Estimate your heating load from fuel you already bought, then check a quote against the published sizing window.
304% of the loadworked example: 100,000 Btu/hr input furnace, 31,500 Btu/hr load720 therms of heating fuel at 80% AFUE over 3,800 HDD65, scaled to a 15°F design temperature, gives 31,500 Btu/hr. The window is 100–140%. Example inputs: use your own.
The short answer
A furnace’s output should be 100% to 140% of your home’s design heating load, per ENERGY STAR and ACCA Manual S. Square footage alone cannot tell you that load. This tool estimates it from a year of fuel bills.
In the worked example, the load is 31,500 Btu/hr, and a 100,000 Btu/hr furnace lands at 304%. That is a reason to ask which load calculation the quote used.
Check a furnace quote against your bills
Nine inputs, most from papers you already have. Nothing is emailed or stored.
—Where the proposal sits
—Design heating load
—100–140% window (ENERGY STAR / Manual S)
—Proposed output capacity
—Implied Btu/hr per sq ft
What this assumed
—
A screening estimate from your fuel bills, not an ACCA Manual J. No room-by-room loads, no duct sizing, and it cannot see changes since the bills. Installing, venting or gas-piping a furnace is licensed work. HyreHVAC does not install, service or sell HVAC equipment.
The formula shown is the formula that runs
How does the furnace sizing calculator work?
It uses your fuel bills, not floor area. A square-foot calculator has one input and prints a number that looks like a specification. This one has six inputs, none of them floor area, and calls its result an estimate.
How far each input moves the estimate, one at a time, against the worked example of 31,500 Btu/hr. Every figure comes from this page’s engine.HyreHVAC calculation using the furnsize engine on this page.
The problem with square-foot rules: Manual J: Residential Load Calculation, 8th edition establishes the heating and cooling load of the dwelling.
ACCA states that "A proper load calculation, performed in accordance with the Manual J 8th Edition procedure, is required by national building codes and most state and local jurisdictions." It has dozens of inputs.
A one-input rule borrows its authority, and it errs toward oversizing, the trade’s usual mistake.
What this does instead: the fuel-bill (degree-day) method treats your house as an instrument that measured itself all winter. Take the heating fuel you bought, subtract what you burn with the furnace off, and convert to delivered heat at the old furnace’s AFUE.
Divide by the degree days for the same period. That gives a heat-loss rate in Btu per hour per degree, a real property of your house. Multiply by the design-day temperature difference to get the load.
The load formula: delivered Btu = (fuel used − non-heating baseload) × Btu per unit × AFUE ÷ 100 Btu/hr per °F = delivered Btu ÷ (HDD65 × 24) design load = Btu/hr per °F × (65 − outdoor design temperature °F)
Why 65 appears twice: degree days are counted from 65°F, so the temperature difference must use the same base. Do not mix in degree days from a source with a different base. Our county degree-day lookup uses base 65.
The worked example: 900 therms of gas over twelve months, minus 180 for water heating, cooking and drying, leaves 720 therms for heat. At 100,000 Btu per therm and 80% AFUE, that is 57.6 million Btu delivered.
Over 3,800 HDD65 (91,200 degree-hours), the house lost 631 Btu per hour per degree. At a 15°F design temperature the gap is 50°F, so the load is 31,570 Btu/hr, reported as 31,500.
The 100–140% window is 31,500 to 44,000 Btu/hr of output. A 100,000 Btu/hr input furnace at 96% AFUE delivers 96,000: 304% of the estimate.
HyreHVAC analysis: floor area appears nowhere in that chain, so the chart has no bar for it. The floor-area box only shows your home’s implied Btu per square foot, so you can compare it with the round numbers you have read elsewhere.
Degree days and design temperature move the answer most, and most calculators never ask for them. A wrong baseload moves it by a quarter, in the direction that makes an oversized proposal look fine.
Where do you find each input?
No technician needed. Six come from papers you have, one from our degree-day lookup, and one you ask for. Asking for it is half the value.
1
Fuel type and quantity
Twelve months of gas, propane or oil bills. Gas bills show therms or ccf; propane and oil show gallons on delivery tickets. Add all twelve. Do not scale up one winter bill.
2
Non-heating baseload
Fuel burned with the furnace off: water heating, cooking, a gas dryer. Take a July or August bill and multiply by twelve. With electric water heating and no gas dryer, it may be near zero.
3
AFUE of the old furnace
On the nameplate inside the furnace door, or from the model number. This is the OLD furnace. If unreadable, 80% is typical for a metal flue and 90%+ for white plastic venting, but that is a guess worth 21%.
4
Heating degree days, base 65°F
From our county degree-day lookup (NOAA 1991–2020 normals). Use the annual figure with twelve months of fuel. NOAA also publishes actual monthly figures for a specific year.
5
Outdoor design temperature
Ask your contractor which one they sized to. It is one number on page one of a Manual J report. If nobody can produce it, you have learned something important.
6
Heated floor area
Excluding an unheated garage or unfinished basement. Not used in the load; it only shows your implied Btu per square foot.
7
Proposed furnace input rating
From the quote, in Btu per hour. It is usually the biggest number on the line, and usually input, not output. The maker’s spec sheet gives both.
8
Proposed furnace AFUE
From the quote or spec sheet. With the input rating it gives the output, which the sizing window applies to.
How does each input change the furnace size?
Each row is the engine run on the worked example with one field changed. Baseline: 900 therms less 180 baseload, 80% AFUE, 3,800 HDD65, 15°F design, 2,200 sq ft, against a 100,000 Btu/hr input furnace at 96% AFUE.
What changed
Load estimate
100–140% window (output)
Where the proposal lands
Nothing: the worked example
31,500 Btu/hr
31,500 – 44,000
304%: more than double the load
Old furnace was 96% AFUE, not 80%
38,000 Btu/hr
38,000 – 53,000
253%: more than double the load
County has 6,000 HDD65, not 3,800
20,000 Btu/hr
20,000 – 28,000
480%: more than double the load
County has 2,200 HDD65, not 3,800
54,500 Btu/hr
54,500 – 76,500
176%: above the 140% ceiling
Design temperature −5°F, not 15°F
44,000 Btu/hr
44,000 – 62,000
217%: more than double the load
Design temperature 30°F, not 15°F
22,000 Btu/hr
22,000 – 31,000
434%: more than double the load
Baseload left at zero
39,500 Btu/hr
39,500 – 55,500
243%: more than double the load
Baseload 300 therms, not 180
26,500 Btu/hr
26,500 – 37,000
365%: more than double the load
Proposal is 80,000 Btu/hr input
31,500 Btu/hr
31,500 – 44,000
243%: more than double the load
Proposal is 60,000 Btu/hr input
31,500 Btu/hr
31,500 – 44,000
182%: above the 140% ceiling
Proposal is 100,000 input at 80% AFUE
31,500 Btu/hr
31,500 – 44,000
253%: more than double the load
950 gallons of propane, 120 baseload
33,500 Btu/hr
33,500 – 46,500
288%: more than double the load
700 gallons of heating oil, 90 baseload
37,000 Btu/hr
37,000 – 51,500
261%: more than double the load
HyreHVAC calculation. Each row is one run of the furnsize engine with one field changed; the figures are the engine’s own rounded output.
Row three surprises people. More degree days on the same fuel gives a smaller load, because degree days divide the fuel: the same fuel over a colder winter means slower heat loss per degree.
What changes the furnace size estimate most?
Four inputs matter, and not in the order most people guess. The input everyone expects first is not in the calculation at all.
1. Degree days (−37% to +73% across the range tested). The biggest lever, and it works backward at first glance because it divides the fuel. A mismatch between the fuel period and the degree-day period is the fastest way to a wrong answer.
2. Design temperature (−30% to +40%). It turns a seasonal average into a design-day figure. It is also the input you are least likely to have, so the tool asks rather than assumes.
3. The baseload (−16% to +25%). Leaving it at zero inflates the estimate by a quarter, in the direction that makes an oversized quote look fine.
4. The old furnace’s AFUE (+21% across the range tested). It converts fuel bought into heat delivered. It is often guessed, because the nameplate is inside the cabinet.
Left out on purpose: floor area. It only drives one context figure: your implied Btu per square foot, 14.4 in the worked example. Compare it with the round numbers online.
Also left out: rounding to a size you can buy. That would turn an estimate into a selection. Where the load falls between products is a Manual S call. Manual S: Residential Equipment Selection selects equipment against the Manual J load at the design conditions.
And it cannot see inside the house. It captures the whole envelope, not its parts, so it cannot say which room loses heat. DOE’s review of 44 field studies found duct leakage in 90-100% of systems, a problem no whole-house figure can see.
The number the whole screen turns on
Where does the 100–140% furnace sizing window come from?
From ENERGY STAR’s HVAC Design Report, which restates ACCA Manual S. The percentage is our math; the window is published and checkable. It is the most useful sentence to take into a sales visit.
ENERGY STAR’s Single-Family New Homes National HVAC Design Report, Version 3.1 / 3.2 / 3.3 (Rev. 14) defines the heating sizing percentage at Item 4.20 as "Total capacity (Item 4.19) divided by total heat loss (Item 3.14)".
Item 4.21 sets the limit. For a heating-only furnace it is 100 – 140%. With cooling, 100 – 140% is recommended and 100 – 400% allowed, because the furnace and AC share one cabinet and blower.
Footnote 26 says: "Equipment shall be selected using the maximum total heat gain in Item 3.12 and the total heat loss in Item 3.14 per ACCA Manual S, Second Edition".
So the window is the trade standard, in a form anyone can check. Manual S: Residential Equipment Selection selects equipment against the Manual J load at the design conditions.
The load in the denominator is a Manual J load: Manual J: Residential Load Calculation, 8th edition (ANSI/ACCA 2 Manual J - 2016).
ACCA states: "A proper load calculation, performed in accordance with the Manual J 8th Edition procedure, is required by national building codes and most state and local jurisdictions." Nothing on this page is a Manual J.
HyreHVAC analysis: the two ends fail differently. Below 100%, the house is cold on the coldest morning. Above 140%, extra capacity buys nothing and adds cycling. On a mid-size house, 40% of the load is about one step on a furnace price list, so which way a selection rounds really matters.
What this window is not: a building code, a warranty condition, or a legal limit. It is a published benchmark for new homes in a certification program. A proposal far outside it gives you a question to ask, not a verdict.
What should you ask to tell a real load calculation from a guess?
You need no heating knowledge for these. Each asks for a document or number that exists or does not, and a deflection is easy to hear.
Ask this
A real answer
A rule of thumb
Was a Manual J done for this house, and may I see it?
Yes, here is the report: several pages, your address, room-by-room assumptions. It can be emailed before you sign
“We do them for new builds.” “It is in our software.” “You do not need one for a swap.” A report that cannot be produced was not done
What outdoor design temperature did you use?
One heating number for your location, the winter design condition, from page one of the report
A pause. Or a number several degrees colder than your climate, a common route to an oversized furnace
What air leakage rate did you assume, and was it measured?
A figure and its source: a blower-door test, or a tightness class based on the house’s age, stated as an assumption
“It is a typical house.” Air leakage is one of the biggest parts of a heating load, worth a whole furnace size
What does Manual S say about the unit you chose?
Output as a percentage of the heat loss, a reason for where it landed, and the exact model number
“It is what we always put in a house this size.” That refers to the street, not your house
Is that rating the input or the output?
Both: the input from the nameplate and the output at the rated AFUE
Used interchangeably. On an 80% AFUE furnace they are 20% apart, half the sizing window
Anyone who did the work can answer each in a sentence. The right-hand column is usually habit in a busy trade, not dishonesty.
HyreHVAC analysis: ask all five before you sign. Afterward, nobody has a reason to produce a document, and the size is locked in for 15 to 20 years. None of these questions is about price, and a good contractor usually enjoys them.
The honest answer is not mainly fuel
What does an oversized furnace actually cost you?
Mostly comfort, cycling and airflow, not gas. Saying it wastes fuel would suit this page, but the best evidence we found does not support that.
The worked example only: four proposed furnaces against one bill-derived load of 31,500 Btu/hr. Each figure is this page’s engine with only the proposed input rating changed. Not real houses.HyreHVAC calculation, using the furnsize engine on this page. Illustrates the math, not the market.
DOE’s Residential HVAC Installation Practices: A Review of Research Findings (June 2018) reviewed 44 field-study reports. It gives a recommended range of 33–48% oversize compared to Manual J.
Across studies from Arizona, southern California, Florida, New Jersey, Nevada, Texas and Wisconsin, 31% to 93% of systems tested were above that range, and 0–9% below. In some samples, oversizing is nearly universal.
This part cuts against us: the same review summarizes an NREL guideline on high-efficiency gas furnaces (Brand and Rose, October 2012) as finding that furnace AFUE "is relatively insensitive to oversizing based on lab testing in the 70% to 120% oversizing range".
So the AC idea that oversizing wastes energy does not carry over cleanly to furnaces.
HyreHVAC analysis: the real costs are four. Temperature swings: the house overshoots, shuts off, and cools again. More cycles: ignition and shutdown are the hardest parts of a furnace’s work. Cold far rooms: short runs never reach the far bedroom.
Duct strain: a bigger furnace moves more air. DOE’s review found low airflow in more than 50% of systems and duct leakage in 90-100%, so a bigger furnace on marginal ducts makes the duct problem the thing you notice.
Undersizing is not free either. Below the design load, the house gets cold on the worst morning of the year. Unlike cooling, there is no upside to being a little small. That is why the trade rounds up, and why “your contractor is lazy” is usually the wrong story.
What to do: if a proposal lands far above the window, do not start with the number. Ask which load calculation it was sized against. That tells you more than the size does.
Even the smallest of the four lands outside the window. The next section explains why.
What if no furnace is small enough for your house?
It happens, more and more, in tight modern houses. Sometimes the honest result is that every furnace on offer is oversized.
HyreHVAC calculation: in the worked example the load is 31,500 Btu/hr, so the window runs from 31,500 to 44,000 Btu/hr of output. A 60,000 Btu/hr input furnace at 96% AFUE delivers 57,500: 182% of the load, outside the window before anyone made a mistake.
We do not publish a figure for the smallest furnace on the market. Ranges vary by maker, fuel, venting and year, and we hold no catalog. The math shows the shape of the problem: a well-sealed house in a moderate climate can need less heat than any single-stage furnace delivers.
HyreHVAC analysis: that changes the question from “is this contractor sizing properly?” to “is a single-stage furnace the right machine?” The usual answers are a two-stage or modulating furnace, which runs well below its nameplate most of the winter, or a heat pump. Both are Manual S decisions for a contractor with a load calculation.
A caveat: a bill-based load can also come out low after a mild winter, a deep thermostat setback, closed-off rooms, or an old furnace that never kept up. Each understates the load. That is why this is a screening figure, and a Manual J tells you the rest.
What do the terms on a furnace proposal mean?
Seven terms make a heating proposal readable. The first two get mixed up constantly, and the gap between them is half the sizing window.
Input rating
The fuel the furnace burns, in Btu per hour. It is the headline nameplate number: “a 100,000 Btu furnace” almost always means input. It is not the heat that reaches the house.
Output capacity
Input times efficiency: the heat that reaches the rooms. 100,000 Btu/hr input at 96% AFUE delivers 96,000; at 80%, 80,000. The sizing window applies to output.
AFUE
Annual Fuel Utilization Efficiency: the share of fuel energy delivered as heat over a season, as a percentage. It has a federal test procedure, so you can compare it between models.
Design temperature
The outdoor temperature the heating load is calculated at: a typical cold condition for your area, not the record low. It is on page one of a Manual J report, and asking for it shows whether one exists.
Heating degree days (HDD65)
How much heating weather a place had, counted from a 65°F base. It measures demand, not how cold the design morning is. This tool uses it to normalize your fuel use, and the design temperature separately to scale it.
Non-heating baseload
Fuel you burn with the furnace off: water heating, cooking, a gas dryer, a fireplace. Take it out of the yearly total first. Skipping it is the most common reason a bill-based load comes out too high.
Temperature rise
The difference between return and supply air temperature, with a range on the furnace nameplate. A reading outside that range on a new install means the blower speed or ducts are wrong.
What mistakes do people make sizing a furnace?
Mistakes in using the tool and in buying a furnace, in the order they happen.
Forgetting to subtract the baseload
In the worked example, 900 therms with no baseload instead of 720 raises the load from 31,500 to 39,500 Btu/hr: 25% too high, in the direction that makes an oversized proposal look fine.
Mixing periods
Fuel and degree days must cover the same span. Twelve months of bills go with annual HDD; one January bill goes with January’s degree days. The tool cannot catch a mismatch.
Using the new furnace’s AFUE instead of the old one’s
The efficiency input is for the furnace that burned the fuel. Entering 96% for an 80% furnace raises the estimate 21% in the worked example.
Comparing the input rating to the load
The window applies to output. Scoring input against the load overstates the proposal by the flue loss: 4% on a condensing furnace, 20% on an 80% one.
Treating a bill-based estimate as a load calculation
It is one whole-house number. It cannot size ducts, explain a cold bedroom, or serve as a design document. Manual J gives room-by-room loads.
Sizing the new furnace from the old one
The most common method in the trade. The old unit may have been oversized, and the house may have been insulated or sealed since. Copying the nameplate locks in an old guess.
Assuming a low result means the contractor is wrong
It is a prompt to ask which load calculation was used. A colder design temperature, a planned addition or a cabinet set by the cooling side can all justify margin.
Getting only one proposal
The best way to see method is to watch two or three contractors approach the same house. The one who measures windows and asks about insulation is doing something different.
What can this furnace sizing tool not do?
Real blind spots, not a disclaimer. Each is why the output is called a screening estimate.
It cannot give room-by-room loads
It gives one whole-house figure. Ducts are sized on room loads, so it cannot explain why one room is cold.
It cannot see the house
No orientation, windows, insulation, air leakage, ceiling height or shade. It infers the whole house from the fuel, and cannot say which part is losing the heat.
It cannot correct for how you lived last winter
Deep setbacks, closed rooms, a wood stove, a mild season or a furnace that never kept up push the estimate down. A full house in a cold winter pushes it up.
It cannot see changes since the bills
New windows, attic insulation, air sealing, an addition. If the house changed, a real load calculation is worth paying for, because the bills describe a different building.
It cannot check ducts, venting or gas supply
A right-size furnace on small ducts underperforms like a sizing error. Venting, combustion air and gas pipe sizing are separate safety questions, not covered here.
It is not an instruction to do any work
Furnaces burn fuel and vent combustion gases. Installing, venting, gas piping or adjusting one is for a licensed contractor, usually with a permit and inspection.
It has no prices in it
HyreHVAC does not install, service or sell HVAC equipment, and holds no data on quotes, invoices or finished jobs. Nothing here is a price we observed. That is also why these tools can tell you to keep the system you have.
It can tell you to keep what you have
If your furnace heats the house, has no fault and burns what it always has, a sizing figure is no reason to replace it.
Questions this calculator answers
What size furnace do I need for a 2,000 square foot house?
Floor area alone cannot say. Two 2,000 sq ft houses on one street can differ twofold in heating load because of insulation, air leaks, windows and orientation. This tool estimates your load from your own fuel bills instead, and shows where a proposal sits against it.
Is my proposed furnace oversized?
Check its output (input rating times AFUE) against your load. ENERGY STAR’s HVAC Design Report sets 100–140% of the calculated heat loss for a heating-only furnace, chosen per ACCA Manual S. Above 140% is worth a question. Far above it, ask in writing.
What is the difference between furnace input and output?
Input is the fuel burned per hour. Output is the heat that reaches the rooms: input times AFUE. A “100,000 Btu furnace” is almost always input; at 96% AFUE it delivers 96,000. The sizing window applies to output.
Does an oversized furnace waste gas?
Less than most pages claim. DOE’s review summarizes NREL finding furnace AFUE relatively insensitive to oversizing across a 70% to 120% range. The real costs are temperature swings, more cycles, cold far rooms, and strain on marginal ducts.
Where do I find my heating degree days?
Use our degree-day lookup: NOAA county normals for 3,107 counties, base 65°F, the same base this calculation uses. Pair the annual figure with twelve months of fuel. The two spans must match.
What is a design temperature and where do I get mine?
The outdoor temperature the load is calculated at: a typical cold condition for your area, not the record low. It is on page one of a Manual J report. Ask your contractor which one they used. If nobody can say, that tells you a lot.
Why do more degree days make the load smaller?
Because degree days divide the fuel. The same fuel over a colder winter means the house lost heat more slowly per degree, so it has a better envelope. The fuel and degree days must cover the same period.
Can I use this instead of paying for a Manual J?
No. This gives one whole-house number. Manual J gives room-by-room loads, which duct sizing and fixing a cold bedroom depend on. Use this to decide whether to challenge a proposal, and a load calculation to decide what to buy.
What if my load is below the smallest furnace sold?
That is increasingly common in tight, well-insulated houses. The answer is usually a two-stage or modulating furnace that runs below its nameplate most of the winter, or a heat pump. Both are Manual S decisions for a contractor with a load calculation.
Are furnace efficiency rules changing?
Yes. Under 10 CFR 430.32(e), non-weatherized and mobile-home gas furnaces must reach 80.0% AFUE until December 18, 2028, and 95.0% AFUE on or after that date. Oil-fired non-weatherized furnaces must reach 83.0%. That is an efficiency rule, not a sizing rule.
Manual J: Residential Load Calculation (Air Conditioning Contractors of America, retrieved 2026-09-05. The ANSI/ACCA 2 Manual J - 2016 designation and ACCA’s statement on building codes. Manual S and D descriptions here are summaries, not quotes.)
Residential HVAC Installation Practices: A Review of Research Findings (US Department of Energy, Building Technologies Office, retrieved 2026-09-04. June 2018 review of 44 field studies. The sizing row of Table 2 (33–48% oversize vs Manual J; 31–93% of systems above that) and the airflow and duct-leak rates.)
Energy conversion calculators: Btu heat content of fuels (US Energy Information Administration, retrieved 2026-09-05. Natural gas 1,036 Btu per cubic foot; distillate fuel oil (15 ppm sulfur or under) 137,381 Btu per gallon. US 2026 averages, estimates.)
10 CFR 430.32(e), energy conservation standards for furnaces (Office of the Federal Register / GPO, eCFR, retrieved 2026-09-05. The 80.0% AFUE gas furnace minimum to December 18, 2028, 95.0% after, and 83.0% for oil. Read through the eCFR content API.)