The course, 2025 to 2045
Each lane’s band is that rail’s own declared evidence corners, re-evaluated. Where a band sits is where that rail binds.Drawn apart because it is not measured like the others: the queue is a Germany-wide transmission-level reading standing in for a state one, its clearance and inflow are assumptions rather than measurements, and it races in years of wait rather than installations. A lane in line with the other three would claim a standing the evidence does not have.
A band reaching 2046 means the rail does not bind within the course. Nothing is computed past 2045, the corpus’s own horizon.
The rates are yours — nine hands, grouped by rail
Copper is the published evidence range. The dark tick is the study’s own setting.Fills three and a half years late, leaks at the dissolution rate, and throttles its own intake through the wage the trade must pay.
A single scenario estimate of deployable flows, converted at the installed cost.
Growth relative to the course, with the 2024 crash and the 2025 recovery as its ends.
It sets what the flows can finance and what households will pay, so it moves capital and demand at once.
The weakest of the four rails. No state queue is published and no clearance rate exists anywhere in the corpus.
The pipe fills three and a half years late
Entry cohorts of the electrical and SHK trades enter a regulated pipe, survive the dissolution leak and the in-trade exit, and accumulate into a stock the retirement wave drains. A change made this autumn is invisible in the stock until the cohort clears.
The completion meter reads 72.5% against a dissolution complement of 61.5% — a gap of 11.0 points. Two readings of one transition, so the meter corroborates the leak and cannot filter the pipe a second time.
The gap ladder is a finding, not a range
11.8-fold spreadFour published readings of one gap. Three of the four come from the same chamber organisation using different instruments, so this is largely one house measuring four ways. Pick a rung; the spread is the readout. Never averaged, never an error bar.
The counting rule moves the displayed gap and nothing else in the world. That separation is the finding: no choice among the four changes which rail binds first.
What this world is weak at
A race whose runners are not equally well measured has to say so, or the picture claims a precision the evidence does not have. These are on the page because they are load-bearing, not in a footnote.
The grid rail's central value is a national figure standing in for a state figure, and that makes it the weakest of the four rails. The official 226 GW queue reading is Germany-wide and transmission-level (C-53); no published state queue exists. The clearance and inflow rates are declared assumptions, not measurements, and the default clearance is set so this rail does not bind at the defaults, as the report's own ruling requires. Treat every grid result as the band it is, not a point.
Course output per installer is an assumption, not a measurement. The corpus holds no published figure for installations per installer per year; the default is the report's central judgement, the position bands span the wider plausible band, and the report text says the rate has no evidenced default. The workforce rail's binding year moves with it more than with any other lever, and that is the honest state of the evidence.
The 2024 installer stock is a stock-flow identity, not a count: the level the current completion flow would sustain, times the evidenced 2015-to-2024 intake ratio. The corpus holds no installer stock count. The pre-2024 intake path is a declared reconstruction at the 2024 growth readings. Both constructions are stated where they are used.
Only two trades are modelled as pipes, the electrical and SHK trades, because those are the two the report names and the two with occupation-level intake evidence. The insulation trades' three-year duration is evidenced (C-81) but the corpus holds no insulation intake series, so there is no third pipe; the structure ends where the measurement ends.
Retirement is modelled as one rate draining both trades evenly, derived from the state-wide retirement wave (C-87 over C-74). The corpus holds no occupation-specific retirement profiles, so cohorts age into the stock through the pipe but leave it by the rate, not by a profile.
The pipe carries two leaks, not three. The corpus pairs 'about 70-75 per cent of apprenticeships completed' with 'about 25-30 per cent dropout/dissolution' in one line, so completion is the complement of the dissolution reading, not a further filter; the meter gap is carried as completion_meter_gap. Two meters are in play and both are evidenced: the prose reasons on the overall NRW dissolution rate of 29.7 per cent (C-33), while the model defaults to the Handwerk rate of 38.5 per cent (C-34), which is why the prose's claim-faithful in-trade band (about 49 to 56 of 100) and the model's default yield (about 45 of 100) differ. Neither reading is hidden.
The state heat pump series does not exist. The course anchor is an estimate scaled from the national series (C-56), and the rival estimate basis sits at the low end of its range; both are labelled estimates. The backcast tests the national evidence base, which is measured, not the state scaling, which is not.
The capital rail is scoped to the heating-conversion course. The corpus's wider financing gap of 8 to 18 billion euros a year (C-64) covers full building decarbonisation and ships as scenario arithmetic; inside the model the rail's deliverable is what the flows finance at the published installed cost. The flow level itself is a single scenario estimate, carried at a zero-width evidence range.