Testing plan: what to buy answers to first
The uncertainty is concentrated in a few estimates, and the question for each is whether buying certainty changes the decision.
- If wrong, the return moves
- 2.8% to 16.3%
- Band today
- 60% obliged, 100% assumed
- What would narrow it, and what that proves
- An indicative term sheet from a sandbox-allocated generator. Strong: it is the term itself, not a proxy
- If wrong, the return moves
- 16.3% to 25.0%
- Band today
- $20.41M benchmark against $2.95M implied by the client
- What would narrow it, and what that proves
- Three vendor quotations at the cluster duty of 1,876 Nm3/h. Strong, and cheap
- If wrong, the return moves
- 16.3% to minus 5.2%
- Band today
- binary, plus $0.50 estimated against a $1.50 gate and a $2.87 breaker
- What would narrow it, and what that proves
- Written positions from PTT and PETRONAS Gas. Strong on the tariff, weaker on acceptance, since a first answer is rarely a final one
- If wrong, the return moves
- 11.1% to 16.3%
- Band today
- four mills within 15 km, assumed
- What would narrow it, and what that proves
- The mill census, plotted. Strong: it is a map question
- If wrong, the return moves
- 12.6% to 19.3%
- Band today
- 280 effective days, inferred
- What would narrow it, and what that proves
- Metering at two mills across a full season. Strong but slow, and the only test here that takes a year
- If wrong, the return moves
- 31.5 to 42.4 mills for the 200 t/day target
- Band today
- 25.7 plus or minus 6.6 Nm3 per m3
- What would narrow it, and what that proves
- The same census, with measured output rather than nameplate. Strong
- If wrong, the return moves
- binary
- Band today
- unknown; not published
- What would narrow it, and what that proves
- A pre-check with the scheme Gas Malaysia uses at Kluang. Moderate: a scheme's view is not a regulator's
- If wrong, the return moves
- 16.3% to above 13% at $32/t
- Band today
- $50 assumed against $238 to $335 merchant
- What would narrow it, and what that proves
- An offer from the industrial gas partner. Moderate: an indication is not a contract
- If wrong, the return moves
- ceiling is 4.83 against 4.17 assumed
- Band today
- mills net 2.18 today
- What would narrow it, and what that proves
- Term sheets with three mills. Strong
The case rests on estimates, and the question for each is not whether it is uncertain but whether buying certainty changes a decision. Three scores, kept separate because they are different questions: how far the answer moves if the estimate is wrong, how wide the honest band already is, and whether a specific piece of work would narrow it.
Sequenced by what each buys rather than by what depends on what.
The first stage buys the four answers that can end the question. The offtake term sheet, because take-or-pay is the largest mover and the most likely to fail. The three upgrading quotations, because they are the cheapest large mover in the plan. The written positions from the two pipeline operators, because a refusal ends the case whatever else is true. And the mill census, which answers cluster geometry, biogas yield and digester status at once, and which starts from the 50-plus mills already under non-disclosure agreement rather than from nothing.
The second stage buys what sizes the build: feedstock term sheets with enough mills to fill 100 tonnes a day, a certification pre-check, and an indication from the industrial gas partner on what the carbon dioxide stream is worth at the fence.
The third buys precision rather than direction: metered seasonal profiles at two mills, and a quotation for the collection fleet against a real geography rather than an assumed 150 km radius.
One of those collides with the brief's only immutable constraint, and it should be said plainly. The brief requires the Go or No-Go by the end of 2026. A metering program that spans a full season cannot report before then, so the seasonal profile behind the 12.6% to 19.3% band will not be measured in time to inform the decision the brief asks for. Two things follow. The year-end decision has to be taken with that band open, which is an argument for deciding to fund stage one rather than to fund the build. And the metering should start now regardless, because it reports in time for the stage-two gate even though it cannot reach the stage-one one.
The same constraint bears on the offtake term sheet. The sandbox award of 28 February 2026 has not been published, so which generators hold quota is unknown, and a term sheet from a counterparty nobody can yet name is not a thing that can be scheduled. If the award remains unpublished through the autumn, the year-end decision will be taken without the single most decision-relevant number in this memo, and the honest response to that is to decide on stage one alone.