cost trajectories). Give weight to expertise; don't treat it as oracular.

When Numbers Conflict

You'll frequently encounter conflicting numbers: "X is cheapest" / "Y is cheapest". How to resolve?

• Check the date: costs change fast in this field. A 2015 study is often outdated
• Check the methodology: what's included, what isn't?
• Check the assumptions: discount rate, capacity factor, lifetime, fuel price, carbon price
• Check incentives: who funded the study?
• Look for meta-analyses: multiple studies converging matter more than a single study

When serious disagreement remains after these checks, it's often because the question is genuinely uncertain (long-term costs, geopolitical risks, technology trajectories). Honest uncertainty is an answer, not a failure.

The Long View

Infrastructure thinks in decades. Individual news cycles think in days. These timescales rarely align.

Useful to cultivate:

• Patience with apparent lack of progress: a pipeline approval that takes 7 years is normal, not broken
• Patience with apparent breakthroughs: "a team announces..." rarely means deployment this decade
• Noticing sustained trends: solar's cost decline was visible for a decade before headlines; the rapid EV adoption was obvious by 2020 to those watching

The readers who end up most-informed about infrastructure aren't reading the most; they're reading consistently over years and building mental models that update gradually.

What Literacy Is and Isn't

Infrastructure literacy isn't:

• Professional engineering knowledge
• Policy prescription
• Certainty about future trajectories
• The ability to win every argument

It is:

• A working vocabulary
• Rough orders of magnitude for core quantities
• Basic models of how the big systems work
• The habit of checking claims against numbers
• Enough context to engage informed people usefully

Don't overstate your expertise. Overstating is common; it produces bad arguments and worse decisions. Literacy plus humility beats false confidence.

A Final Note

Infrastructure is the quiet layer under most of what matters. It's easier to ignore than to engage with. Engaging with it makes you a more useful citizen, investor, parent, and neighbour.

There's also something satisfying about understanding how the world works physically. Cold winter evening; light comes on; water comes out of the tap; message reaches a friend across the planet. All of this runs on systems built over generations by people you mostly don't know.

Knowing how any of this works is a form of respect for the work. It's also a form of participation in choices about what gets built next.

Where to Go From Here

• Find out where your electricity comes from this year; check how that compares to ten years ago
• Read one of the books in the README's resource list
• Pick one local infrastructure project (a transmission line, a transit build, a water system upgrade) and follow it
• Do the math on your own household's energy use; see if it surprises you
• Revisit this tutorial in a year; the numbers will have aged; the frameworks less so

Infrastructure moves slowly. That's part of its appeal: what you learn stays useful. Good luck.