I want to walk through some math today because I think most residential HVAC sizing is wrong in the same way and for the same reasons, and the cost of getting it wrong is paid by customers over 10-15 years.

Caveat upfront: this isn’t a Manual J replacement. Manual J is a real methodology that involves about 40 inputs and produces a real answer. What I’m walking through is the back-of-envelope math that explains why the rules of thumb most installers use are systematically wrong, and what the rough magnitude of the error is.

The rule of thumb that broke

For decades, the residential HVAC sizing rule of thumb in the southern US was something like:

1 ton of cooling per 600 square feet of floor area.

A 2,400 square foot house in Texas would get a 4-ton system. A 1,800 square foot house would get a 3-ton. Easy, fast, requires no calculation, and made enough sense that it became the default installation approach across the trade.

This rule of thumb was correct for the housing stock it was developed against. Let’s call that the 1970-1985 housing stock: R-11 wall insulation if you were lucky, R-19 attic insulation, single-pane aluminum-frame windows, leaky duct systems, single-pane sliding glass doors, often no continuous air barrier.

A 2,400 square foot house in 1980 in Dallas had a peak cooling load of approximately 4 tons. The rule of thumb was actually a rule.

What changed

The same 2,400 square foot house built in 2015 has a peak cooling load of approximately 2.5 tons. Built to 2020+ standards, closer to 2 tons.

Why? Compounding improvements:

• Insulation went from R-11/R-19 to R-21/R-49 or better
• Windows went from single-pane U-1.2 to double-pane low-E U-0.30
• Air sealing went from “however the framers left it” to mandated 3-5 ACH50 in some jurisdictions, with continuous air barriers and verified blower-door testing
• Duct sealing went from leaky to mastic-sealed with leakage testing required by code in many areas
• Window orientation and shading got better attention in design even where not strictly required

The peak heat load of a modern 2,400 sq ft house can be roughly half what it was in 1980. The rule of thumb didn’t change. Installers kept matching size based on square footage, sometimes based on “what was there before.” So a 2015 house gets a 4-ton system to handle a 2.5-ton load. The system is sized 60% larger than the actual peak demand.

What “60% oversized” actually does

A 4-ton system serving a 2.5-ton load runs as follows during a hot day:

The system kicks on. It produces 4 tons of cooling instantaneously. The load is 2.5 tons. The thermostat satisfies in about 8-10 minutes of compressor runtime. Compressor shuts off. House sits. Temperature climbs back to setpoint maybe 20-30 minutes later. Compressor cycles on. Eight more minutes. Off.

This is short-cycling. Two specific problems result.

Problem 1: dehumidification failure. An AC system removes moisture from indoor air during the first 15-20 minutes of a runtime cycle. Before that, the system is mostly cooling the coil down to where the coil temperature drops below dew point and moisture starts condensing. Short cycles that end at 8-10 minutes never reach effective dehumidification mode. The indoor humidity climbs.

In Texas or Florida, this is the customer complaint of “the house feels muggy even when it’s cold.” Their new oversized system is colder and wetter than the old system. They blame the equipment. The equipment is correctly responding to bad sizing.

Problem 2: compressor wear. Compressors are designed for a certain number of starts per hour. Starting a compressor is mechanically harder than running it — the inrush load and the lubrication stabilization on a cold compressor are stressors that running steady-state doesn’t impose. Short-cycling 4-6 times per hour vs. running at 60-70% duty cycle once-per-hour produces 3-4x the start count over the same conditioning period.

The result: compressors in oversized systems fail years earlier than the same compressor in a properly sized system. I see 8-10 year compressor failures in oversized systems on what’s nominally a 15-year design life.

The math that proves the rule of thumb is wrong

Let me do this concretely.

A 2,400 sq ft single-story house in Dallas, built in 2020 to current Texas code:

• Envelope conductive load (walls, ceiling, floor at design temperature differential): ~7,000 BTU/hr
• Window solar gain (assuming 300 sq ft of double-pane low-E with reasonable orientation): ~9,000 BTU/hr
• Infiltration load (at 3 ACH50, conservatively): ~4,500 BTU/hr
• Internal gains (occupants, lights, equipment): ~5,000 BTU/hr

Total: 25,500 BTU/hr ≈ 2.1 tons.

A correct 2-ton system handles this load with proper runtime, full dehumidification, and reasonable compressor cycling.

The same house under the rule of thumb gets 4 tons. The customer’s installer says “you’ll be glad you have the headroom on the worst day.” On the worst day, the system runs full-blast for 12-14 minutes and shuts off. On the 99% of days that aren’t the worst day, the system never approaches its capacity.

The customer pays for a 4-ton system, pays the operating cost of a 4-ton system’s electrical inrush, replaces the compressor in year 9 instead of year 14, and lives in a house with bad humidity control for the duration.

What I’d actually do

If you’re an HVAC installer and you’ve been using square-feet-per-ton sizing, here’s a practical upgrade:

1. Buy or subscribe to Manual J software. Wrightsoft, Cool Calc, or HVAC-Calc are the main options. Cost is $200-$800 a year. Time per house: 2-4 hours initially, dropping to under an hour once you’re fluent.

2. Charge for it. Manual J is real work and the customer benefits from it. A $200-$400 line item on a $10,000 system is reasonable and is now common in the upper end of the residential market.

3. Document the inputs. When you save a Manual J report, you have a defensible record of why the system was sized the way it was. If a customer later complains about humidity, you have receipts.

4. Resist the customer who demands oversizing. Some customers will push back. “I want the bigger one to be sure.” The right response is to walk through the dehumidification and cycling issues. Many will hear it. The ones who don’t may not be customers you want.

5. Consider variable-capacity equipment. Two-stage and inverter-driven equipment handles sizing-margin variation much better than single-stage. If you’re not confident in your load calc, variable capacity gives you a margin of safety without the short-cycling penalty.

The rule of thumb made sense in 1980. It doesn’t now. The math doesn’t math. The customer outcomes are worse. The industry has the tools to do this right — most of us just haven’t updated.

— Cal