Realize that the load estimating tools they use can be off by that much. If you post your zip code and your current heating energy usage, you might get someone here to estimate the tons needed.

They don't measure soil thermal conductivity, so there is guessing involved there too. Grout used, double U tubes and spacers also make a difference (ie, 300' isn't the whole story). Horizontal loops are not possible?

When in doubt, go with the one with the smallest numbers, but not before you ask them to explain why it's not yet another ton smaller than what they proposed.

Upsizing the compressor isn't a huge cost adder, but upsizing the well's thermal capacity can be.

But heating energy use correlated against degree-day data will tell all... Unlike a heat loss estimation, it's a heat loss MEASUREMENT, a snapshot of the as-used thermal performance of the building. Most heat-loss calculation tools overestimate, based on assumptions of interior temperatures that may not be valid. Paying for an extra ton of geo that YOU will never use on the basis that some owner 3 tenancies from now MIGHT want to run it warmer/cooler than you do isn't a great investment. There are many homes with heating systems sized at 75% of a meticulously done Manual-J estimate that are never cold. If sized perfectly for how YOU use it it'll run fine, and if the next owner wants to run it hotter, it only means they'll be running the auxiliary heating strips at a higher duty-cycle.

Also, it's often pretty easy to peel another half-ton to a ton off the heat load with a round of air-sealing and padding out the insulation where it's cheap & easy (or missing.) Even if it's 50% more expensive to reduce the load by a ton than the cost of an extra ton of geo, it's usually worth it from both a comfort and operating cost point of view. A ton of geo has an operating cost, whereas air-sealing & insulation has none.

BTW: NOTTA? (Meant you "nada", mayhaps?)

Oversizing is a national disease that needs to be stamped out one house at a time.

Be certain that the system design and execution includes / ensures that every room gets the airflow it needs via the ductwork. In order to sucessfully execute a system downsized to the right size for a home requires that the ductwork be rightsized as well. In the South crappy ductwork is often masked under an oversized system.

I strongly recommend having a blower door test performed before finalizing system capacity. That data will refine the load calculation as well as guide the air-sealing and insulation efforts Dana recommends. In fact, I won't commit to a sizing design without blower door data.

Some one with a 70MBH load may get all three sizes suggested to them, but load estimators will not be off by 24MBH if entered data is the same and reasonably accurate. Nor would local pros require a theralconductivity test for a small job (these are employed only at large jobs where it's worth paying a bunch of money to minimize loop drilling required).

I don't suggest "when in doubt" going with the lowest bidder as many things impact size design besides load (i.e. cost per kwh, existing duct size etc.).

I spoke with Curt the other day about a house that fooled me by using more btu's than predicted. In this one instance (in 22 years) a blower door test would have prevented the trouble of augmenting the insulation, however most times an experienced pro knows what will work. Curt on the other hand is marketing envelope improvement with geo which with the blower door test makes a ton of sense. That is different than suggesting those who do not do the test with their proposal are guessing or errant.

Before any of us can help you with the size question, we would need to know the loss/gain of your home. Hopefully one or 2 of your bidders shared it with you.
J

For homes of that vintage that havn't already undergone a round of blower door testing and air-sealing, the cost of a blower door test along with air sealing remediation for a large fraction of the leakage is usually well under the cost of an additional ton of geo. Air sealing WILL reduce the load, while reducing the error bars around what to enter for infiltration factor in the heat loss calc.

As long as it doesn't require ripping apart walls or completely re-siding/re-roofing the place, augmenting the insulation is usually a better NPV bet than the upsized geo + operating cost required to handle the larger, less-insulated load.

It's not the geo contractor's job to treat your building-envelope deficiencies, but from a rational economics point of view knowing the costs (including lifecycle costs, not just the up-front price) the homeowner is well advised to look carefully at both before pulling the trigger on an expensive heating system. Treating the envelope first, then sizing the heating system accordingly is usually the right order in which to stage the upgrades in terms of optimizing comfort, cost, and value.

On a home with a 22 year heating history it's usually possible to measure the heat load based on (recent years) heating fuel use, which is inherently more accurate than any heat-loss calculator. If the proposal were accompanied with a heat loss number, look carefully comparing not only the design BTU/hr but also the indoor & outdoor design temperature(s) on which they were based. The 6 ton proposal could easily have been for a design temp 10F or more below the ASHRAE 97.5th percentile design temp and designed to provide 120% of that load with the geo alone, whereas the 3 ton proposal might have used a 95th percentile number and larger assumed duty cycle on the aux heat based on a value-proposition relative to the cost of the additional geo vs. the 10 or 20 year operating cost of the resistance heating. But if we have the heating-fuel history and a nearby town or city from which to derive weather history and design temps it's pretty simple math to put an upper bound on the system output requirements.

Down here, cooling loads dominate, and there are no "aux cooling strips" to be had, so to be certain of sizing I need the blower door data.

Another oddity is that we stay very near design cooling conditions for months at a time. Design is 94 and typical afternoon high is 91-93, unless it rains.

Up north I'd round down and let the strips make up the diff on the relatively fewer design days. Some of the loss to the strips would be offset by more efficient operation during cooling season