Where the heat actually goes
A heated house loses warmth three ways. Conduction through walls, ceilings, floors and glass, which is what insulation resists. Air exchange, where warm indoor air leaves through gaps and cold outdoor air replaces it, which is what air sealing addresses. And distribution losses, where heat is generated correctly and then delivered somewhere other than the rooms — leaky ducts in an unheated attic, uninsulated hydronic pipe in a crawl space.
The third one is the one nobody checks, and it is frequently the largest single correctable loss in a house that already has some insulation. Duct systems that run through vented attics or crawl spaces routinely lose a substantial fraction of the air passing through them — in poorly sealed systems, a fifth or more, delivered directly to the outdoors. You paid to heat that air. You cannot see it happening, no gauge reports it, and the only symptoms are rooms that never quite warm up and a furnace that runs longer than it should.
The practical consequence is an ordering rule worth stating on its own: putting a bigger, more efficient furnace in a leaky house with leaky ducts is spending money to lose heat faster. Seal the envelope first, seal the ducts, then size the equipment to the house you now have. That house needs a smaller unit, which costs less to buy and runs longer at low fire, which is where a modulating furnace is efficient anyway.
Comparing fuels without fooling yourself
Price per unit is meaningless across fuels because the units are different sizes and the equipment converting them has different efficiencies. A therm of natural gas, a gallon of heating oil, a gallon of propane and a kilowatt-hour of electricity all contain different amounts of heat, and a furnace at 80 percent efficiency delivers less of it into the house than one at 95 percent.
The comparable figure is cost per unit of delivered heat, and the method is the same for every fuel: take the price you actually pay per unit, divide by the heat content of that unit, then divide again by the efficiency of the equipment that burns it. Whatever consistent heat unit you pick, the ranking that comes out is the true one for your house, at your prices, this year. Prices move and the ranking moves with them, which is why no page should tell you which fuel is cheapest.
| What you are given | What it means | Where it misleads |
|---|---|---|
| AFUE, on a furnace or boiler | Annual fuel utilization efficiency — the share of the fuel energy that ends up as heat in the distribution system over a season | It is measured at the equipment. Duct losses happen after this number is taken, so an AFUE 95 furnace on ducts losing a fifth of their air is not delivering 95 percent of anything to your rooms. |
| HSPF, on a heat pump | A seasonal heating efficiency rating for a heat pump | Seasonal ratings are taken against a standard climate profile. Yours is not that profile. Cold-climate performance in particular depends heavily on the specific model and on outdoor temperature. |
| Electric resistance heat | Converts essentially all the electricity to heat in the room | Sounds unbeatable and often is not, because the comparison that matters is cost per delivered heat unit, and electricity is usually priced well above fuels per unit of energy. A heat pump moves several units of heat per unit of electricity, which is a different game entirely. |
The gas bill calculator handles the therm, CCF and MCF conversions if your statement uses one and your equipment documentation uses another, and heating cost savings will price a change against whatever you are actually paying.
What a setback is worth
Turning the thermostat down when nobody is home or everyone is asleep saves money, and the reason is duller than the arguments about it: heat loss through the shell is proportional to the temperature difference across it. A house held at 62 rather than 70 is losing less heat every minute it sits there. The recovery cost when you turn it back up is real but does not erase the saving, because the house was losing less the entire time it was cool.
Two things change the size of the effect. Duration matters more than depth — eight hours at four degrees down does more than one hour at eight degrees down. And a well insulated house saves less in percentage terms from a setback, because it was losing less to begin with, which is not an argument against the setback but an argument for the insulation.
The exception people get wrong is the heat pump. Deep setbacks on a heat pump can trigger the electric resistance backup on recovery, and resistance heat costs far more per unit of heat than the compressor does. Some thermostats handle this well with intelligent recovery, some do not. With a heat pump, modest setbacks or a steady setting, and a thermostat that knows it is driving a heat pump, is the safer default.
Portable heaters, honestly
The arithmetic from the energy audit page settles most of the marketing here. Watts times hours divided by 1,000 gives kWh; times your rate gives dollars.
A 1,500-watt heater draws 1,500 watts. So does every other 1,500-watt heater, whether it is ceramic, oil-filled, infrared, quartz, or sold with a wooden cabinet and a remote control. They all convert essentially all of that electricity into heat in the room, so at equal watts they produce equal heat and cost equal money. What differs between them is how the heat is distributed and how it feels — radiant models warm objects and people directly, so they can feel effective at lower output; oil-filled models stay warm after cycling off. Neither of those is a different efficiency.
Run one eight hours a day and it is 12 kWh a day. At 16 cents a kWh that is about $1.92 a day and $58 in a month, for one room. That can be a genuine bargain if it lets you keep the whole house six degrees cooler, and it is an expensive mistake if it runs alongside the central system rather than instead of it. The appliance running cost calculator does this for whatever wattage is on the label.
The safety points are not negotiable: plugged directly into a wall receptacle, never into an extension cord or a power strip, three feet of clearance from anything that burns, off when you leave the room and off when you sleep. Portable heaters and the circuits they overload are a leading cause of home heating fires. A 1,500-watt heater is 12.5 amps, which is most of a 15-amp circuit by itself — see the circuit load calculator before adding it to a circuit already carrying something.
Ducts and distribution
| Symptom | Likely cause | What to do |
|---|---|---|
| One room never gets warm, others are fine | A disconnected or crushed branch duct, a closed damper, or a run too long for the CFM it is given | Trace the run in the attic or crawl space. A branch that has come off its takeoff is common and takes ten minutes to reattach and seal. |
| Every room is weak, the furnace runs constantly | System-wide duct leakage, a filthy filter, or equipment undersized for a house that has become leakier | Change the filter first, then look at joints. Duct sizing shows what a given CFM actually needs. |
| The attic or crawl space is noticeably warm in winter | You are heating it | Every joint gets mastic, then the duct gets insulated. Cloth duct tape fails on ducts, which is one of the better jokes in the trade. |
| Dust everywhere, and the house smells like the crawl space | Return-side leakage pulling unconditioned air into the system | Return leaks are worse than supply leaks because they drag in whatever is in that space. Seal the return plenum and any panned joist cavity used as return. |
The things that cost nothing
Close the fireplace damper when there is no fire in it, because an open flue is a chimney-sized hole in the ceiling. Reverse ceiling fans to push warm air down off the ceiling on low speed, and turn them off in empty rooms, since a fan cools people rather than air. Open south-facing curtains during sun hours and close everything at dusk. Keep furniture and rugs off supply registers and return grilles. Change the filter, because a loaded one starves the blower and costs both efficiency and equipment life. Put a hand behind every outlet cover on an exterior wall and fit foam gaskets where you feel air. And in a house with rooms nobody uses, closing off a room is worth doing with the vents left mostly open, because sealing a room completely in a duct system raises pressure in the rest of the system and increases leakage elsewhere.
Dressing for 66 rather than heating for 72 is not a moral position, it is the cheapest degree of comfort available, and the number of degrees is genuinely worth money. If cold weather is a health question in your household rather than a comfort one — an infant, an older adult, anyone with a cardiac or respiratory condition — that changes the calculation entirely, and the cold weather guide covers where the safety line sits.
When equipment replacement is the right answer
Replace when the existing unit has failed, when a repair estimate approaches a meaningful fraction of replacement, when a heat exchanger inspection turns up a crack, or when the unit is old enough that parts availability is a gamble in January. Replacing a functioning mid-efficiency furnace purely on projected savings rarely pencils out, and the projections are made by people selling the furnace.
When you do replace, insist on a load calculation for the house rather than a rule of thumb from the square footage or a match to whatever is there now. Oversized equipment short-cycles: it satisfies the thermostat quickly, shuts off, and never reaches steady-state efficiency, which is both less efficient and less comfortable than a smaller unit running longer. If you have air sealed and insulated first, the correct size is smaller than the old one, and that is the point.
The through-line here is that the heating bill is mostly determined before the furnace ever fires — by the shell, and by whether the heat reaches the rooms. Fix those two and the equipment question gets smaller and cheaper. Skip them and you buy a better machine to do the same losing.
Questions people ask
Does turning the thermostat down at night actually save money?
Yes, and the physics is straightforward: heat loss through the walls and ceiling is proportional to the temperature difference between inside and outside, so a cooler house loses less every minute it stays cool. The energy spent bringing it back up does not cancel that out, because the reduced loss ran the whole time. Duration matters more than depth — eight hours down four degrees beats one hour down eight. The exception is a heat pump, where a deep setback can bring on electric resistance backup during recovery and cost more than it saved. With a heat pump, use modest setbacks and a thermostat that is configured for one.
Which heating fuel is cheapest?
It depends on your prices and your equipment, and it changes from year to year, so no page can answer it for you. The method is what transfers: take the price you pay per unit, divide by the heat content of that unit, and divide again by the efficiency of the equipment burning it. That gives cost per unit of heat delivered, which is the only comparable figure across natural gas, propane, oil, electric resistance and a heat pump. Note that a heat pump does not convert energy to heat, it moves heat, so it can deliver several units of heat per unit of electricity — which is why electricity being expensive per unit does not settle the question.
Are infrared or oil-filled heaters cheaper to run than a regular space heater?
No. A 1,500-watt heater draws 1,500 watts and turns essentially all of it into heat in the room, whatever the heating element is made of and whatever the box claims. At equal wattage they cost the same to run. What differs is how the warmth is delivered: radiant and infrared models heat people and objects directly, so they can feel adequate at a lower setting, and oil-filled models coast for a while after the element cycles off. Those are comfort differences, not efficiency differences. Run the wattage on the label through watts times hours divided by 1,000, times your rate.
How much heat am I losing through leaky ducts?
Without a pressurization test, you cannot know the number, but the range in poorly sealed systems is large enough to matter more than most other single fixes — a substantial fraction of the air, delivered to an attic or crawl space instead of a room. The symptoms are rooms that never warm properly, a system that runs much longer than it should, an unheated space that feels oddly warm in winter, and heavy dust if the leaks are on the return side. Feel every reachable joint while the blower runs, seal with duct mastic rather than cloth tape, and insulate the runs afterward. It is one of the cheapest large improvements available in a house with ducts outside the conditioned space.
Should I close vents in rooms I do not use?
Partially, at most, and it helps less than people expect. A forced-air system was balanced for a certain airflow; closing registers raises static pressure across the whole system, which increases leakage at every existing duct seam and can strain the blower. If a room is genuinely unused, damping its supply down somewhat and keeping the door closed is reasonable, but do not shut off a third of the house and expect to save a third of the heat. Watch for condensation and mold in any room you let get cold, particularly on exterior walls, and do not let a room with plumbing in an exterior wall drop toward freezing.