Oil processing ratios

Why "crude per petroleum gas" has no answer

Ask a calculator how much iron ore one gear costs and it walks the recipe backwards. Ask how much crude one unit of petroleum gas costs and the question is malformed: the same refinery craft that produced your gas also produced heavy and light oil. Assigning a share of the crude to the gas alone requires deciding what the other two outputs are worth, and that is a judgement, not a calculation.

This is why our calculator stops at fluids and reports them as inputs rather than expanding further. A tool that quietly picked an allocation would be inventing a number and presenting it as derived.

The forward question does have a clean answer, and it is the one worth asking: given a refinery, how many cracking plants stop the outputs backing up?

What one refinery produces

Advanced oil processing, running in an oil refinery at speed 1:

Output	Per second
heavy oil	5.00
light oil	9.00
petroleum gas	11.0

Those proportions are fixed. If your factory needs gas and not heavy oil, the heavy oil still arrives, and a full heavy oil pipe stops the refinery dead. Cracking exists to solve that.

The cracking balance

Solve it by conservation: every unit of heavy oil goes to a heavy cracker, and every unit of light oil — both the refinery's own output and the light oil the heavy crackers produce — goes to a light cracker. Per refinery that gives:

Machine	Per refinery	Whole numbers
Oil refinery	1	20
Heavy oil cracking plant	0.25	5.00
Light oil cracking plant	0.85	17.0

In whole numbers that is 20 refineries to 5.00 heavy crackers to 17.0 light crackers. Unlike most ratios in the game this one does land on integers, which is why it circulates as a fixed recipe rather than as a formula.

What cracking actually buys

The usual framing is that cracking prevents backups. It does, but the yield figures show it is also the single largest efficiency decision in the oil chain:

Setup	Petroleum gas per crude	Relative
Basic oil processing	0.450	1.00x
Advanced, no cracking	0.550	1.22x
Advanced, fully cracked to gas	0.975	2.17x

Advanced processing on its own is only a 22% improvement on basic if you are only after gas. Add the cracking chain and the same crude yields 2.17 times as much — nearly 98 units of gas per hundred crude, against 45 on basic. Researching advanced processing and not building the crackers captures a small fraction of what the upgrade is worth.

Water is the input people forget

A fully cracked refinery line consumes 26.5 water per second against 20.0 crude — more water than oil. The water goes into advanced processing itself and into both cracking recipes. Oil setups that stall for no visible reason are often short of water rather than short of crude.

Solid fuel: the one place the choice is obvious

Solid fuel can be made from any of the three fluids, and the recipes are not equivalent:

Recipe	From	Fluid per solid fuel
solid fuel from light oil	light oil	10
solid fuel from heavy oil	heavy oil	20
solid fuel from petroleum gas	petroleum gas	20

Light oil is the efficient route at 10 units per solid fuel, half the cost of the petroleum gas recipe. This is also why solid fuel and full cracking compete: light oil sent to solid fuel is light oil not cracked into gas, so the two decisions have to be made together rather than separately.

Solid fuel versus cracking: the decision criterion

Light oil makes solid fuel at ten units per fuel, half the cost of making it from petroleum gas at twenty. But light oil cracked down to gas produces petroleum gas, and gas is what most of the factory actually consumes — plastics, sulphur, rocket fuel components. The two uses compete directly for the same light oil, and the right allocation depends on what the factory is short of.

The decision criterion is straightforward. If the factory has more petroleum gas than it consumes and gas tanks sit full, diverting light oil to solid fuel is free: the light oil would otherwise back up and stall the refineries anyway. If gas tanks run low, every unit of light oil sent to solid fuel is gas the factory does not get, and the cracking balance shifts — fewer light crackers are needed because less light oil arrives at them.

Heavy oil should never be used for solid fuel directly if a heavy cracker is available. Cracking heavy to light and then making solid fuel from light uses twenty units of heavy to produce fifteen units of light (via the heavy cracker), enough for one and a half solid fuel. Making solid fuel directly from heavy costs twenty heavy for one fuel. The cracked route is fifty percent more efficient, and the only reason to make fuel from heavy is if the heavy cracker itself is unavailable or unbuilt.

Route	Fluid per solid fuel	When to use
Light oil → solid fuel	10 light oil	Default; light oil is the efficient feed
Heavy → crack to light → fuel	20 heavy → 15 light → 1.5 fuel	When heavy crackers are running
Heavy oil → solid fuel directly	20 heavy oil	Only if heavy cracking is unavailable
Petroleum gas → solid fuel	20 gas	Worst ratio; only when gas is surplus

What allocation does to the cracking balance

The 20-to-5-to-17 ratio assumes every drop of light oil is cracked. Diverting even a small fraction to solid fuel changes the downstream counts because the light crackers receive less input. This is not a subtle effect: if half the light oil goes to fuel, half the light crackers are idle and the petroleum gas output drops accordingly.

Worked through for a single refinery: advanced processing produces nine units of light oil per second. If a chemical plant makes solid fuel from light oil at one craft per second consuming ten units, one solid fuel plant consumes ten units of light oil per second — more than the entire light oil output of one refinery. Scaled to the twenty-refinery block, that is 180 units of light oil per second, of which ten solid fuel plants consume 100. The remaining 80 units feed roughly five and a third light crackers instead of the seventeen the full-cracking ratio calls for. The gas output falls by the amount those twelve idle crackers would have produced.

The practical consequence is that solid fuel production should be treated as a deliberate reduction in gas capacity, not bolted on as an afterthought. Build the fuel plants, count how much light oil they consume, and reduce the light cracker count by that fraction. A factory that builds the full 20-to-5-to-17 block and then adds solid fuel plants without reducing crackers ends up with light crackers that starve intermittently — not broken, but never running at the rate the ratio assumes.

Water: the input with no pump count

A fully cracked refinery block consumes 26.5 units of water per second per refinery, against 20 units of crude. For the twenty-refinery whole number block that is 530 water per second against 400 crude. Water is the larger fluid input, and it is the one most often undersupplied.

The dataset records the offshore pump's energy draw (60 kilowatts) and its void energy source, but it does not record a pumping rate. The prototype entity in the game data has no fluid output or pumping speed field in the extracted JSON — those values live in a part of the prototype definition this dataset does not capture. This page therefore cannot tell you how many offshore pumps the 530 water-per-second block needs. Giving a number would mean inventing one.

What can be said without that data is that water is a flow constraint, not a capacity constraint. Unlike crude oil, which is extracted from mines at a rate that depends on the patch and can be buffered in tanks, water arrives directly from the pump at whatever rate the pump produces. There is no water patch richness to overbuild. If the combined pump output is less than the refinery block's demand, the refineries run dry and stall, and the stall looks identical to a crude shortage from the outside.

The practical layout rule is to run water on its own pipe rather than teeing into an existing water line. Water pipes that also feed boilers or concrete production compete for the same flow, and a refinery block starting up can drain a shared line faster than the boilers recover. A dedicated water line from shore to refineries removes that coupling even without knowing the exact pump count.

Pipe throughput as the invisible ratio breaker

The 20-to-5-to-17 ratio assumes fluid can actually reach the machines. Pipes in Factorio have a throughput that falls with length: a long straight pipe carries less fluid per second than a short one, because each pipe segment adds resistance. A correctly ratioed oil block can underperform for no reason visible in the machine counts if the crude oil pipe from the pumpjacks is too long, or if the petroleum gas line to the plastics plant snakes across the base.

The dataset does not include pipe resistance or fluidbox connection data, so this page cannot give a maximum pipe length for a given flow. What can be stated is the shape of the problem: fluids are not belts. A belt carries its rated throughput regardless of distance. A pipe does not. The off-gas line from a twenty-refinery block carries 390 units of petroleum gas per second, and that flow will not survive an arbitrarily long pipe.

The standard mitigation is a pump every few pipe segments to reset the flow pressure, and placing chemical plants close to the refineries rather than running gas across the base. Storage tanks buffer volume but do not increase throughput; a tank at the end of a long pipe fills slowly even though it can hold a lot. If an oil block is ratioed correctly but runs below its rated output, pipe length is the first thing to check after water supply.

Why "crude per gas" is genuinely unanswerable: a concrete example

The earlier section explained the allocation problem in the abstract. Here is what it looks like with actual numbers. One refinery running advanced processing consumes 100 crude oil and 50 water over five seconds, producing 25 heavy oil, 45 light oil and 55 petroleum gas. Suppose you want to know how much crude "one petroleum gas cost".

Answering requires deciding what the 25 heavy oil and 45 light oil are worth. If they are waste — cracked to gas with no other use — then all 100 crude are attributable to the 97.5 total gas (55 directly plus what cracking yields), giving roughly 1.03 crude per gas. If heavy oil is diverted to lubricant for express belts, and light oil to rocket fuel, then the 55 gas is a byproduct and its "share" of the crude depends on how you value lubricant and rocket fuel. A calculator that picked a single allocation would be burying that judgement inside a number.

This is not a theoretical edge case. A mid-game base typically uses heavy oil for lubricant, light oil for both solid fuel and some cracking, and gas for plastics and sulphur. The proportions shift as the base grows. Any fixed "crude per gas" figure would be wrong as soon as the base changed which fluid it was short of. The honest answer is the forward one: given the refineries, here are the three output rates; decide what to do with each.

Productivity modules change the balance

The cracking ratio assumes no modules. Productivity modules in a refinery increase output without increasing input, which sounds universally good but shifts the downstream counts. A refinery with +30% productivity produces 30% more of all three fluids per unit of crude, but the crackers processing those fluids are a separate set of machines. If the crackers are unmodded, 30% more heavy output needs 30% more heavy crackers. If the crackers also carry productivity, the shift compounds differently because the cracker's own output bonus changes how much light oil arrives at the light crackers.

Speed modules have a simpler effect: they increase both input and output proportionally, so the ratio between refineries and crackers stays the same — everything just runs faster and uses more crude. Productivity is the module that breaks the fixed ratio because it acts on output only. For a beacons-and-modules megabase, the 20-to-5-to-17 figure is a starting point, not the final count; the actual numbers depend on which machines carry which modules.

Applies when…

• No modules or beacons are fitted. Productivity modules change the cracking balance, because they alter output amounts without altering input amounts.
• The cracking ratio assumes every drop is cracked down to petroleum gas. A factory consuming light oil directly — for solid fuel, rocket fuel or flamethrower ammunition — needs fewer light crackers in proportion.
• Machine counts assume pipe throughput can carry the flow. Fluid throughput falls with pipe length and is a common reason a correctly-ratioed oil setup underperforms.
• Coal liquefaction is a separate route with its own balance, including a heavy oil output that feeds back into its own input. It is not covered here.

Related

• Production Ratio Calculator — stops at fluids, for the reason above
• How production ratios actually work
• All ratio pages