Rheem Classic vs Prestige: is the higher-efficiency heat pump worth $1,250 more?

## Assumptions (stated explicitly — the whole answer hangs on these)

- Blended rate: **$0.17/kWh** base; sensitivity at **$0.14** and **$0.20**.
- Service life: **15 years**.
- Post-insulation **cooling design load ≈ 24,000 BTU/h (2 tons)**, cooling **EFLH = 800 h** (typical zone-4A Baltimore/DC).
- **Heating design load ≈ 28,000 BTU/h**, heating **EFLH = 1,400 h** — heating is the larger annual energy user, as it should be in 4A.
- Both units deliver the **same annual BTU to the same house**. Energy = delivered BTU ÷ efficiency metric (SEER2 for cooling, HSPF2 for heating), since SEER2/HSPF2 are both BTU-out per Wh-in.

## (a) The arithmetic

**Delivered loads**
- Cooling: 24,000 × 800 = **19.2 MMBtu/yr**
- Heating: 28,000 × 1,400 = **39.2 MMBtu/yr**

**Cooling kWh**
- Classic (SEER2 15.2): 19,200,000 ÷ 15.2 ÷ 1,000 = **1,263 kWh**
- Prestige (SEER2 20.0): 19,200,000 ÷ 20.0 ÷ 1,000 = **960 kWh**
- Δ = **303 kWh/yr**

**Heating kWh**
- Classic (HSPF2 7.5): 39,200,000 ÷ 7.5 ÷ 1,000 = **5,227 kWh**
- Prestige (HSPF2 8.5): 39,200,000 ÷ 8.5 ÷ 1,000 = **4,612 kWh**
- Δ = **615 kWh/yr**

**Total savings = 303 + 615 = 918 kWh/yr.**

Note the split: **heating is ~67% of the savings.** The flashy 15.2→20 SEER2 number is the *smaller* half of this decision.

**Dollars and payback**

| Rate | Annual $ saved | Simple payback on $1,250 | 15-yr savings | Net over life |
|---|---|---|---|---|
| $0.14 | $128.5 | **9.7 yr** | $1,928 | +$678 |
| $0.17 | $156.1 | **8.0 yr** | $2,341 | +$1,091 |
| $0.20 | $183.6 | **6.8 yr** | $2,754 | +$1,504 |

**Verdict on (a):** Yes, it pays back — but it is *marginal, not a slam-dunk*. Simple payback lands at 7–10 years against a 15-year life. That means the pure-energy case survives, but it is fragile: it dies if you undershoot run-hours, if the unit fails early, or if you apply any real discount rate (at ~8% NPV the net gain shrinks toward break-even). It gets *better* if Maryland electricity keeps outpacing inflation, which it historically has.

**Two effects my table under-counts, both favoring the Prestige — and this is where I diverge from a pure spreadsheet answer:**

1. **Backup-heat avoidance.** The Prestige delivers **31,800 BTU/h at 47°F vs the Classic's 27,800**, and inverters hold capacity better as temperature drops. In a 4A cold snap, the Classic hits its capacity wall sooner and hands the load to **resistance strips at an effective COP of 1.0** — the most expensive kWh in the house. HSPF2 only partially captures this. Real-world, the heating-dollar gap is likely *wider* than the $105/yr I computed.

2. Every kWh saved on the heating side compounds, and heating is the bigger bucket. This is a **heating-efficiency purchase wearing a cooling-efficiency label.**

## (b) The 3-ton sizing question — the homeowner is asking the wrong question

The instinct — "2.5-ton house losing load, why buy 3-ton?" — is **correct for a single-stage unit and wrong for an inverter.** These are not comparable objects.

- The **Classic is fixed-output**: 28,400 BTU/h on/off. On a 24,000 load it's ~18% oversized and *this* is your genuine short-cycling/poor-dehumidification risk. A 3-ton single-stage (33% oversize) would be worse — but you're not being offered that.
- The **Prestige's 31,800 is a *ceiling*, not an operating point.** An inverter modulates down to roughly **30–40% of capacity (~10,000–13,000 BTU/h)**. On a 2-ton load it runs **long, low, steady cycles** — which is precisely what removes latent moisture in a humid climate. It will **dehumidify better than the "correctly sized" single-stage**, not worse.

So the tonnage comparison between the two is a **category error.** The number that governs short-cycling and humidity is **minimum modulated capacity (turndown)**, not the nameplate ton. The Prestige's high ceiling is *headroom* — useful for cold-weather heating and rapid recovery — with essentially no comfort penalty at part load. The oversizing fear does not just disappear for the Prestige; **it inverts** — the single-stage is the machine actually at risk of poor humidity control on the shrunken load.

One caveat I'd insist on: **get a real Manual J** after the envelope work. If the true load comes in near 18,000 BTU/h, even the inverter's minimum starts to look high, and you'd want to confirm its documented turndown ratio.

## What it turns on / where it flips

**Recommend the Prestige.** The decision turns on:

- **Buy it** if you value guaranteed day-one comfort (humidity control + fewer resistance-heat cold snaps) and plan to stay past ~8 years. The comfort value is *certain*; the money is *probable*.
- **Flips to the Classic** if: your rate is truly ~$0.14 *and* run-hours are low, you expect to sell within ~7 years, or you'd rather redirect the $1,250 to further envelope work (whose ROI, if any weak spots remain, usually beats an equipment efficiency bump).
- The dominant sensitivity is **heating hours and backup-strip usage, not SEER2.** Anyone answering this purely off the cooling efficiency delta has solved the smaller third of the problem.

## Assumptions

I will use the quoted matched-system ratings exactly as provided:

- Electricity: **$0.17/kWh** base case.
- Sensitivity: **$0.14/kWh** and **$0.20/kWh**.
- Equipment life: **15 years**.
- Post-retrofit equivalent cooling load: **24,000 Btu/h**, or approximately two tons.
- Cooling equivalent full-load hours: **800 hours/year**.
- Heating equivalent full-load hours: **1,800 hours/year**.
- For a transparent apples-to-apples calculation, I will use **24,000 Btu/h as the equivalent heating-load basis too**.

That produces:

- Annual cooling delivered:
**24,000 × 800 = 19,200,000 Btu/year**
- Annual heating delivered:
**24,000 × 1,800 = 43,200,000 Btu/year**

This is an illustrative load model, not a substitute for a post-retrofit Manual J. The result turns mainly on the house’s actual annual delivered heating and cooling Btu, not on the equipment nameplates.

## 1. Annual cooling electricity

SEER2 is being used as Btu delivered per watt-hour.

### Classic

\[
19{,}200{,}000 \div 15.2 = 1{,}263{,}158\text{ Wh}
\]

\[
= 1{,}263\text{ kWh/year}
\]

At $0.17/kWh:

\[
1{,}263 \times \$0.17 = \$214.74/year
\]

### Prestige

\[
19{,}200{,}000 \div 20.0 = 960{,}000\text{ Wh}
\]

\[
= 960\text{ kWh/year}
\]

At $0.17/kWh:

\[
960 \times \$0.17 = \$163.20/year
\]

### Cooling savings

\[
1{,}263 - 960 = 303\text{ kWh/year}
\]

\[
\$214.74 - \$163.20 = \boxed{\$51.54/year}
\]

## 2. Annual heating electricity

HSPF2 is being used as Btu delivered per watt-hour.

### Classic

\[
43{,}200{,}000 \div 7.5 = 5{,}760{,}000\text{ Wh}
\]

\[
= 5{,}760\text{ kWh/year}
\]

At $0.17/kWh:

\[
5{,}760 \times \$0.17 = \$979.20/year
\]

### Prestige

\[
43{,}200{,}000 \div 8.5 = 5{,}082{,}353\text{ Wh}
\]

\[
= 5{,}082\text{ kWh/year}
\]

At $0.17/kWh:

\[
5{,}082 \times \$0.17 = \$864.00/year
\]

### Heating savings

\[
5{,}760 - 5{,}082 = 678\text{ kWh/year}
\]

\[
\$979.20 - \$864.00 = \boxed{\$115.20/year}
\]

Heating is the larger energy user under these Zone-4A assumptions. It supplies about 69 percent of the calculated annual energy savings.

## 3. Base-case annual comparison

| | Classic | Prestige |
|---|---:|---:|
| Cooling electricity | 1,263 kWh | 960 kWh |
| Heating electricity | 5,760 kWh | 5,082 kWh |
| **Total electricity** | **7,023 kWh** | **6,042 kWh** |
| Cooling cost at $0.17 | $214.74 | $163.20 |
| Heating cost at $0.17 | $979.20 | $864.00 |
| **Total annual cost** | **$1,193.94** | **$1,027.20** |

Annual savings:

\[
7{,}023 - 6{,}042 = 981\text{ kWh/year}
\]

\[
\$1{,}193.94 - \$1{,}027.20
= \boxed{\$166.74/year}
\]

## 4. Payback of the $1,250 premium

\[
\$1{,}250 \div \$166.74
= \boxed{7.5\text{ years}}
\]

Over 15 years, ignoring discounting, maintenance, repairs, and future rate changes:

\[
15 \times \$166.74 = \$2{,}501
\]

Net savings after recovering the premium:

\[
\$2{,}501 - \$1{,}250
= \boxed{\$1{,}251}
\]

### Electricity-rate sensitivity

| Electricity rate | Annual savings | Simple payback | 15-year gross savings | 15-year savings after premium |
|---:|---:|---:|---:|---:|
| $0.14/kWh | $137.31 | 9.1 years | $2,060 | $810 |
| **$0.17/kWh** | **$166.74** | **7.5 years** | **$2,501** | **$1,251** |
| $0.20/kWh | $196.16 | 6.4 years | $2,942 | $1,692 |

### Answer to question (a)

**Under the stated load and rate assumptions, yes: the Prestige pays back the $1,250 premium within its assumed 15-year life.** The modeled payback is approximately **6.4 to 9.1 years** across the requested electricity-rate range.

That conclusion becomes less favorable if the post-retrofit loads are much lower than assumed. The useful scaling formula at the $0.17/kWh rate is:

\[
\text{Annual savings}
=
\$51.54
\left(\frac{\text{actual cooling equivalent load}}{24{,}000}\right)
\left(\frac{\text{cooling hours}}{800}\right)
\]

\[
+
\$115.20
\left(\frac{\text{actual heating equivalent load}}{24{,}000}\right)
\left(\frac{\text{heating hours}}{1{,}800}\right)
\]

For example, if both equivalent loads were only 18,000 Btu/h rather than 24,000, the modeled savings would fall to about:

\[
\$166.74 \times \frac{18{,}000}{24{,}000}
= \$125.06/year
\]

That would produce roughly a **10-year payback at $0.17/kWh**, still within 15 years. At very low loads, low electricity rates, substantial auxiliary-resistance operation, or a shorter actual service life, the payback could extend beyond 15 years.

This calculation also assumes the quoted HSPF2 and SEER2 values represent the actual certified matched combination. Use the exact AHRI certificate for the outdoor unit, indoor coil, and air handler—not a generic catalog rating.

## 5. Does the house need a three-ton Prestige?

**No—not based merely on 2,100 square feet, the old equipment, or the three-ton label.** The decision should turn on:

1. A post-insulation, post-air-sealing **Manual J cooling load**.
2. The Prestige’s actual minimum cooling capacity at relevant conditions.
3. Exact matched-system latent-capacity and airflow data.
4. Duct capacity and static pressure at the selected airflow.
5. The winter capacity and backup-heat strategy at Maryland design temperatures.

Using the supplied cooling-load assumption:

\[
31{,}800 \div 24{,}000 = 1.325
\]

The Prestige’s maximum quoted cooling capacity is about **33 percent above** the assumed 24,000 Btu/h design load.

That would be concerning for a conventional single-stage unit. But Rheem describes the RD18AY as an inverter system operating between approximately **45 percent and 100 percent of capacity**. ([rheem.com](https://www.rheem.com/product/rheem-rd18ay-endeavor-line-prestige-series-side-discharge-universal-heat-pumps-rd18ay36ajvca/?utm_source=openai))

Using the quoted 31,800 Btu/h capacity as a rough reference:

\[
31{,}800 \times 0.45 \approx 14{,}300\text{ Btu/h}
\]

So the approximate operating range suggested by the manufacturer’s modulation description is:

\[
14{,}300\text{ to }31{,}800\text{ Btu/h}
\]

The assumed 24,000 Btu/h design load lies comfortably inside that range. That makes the three-ton inverter **potentially reasonable**, even though its maximum capacity is larger than the load. It can spend much of the season operating below its nominal three-ton output rather than turning fully on and off.

The 14,300 Btu/h figure is only a rough interpretation of the stated modulation range, not a guarantee of exact minimum delivered capacity at every outdoor temperature, indoor airflow, or humidity condition. The contractor should provide the exact performance data for the quoted pairing.

## 6. Humidity-control risk

Maryland is humid, so the important question is not merely whether the compressor can reduce speed. It is whether the complete matched system can:

- Operate for long enough cycles at low load.
- Maintain adequate evaporator temperature for latent removal.
- Use an appropriate blower airflow.
- Avoid excessive airflow that raises sensible capacity while reducing dehumidification.
- Provide acceptable latent capacity at the actual low-stage operating point.

A properly commissioned inverter can be better for humidity than an oversized fixed-capacity system because it can run longer at lower output. But “variable speed” is not a blanket guarantee of good humidity control. A poor blower setting, excessive duct airflow, or inadequate low-stage latent performance can still produce high indoor relative humidity.

For this house, I would require the installer to document:

- Post-retrofit Manual J cooling load.
- Exact AHRI match.
- Minimum and maximum cooling capacity at representative conditions.
- Low-stage sensible and latent capacity, or sensible heat ratio.
- Required airflow and measured external static pressure.
- How the thermostat and air handler control dehumidification.
- Whether the existing duct system can handle the selected airflow quietly and efficiently.

## 7. A significant quote-check

There is a possible description mismatch in the proposal. Rheem’s official listing for the exact model **RP14AY30AJ2N** identifies it as a **two-stage** heat pump, not a single-stage heat pump. ([bim.rheem.com](https://bim.rheem.com/item/residential-split-heat-pump/rp14ay-endeavor-line-classic-series-heat-pump/rp14ay30aj2n?utm_source=openai))

That does not change the arithmetic because I used the quoted 15.20 SEER2, 7.50 HSPF2, and 28,400 Btu/h figures. But it does affect the cycling comparison:

- If the Classic really is single-stage, its 28,400 Btu/h output is about 18 percent above a 24,000 Btu/h load.
- If it is actually the Rheem two-stage RP14AY30AJ2N, its low-stage behavior may be better than the proposal’s description suggests.
- The homeowner should ask for the exact model numbers, certified match, stage/control configuration, and low-stage capacity before deciding.

## Bottom line

### Efficiency and payback

Choose the Prestige on economics if the assumptions are reasonably close:

- Base-case annual savings: **about $167**
- Base-case simple payback: **about 7.5 years**
- 15-year gross energy savings: **about $2,500**
- Net savings after the $1,250 premium: **about $1,250**

Heating savings dominate because the assumed annual heating load is much larger than the annual cooling load.

### Sizing

The Prestige is **not automatically too large** merely because it is nominally three tons. At an assumed 24,000 Btu/h design load, its advertised variable-capacity range makes it potentially suitable.

My recommendation would be:

- **Proceed with the Prestige** if the post-retrofit Manual J is roughly 18,000–24,000 Btu/h, the exact matched system can modulate below that load, and the contractor documents acceptable low-stage latent performance and duct static pressure.
- **Prefer a smaller system** if the actual cooling load is near or below roughly 14,000 Btu/h, if the inverter’s true minimum capacity is materially higher than expected, or if the contractor cannot document low-stage humidity performance.
- **Do not approve either proposal solely from square footage or the old 2.5-ton nameplate.**

The recommendation turns on the **post-retrofit Manual J and the Prestige’s minimum/low-stage performance**. The three-ton nameplate is secondary.

**Explicit assumptions (as required):**
- Post-insulation design cooling load = 24,000 BTU/h (~2 tons).
- Reasonable zone-4A EFLH (Baltimore metro): cooling 1,000 h; heating 2,000 h (heating is the larger annual user, consistent with HDD ~4,500–5,000 vs CDD ~1,200–1,500 and mixed-humid data).
- Annual delivered cooling load = 24,000 × 1,000 = 24 million BTU.
- Annual delivered heating load = 24,000 × 2,000 = 48 million BTU (same for both units).
- Maryland blended residential rate $0.17/kWh (sensitivity $0.14 / $0.20).
- 15-year service life.
- Both units deliver identical annual BTU to the house; only efficiency differs. Prestige is variable-speed inverter; Classic is single-stage.

**Cooling arithmetic**
Option 1 (SEER2 15.2): 24,000,000 BTU / 15.2 = 1,578,947 Wh ≈ **1,579 kWh**
Option 2 (SEER2 20.0): 24,000,000 / 20.0 = 1,200,000 Wh = **1,200 kWh**
Cooling savings: 379 kWh/yr

**Heating arithmetic**
Option 1 (HSPF2 7.5): 48,000,000 BTU / 7.5 = 6,400,000 Wh = **6,400 kWh**
Option 2 (HSPF2 8.5): 48,000,000 / 8.5 ≈ 5,647,059 Wh ≈ **5,647 kWh**
Heating savings: 753 kWh/yr

**Total annual savings**
1,132 kWh/yr
- At $0.17/kWh: **$192/yr**
- At $0.14/kWh: **$158/yr**
- At $0.20/kWh: **$226/yr**

**Payback on $1,250 premium**
- $0.17: 1,250 / 192 ≈ **6.5 years**
- $0.14: 1,250 / 158 ≈ **7.9 years**
- $0.20: 1,250 / 226 ≈ **5.5 years**

All scenarios pay back well inside the 15-year life (remaining 7–9+ years of net savings). The result turns on the EFLH split and delivered-load equality; if actual post-sealing cooling load drops below ~20k BTU/h or heating EFLH is <1,600 h the payback lengthens modestly but stays attractive at $0.17+.

**3-ton sizing question**
Design load is shrinking to ~2 tons. The Prestige (nominal 3-ton, inverter) can modulate well below 50% capacity on most models (typical 30–100%+ range for this class), allowing it to run longer, steadier cycles. This improves latent (humidity) removal in humid 4A summers versus a single-stage unit that would short-cycle if oversized. The Classic 2.5-ton single-stage has less modulation headroom if the load is now closer to 2 tons.

The inverter largely neutralizes oversizing risk for comfort and efficiency; short-cycling/humidity problems are more likely with the single-stage unit. Recommendation favors the Prestige unless the homeowner is extremely budget-constrained and willing to accept lower efficiency plus potential dehumidification trade-offs. It would flip only if real-world modulation range were poor (<25% turndown) or if measured loads proved far smaller than assumed.

Starting anthropic/claude-opus-4.8

Starting x-ai/grok-4.3

Starting openai/gpt-5.6-luna

x-ai/grok-4.3 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

Initial responses are now shared. Each council member will critique every peer and endorse the strongest answer.

- x-ai/grok-4.3: STRENGTH: Cleanest, most readable arithmetic and correctly flags that heating dominates the savings. WEAKNESS: 1,000 cooling + 2,000 heating EFLH are both high for an interior, air-sealed 4A townhouse, which inflates the payback into the rosiest bracket; ignores backup-strip economics entirely. FACTUAL-CONCERN: yes — the EFLH choices are aggressive and make the payback look better than the more defensible 800/1,400 range would.

- anthropic/claude-opus-4.8: STRENGTH: Only answer to price the resistance-backup effect (COP≈1.0 in cold snaps) and to name the sizing question as a category error — turndown, not nameplate, governs humidity; correctly identifies heating as ~two-thirds of the savings. WEAKNESS: Presents payback as more "marginal/fragile" than the numbers strictly require, and like the others treats the Classic as pure on/off despite the variable-speed air handler. FACTUAL-CONCERN: no.

- openai/gpt-5.6-luna: STRENGTH: Most rigorous — insists on AHRI-matched certificate, cites Rheem's actual ~45% modulation floor (more grounded than the 30–40% others guessed), and flags a possible single-stage-vs-two-stage discrepancy on the RP14AY. WEAKNESS: Uses 24,000 Btu/h as the heating-load basis (overstates heating energy versus a lower part-load-weighted heating figure), and is long enough to bury its own best points. FACTUAL-CONCERN: yes — the "RP14AY is actually two-stage" claim rests on a catalog citation I cannot verify against the prompt's stated single-stage; if wrong, it introduces confusion rather than resolving it.

I endorse anthropic/claude-opus-4.8 because it isolates the actual economic driver — heating plus avoided resistance strips, not the headline SEER2 jump — and reframes the sizing question correctly (minimum modulated capacity, not tons). Luna's spec-verification discipline is the strongest thing the council produced and nearly won my endorsement, but its unverifiable two-stage claim and inflated heating basis hold it back. The whole council, including me, missed that the identical variable-speed ECM air handler gives the Classic a low-airflow dehumidification mode, which materially undercuts the shared premise that the single-stage unit must handle humidity worse.

[x-ai/grok-4.3]: STRENGTH: Highest heating EFLH produces fastest payback and clear heating-dominance math. | WEAKNESS: Overstates heating hours vs typical zone-4A data; no model verification. | FACTUAL-CONCERN: no
[anthropic/claude-opus-4.8]: STRENGTH: Highlights backup-strip risk and heating savings share; good qualitative inverter point. | WEAKNESS: Lowest heating EFLH weakens payback case; no sensitivity scaling. | FACTUAL-CONCERN: no
[openai/gpt-5.6-luna]: STRENGTH: Only peer that flags possible two-stage Classic mismatch and supplies scaling formula plus explicit Manual J caveats. | WEAKNESS: Slightly lower heating hours than my calc but still realistic. | FACTUAL-CONCERN: no

Endorsed Luna for catching the RP14AY discrepancy and providing the most actionable real-world checks; council collectively missed low-stage dehumidification metrics that could independently favor the inverter regardless of payback.

[x-ai/grok-4.3]: STRENGTH: The arithmetic is clear, heating is correctly treated as the larger energy bucket, and rate sensitivity is directly presented. | WEAKNESS: It confidently recommends the inverter without documenting this exact model’s minimum capacity, low-stage latent performance, or whether the shared air handler is properly matched. | FACTUAL-CONCERN: yes — “typical 30–100%” modulation and the claim that the inverter largely neutralizes oversizing are unsupported for the quoted equipment.

[anthropic/claude-opus-4.8]: STRENGTH: It correctly emphasizes that heating efficiency, not the eye-catching SEER2 increase, drives much of the modeled savings and identifies Manual J and turndown as decisive. | WEAKNESS: Its assumed 28,000-Btu/h heating basis and 1,400 heating hours are relatively conservative without enough justification, while its backup-heat argument goes beyond the user’s stipulated efficiency-only comparison. | FACTUAL-CONCERN: yes — claiming real-world heating savings will likely exceed the HSPF2 calculation is not established and may double-count effects already represented by the rating.

[openai/gpt-5.6-luna]: STRENGTH: It fulfills nearly every requested calculation, gives the requested rate sensitivity, scales results to lower loads, and properly conditions the sizing recommendation on minimum capacity, latent performance, ducts, and Manual J. | WEAKNESS: Using 24,000 Btu/h as the heating-load basis is explicitly simplified rather than independently justified, and the discussion of the Classic’s staging risks distracting from the prompt’s stated single-stage assumption. | FACTUAL-CONCERN: yes — the asserted 45% modulation range and possible two-stage identity of the Classic require verification for the exact certified match.

The endorsement goes to openai/gpt-5.6-luna because it is the most complete and appropriately conditional response, despite those verifiability caveats. Collectively, the council provided scenarios but did not turn them into a precise 15-year break-even load or run-hour threshold.

Starting anthropic/claude-opus-4.8

Starting x-ai/grok-4.3

Starting openai/gpt-5.6-luna

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

x-ai/grok-4.3 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 serves as Chairperson and will now synthesize the council's reasoning into a single canonical answer.

# The synthesized answer

## Assumptions (committed, not hedged)

- Blended rate **$0.17/kWh**, with sensitivity at **$0.14** and **$0.20**.
- Service life **15 years**, simple payback (no discounting).
- Post-retrofit **cooling design load 24,000 BTU/h**; **heating design load 26,000 BTU/h** (in 4A the heating design load slightly exceeds the cooling load).
- **Cooling EFLH = 800 h**; **heating EFLH = 1,600 h.** This deliberately splits the council's spread. Grok's 1,000/2,000 is too hot for an *interior, freshly air-sealed* townhouse (shared walls, low solar/skin exposure); my own Phase-1 1,400 heating hours was too cold. 800/1,600 is the defensible middle for Baltimore/DC-metro.
- Both units deliver identical annual BTU to the same house. Energy = delivered BTU ÷ (SEER2 for cooling, HSPF2 for heating).

Delivered loads:
- Cooling: 24,000 × 800 = **19.2 MMBtu/yr**
- Heating: 26,000 × 1,600 = **41.6 MMBtu/yr**

## (a) The arithmetic

**Cooling kWh**
- Classic (SEER2 15.2): 19,200,000 ÷ 15.2 ÷ 1,000 = **1,263 kWh**
- Prestige (SEER2 20.0): 19,200,000 ÷ 20.0 ÷ 1,000 = **960 kWh**
- Δ = **303 kWh/yr**

**Heating kWh**
- Classic (HSPF2 7.5): 41,600,000 ÷ 7.5 ÷ 1,000 = **5,547 kWh**
- Prestige (HSPF2 8.5): 41,600,000 ÷ 8.5 ÷ 1,000 = **4,894 kWh**
- Δ = **653 kWh/yr**

**Total savings = 956 kWh/yr. Heating is 68% of it.**

| Rate | Annual $ saved | Simple payback | 15-yr gross | Net over life |
|---|---|---|---|---|
| $0.14 | $133.8 | **9.3 yr** | $2,007 | +$757 |
| $0.17 | $162.5 | **7.7 yr** | $2,438 | +$1,188 |
| $0.20 | $191.2 | **6.5 yr** | $2,868 | +$1,618 |

**Verdict: Yes, it pays back — 6.5–9.3 years against a 15-year life. Real but not a slam-dunk; apply an 8% discount rate and the net gain shrinks toward break-even.**

### The insight the council circled but never stated

Compute the savings *per MMBtu delivered*:
- Cooling: (1/15.2 − 1/20.0) × 1,000,000 ÷ 1,000 = **15.8 kWh/MMBtu**
- Heating: (1/7.5 − 1/8.5) × 1,000,000 ÷ 1,000 = **15.7 kWh/MMBtu**

They are **nearly identical.** The eye-catching SEER2 jump (+31.6%) and the modest HSPF2 jump (+13.3%) produce the *same* absolute kWh saved per unit of delivered energy, because heating's baseline consumption per MMBtu is roughly double cooling's. **This dissolves the entire EFLH argument the council spent its energy on:** total savings ≈ 15.7 kWh × (total MMBtu delivered), regardless of the heating/cooling split. You don't need to win the run-hours debate — you only need total delivered energy.

### The exact break-even threshold (Luna's flagged gap, now computed)

The premium recovers in 15 years if annual savings ≥ $1,250 ÷ 15 = **$83.33/yr**. Converting to delivered energy at each rate:

- **$0.17/kWh:** need 490 kWh/yr → **≈ 31 MMBtu/yr delivered** (about half my base case, 60.8 MMBtu)
- **$0.14/kWh:** need 595 kWh/yr → **≈ 38 MMBtu/yr**
- **$0.20/kWh:** need 417 kWh/yr → **≈ 27 MMBtu/yr**

**So the Prestige fails to recover its premium only if total annual delivered energy falls below roughly half of what a normally-occupied 2,100 sqft 4A townhouse uses.** For a lightly-occupied, very-low-load house at the cheap $0.14 rate, that failure is plausible — everywhere else, it clears.

## (b) The 3-ton sizing question

The homeowner's instinct — "2.5-ton house losing load, why a 3-ton?" — is **correct for a single-stage unit and a category error for an inverter.**

- The **Classic is fixed-output**: 28,400 BTU/h on/off, ~18% oversized on a 24,000 load. *This* is the genuine short-cycle candidate.
- The **Prestige's 31,800 is a ceiling, not an operating point.** Rheem describes the RD18AY as modulating down to roughly **45% of capacity (~14,300 BTU/h)** — verify against the exact AHRI certificate. A 24,000 load sits comfortably inside a ~14,300–31,800 range, so it runs long, low, steady cycles: the ceiling is *heating headroom* (cold-snap capacity + rapid recovery), not a comfort penalty.

The number that governs short-cycling and humidity is **minimum modulated capacity**, not nameplate tons.

### Three things the council collectively missed

1. **The shared air handler undercuts the whole "single-stage = bad humidity" premise.** Both options use the *same* RH2VY variable-speed ECM blower. That blower gives the **Classic** an enhanced-dehumidification low-airflow mode — dropping CFM/ton lowers coil temperature and raises latent removal. The single-stage's humidity disadvantage is therefore **smaller than all three Phase-1 answers assumed.** The inverter's real dehum edge is longer part-load runtime, not a monopoly on low airflow.

2. **Nobody has the latent number.** "Variable speed" is not a humidity guarantee. Require the installer to document **low-stage sensible and latent capacity (or SHR) at part load** for the certified match — that, not the tonnage, decides summer comfort.

3. **The backup-strip effect widens the heating gap beyond HSPF2.** The Prestige's higher 47°F capacity (31,800 vs 27,800) and better cold-temperature capacity retention mean it hands off to resistance strips (COP ≈ 1.0) less often in a 4A cold snap. HSPF2 only partially captures this, so the real heating-dollar gap likely runs a bit *above* my $111/yr line — a genuine effect, but don't double-count it into the certified rating.

### Spec-verification items (do not sign without these)
- Post-retrofit **Manual J** (if true load is ~18,000 BTU/h, even the inverter's ~14,300 minimum warrants a hard look at turndown).
- Exact **AHRI-certified match** for the outdoor unit + RH2VY coil/blower.
- Documented **minimum capacity and low-stage SHR/latent** at representative conditions.
- Duct static pressure at the selected airflow.
- **Confirm the Classic's staging** — the RP14AY model line is described in some Rheem catalogs as two-stage; if so, its cycling/humidity behavior is better than the prompt's "single-stage" label implies, narrowing the comfort gap further (does not change the arithmetic).

## Bottom line — what it turns on and where it flips

**Recommend the Prestige.** It turns on two things, in order: **(1) total annual delivered energy** (the payback is essentially linear in it — clears everywhere above ~31 MMBtu at $0.17), and **(2) how long you keep the house** (past ~8 years the money is favorable; the day-one comfort value is favorable regardless).

**It flips to the Classic if:** your rate is genuinely ~$0.14 *and* the house is a very-low-use, sub-35-MMBtu load; you expect to sell within ~7 years; or you'd rather put the $1,250 into remaining envelope weak spots (whose ROI usually beats an equipment-efficiency bump).

The one thing it does **not** turn on is the 3-ton nameplate. That's the wrong number to worry about — worry instead about the documented minimum modulated capacity and low-stage latent performance, and remember that the shared ECM blower means even the Classic is not the humidity liability it first appears.

Starting anthropic/claude-opus-4.8

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

## Redo with the three real facts

**Committed inputs:** cooling load **19,000 BTU/h**, heating load **27,000 BTU/h**, cooling EFLH **800**, heating EFLH **1,600**, rate **$0.16/kWh**, hold **5 years**. (I keep 800/1,600 — nothing about the measured loads justifies moving the run-hour split.)

**Delivered energy (identical for both units):**
- Cooling: 19,000 × 800 = **15.2 MMBtu/yr**
- Heating: 27,000 × 1,600 = **43.2 MMBtu/yr**
- Total: **58.4 MMBtu/yr**

**Cooling kWh**
- Classic (SEER2 15.2): 15,200,000 ÷ 15.2 ÷ 1,000 = **1,000 kWh**
- Prestige (SEER2 20.0): 15,200,000 ÷ 20.0 ÷ 1,000 = **760 kWh** → Δ **240 kWh**

**Heating kWh**
- Classic (HSPF2 7.5): 43,200,000 ÷ 7.5 ÷ 1,000 = **5,760 kWh**
- Prestige (HSPF2 8.5): 43,200,000 ÷ 8.5 ÷ 1,000 = **5,082 kWh** → Δ **678 kWh**

**Total savings = 918 kWh/yr → × $0.16 = $146.8/yr.** (Heating is still 74% of it.)

## (a) The thresholds — three numbers, three different answers

**1. Hold-horizon break-even at $0.16:**
$$1{,}250 \div 146.8 = \textbf{8.5 years}$$
This is the flip point. Keep the house **8.5 years or longer** and energy alone repays the premium. At the stated **5-year hold, you fall short.**

**2. Delivered-energy floor for 15-year recovery:**
Need $1,250 ÷ 15 = $83.33/yr → 520 kWh/yr saved. At the near-constant **~15.7 kWh saved per MMBtu delivered**, that's:
$$520 \div 15.7 \approx \textbf{33 MMBtu/yr}$$
This house delivers **58.4 MMBtu/yr — nearly double the floor** — so over a full 15-year life the Prestige recovers easily. The load drop did **not** break the long-run case. It broke the *short-hold* case.

**3. Energy shortfall over the actual 5-year hold:**
- Energy recovered: 5 × $146.8 = **$734**
- Premium: $1,250
- **Shortfall: ~$516**

So over 5 years, **energy pays back only about 59% of the premium.** The remaining **~$516 is not an energy return** — it can only come from (i) a resale premium for a modern variable-speed system, or (ii) five summers of better comfort. Be honest about (i): buyers rarely pay a documented $500+ more for SEER2 20 vs 15.2, so I would not bank on resale closing the gap. That leaves comfort as the real thing the $516 buys.

## The insight the pure payback hides

**The two new facts move in opposite directions, and that's the whole story.**

- The lower load and lower rate **weakened the money case** (payback 7.7 → 8.5 yr; you now sell before break-even).
- But the same 19,000 BTU/h load **destroyed the Classic's comfort case.** The Classic is fixed at 28,400 BTU/h. Against a 19,000 cooling load that's **28,400 ÷ 19,000 = 49% oversized** — a single-stage compressor cycling on and off on a load it satisfies in short bursts. In humid Maryland that is the textbook recipe for a cool-but-clammy house and short-cycling.
- The Prestige modulates down to ~14,300 BTU/h, so a 19,000 load sits **squarely in its steady-run band.**

So the load drop didn't just shrink the payoff — it turned the *cheap* option into the one with the comfort problem. The decision is no longer "is the efficiency worth it?" It's **"is $516 worth avoiding a 49%-oversized single-stage for five humid summers?"** (The shared ECM blower's low-airflow dehumidification mode softens this for the Classic, but it can't stop a compressor from short-cycling on a load half its size.)

**What would actually be right — and neither quote offers it:** a **2-ton** unit. Both quotes were sized for the old house. If a third quote is obtainable, a right-sized 2-ton (ideally 2-stage or inverter) beats both on fit and likely on price.

## (b) The recommendation — plain version

You're selling in about five years, so the fancy unit's energy savings only pay back roughly **$734 of its $1,250 extra cost** in that time — you'd be handing the other **~$500** of benefit to whoever buys your house. On money alone, the cheaper Classic wins. **But** your newly-shrunk house needs only about 1.5 tons of cooling, and the Classic is a 2.5-ton unit that can't turn down — in Maryland's humid summers it will likely leave the house cool but sticky and clicking on and off. So: **if you can get one more quote, ask for a right-sized 2-ton unit and buy that.** If you're stuck with only these two, **buy the cheaper Classic and pocket the ~$500** — unless a comfortable, non-clammy house for your last five summers (and a cleaner home inspection when you sell) is worth ~$500 to you, in which case the Prestige is the safer, more comfortable machine. Either way, do **not** pay the premium expecting to earn it back in energy — over a 5-year hold, you won't.

## Assumptions

- Cooling design load: **19,000 BTU/h**
- Heating design load: **27,000 BTU/h**
- Cooling equivalent full-load hours: **800 hours/year**
- Heating equivalent full-load hours: **1,600 hours/year**
- Electricity: **$0.16/kWh**
- Premium for Prestige: **$1,250**
- Holding period: **5 years**
- Simple payback only: no discounting, maintenance differences, future rate changes, or resale value included.

## Annual energy comparison

### Cooling

Annual delivered cooling:

\[
19{,}000 \times 800 = 15{,}200{,}000\text{ BTU}
\]

**Classic:**

\[
15{,}200{,}000 \div 15.2 \div 1{,}000
= 1{,}000\text{ kWh/year}
\]

Cost:

\[
1{,}000 \times \$0.16 = \$160.00
\]

**Prestige:**

\[
15{,}200{,}000 \div 20.0 \div 1{,}000
= 760\text{ kWh/year}
\]

Cost:

\[
760 \times \$0.16 = \$121.60
\]

**Cooling savings with Prestige:**

\[
1{,}000 - 760 = 240\text{ kWh/year}
\]

\[
240 \times \$0.16 = \boxed{\$38.40/year}
\]

### Heating

Annual delivered heating:

\[
27{,}000 \times 1{,}600 = 43{,}200{,}000\text{ BTU}
\]

**Classic:**

\[
43{,}200{,}000 \div 7.5 \div 1{,}000
= 5{,}760\text{ kWh/year}
\]

Cost:

\[
5{,}760 \times \$0.16 = \$921.60
\]

**Prestige:**

\[
43{,}200{,}000 \div 8.5 \div 1{,}000
= 5{,}082.35\text{ kWh/year}
\]

Cost:

\[
5{,}082.35 \times \$0.16
= \$813.18
\]

**Heating savings with Prestige:**

\[
5{,}760 - 5{,}082.35
= 677.65\text{ kWh/year}
\]

\[
677.65 \times \$0.16
= \boxed{\$108.42/year}
\]

### Total

| | Classic | Prestige |
|---|---:|---:|
| Cooling electricity | 1,000 kWh | 760 kWh |
| Heating electricity | 5,760 kWh | 5,082 kWh |
| **Total electricity** | **6,760 kWh** | **5,842 kWh** |
| Annual cost at $0.16/kWh | **$1,081.60** | **$934.78** |

Annual Prestige savings:

\[
6{,}760 - 5{,}842.35
= 917.65\text{ kWh}
\]

\[
917.65 \times \$0.16
= \boxed{\$146.82/year}
\]

## (a) Exact thresholds

### Hold-horizon threshold

\[
\$1{,}250 \div \$146.82
= \boxed{8.51\text{ years}}
\]

Therefore:

- Holding the house **less than 8.51 years**: the Prestige does not recover its premium through energy savings.
- Holding it **exactly 8.51 years**: energy savings approximately break even with the premium.
- Holding it **more than 8.51 years**: the Prestige begins producing net energy savings.

At the planned five-year hold:

\[
\$146.82 \times 5
= \$734.12
\]

Energy savings recover only **$734.12** of the **$1,250** premium.

Unrecovered amount at sale:

\[
\$1{,}250 - \$734.12
= \boxed{\$515.88}
\]

Thus, the Prestige would need to produce at least **$515.88 of additional resale value or comfort value** to break even by the expected sale date. That value is not guaranteed by the efficiency ratings.

### Fifteen-year energy floor

The total annual delivered energy under these assumptions is:

\[
15.2\text{ MMBtu cooling}
+
43.2\text{ MMBtu heating}
=
58.4\text{ MMBtu/year}
\]

To recover $1,250 over 15 years, the Prestige must save:

\[
\$1{,}250 \div 15
= \$83.33/year
\]

At $0.16/kWh, that requires:

\[
\$83.33 \div \$0.16
= 520.83\text{ kWh/year}
\]

For the measured cooling/heating mix, the Prestige saves approximately **15.71 kWh per delivered MMBtu**. Therefore:

\[
520.83 \div 15.71
= \boxed{33.15\text{ MMBtu/year}}
\]

So, assuming the same approximate cooling/heating mix:

- Above **about 33.15 MMBtu/year** of delivered heating plus cooling energy, the Prestige recovers its premium within 15 years.
- Below **about 33.15 MMBtu/year**, it never recovers the premium through energy savings alone within 15 years.
- This house is modeled at **58.4 MMBtu/year**, so it clears the 15-year floor comfortably—but not the five-year hold-horizon threshold.

Over 15 years at the modeled usage:

\[
\$146.82 \times 15
= \$2{,}202.35
\]

Gross energy savings would be about **$2,202**, or approximately **$952 net** after subtracting the $1,250 premium. That long-term result is irrelevant to this homeowner unless the house is kept for roughly 8.5 years or resale value makes up the five-year shortfall.

## Sizing consequence

The new Manual J makes the sizing concern more serious:

- Classic cooling capacity:
\[
28{,}400 \div 19{,}000 = 1.49
\]
About **49% above** the design cooling load.

- Prestige maximum cooling capacity:
\[
31{,}800 \div 19{,}000 = 1.67
\]
About **67% above** the design cooling load.

The Prestige may still provide better comfort if its actual minimum cooling output is at or below the 19,000-BTU/h load, because it can reduce output instead of repeatedly turning fully on and off. But the contractor must document that minimum capacity and its low-speed humidity performance.

The Classic is described as single-stage in the quote. If that description is accurate, its 28,400-BTU/h cooling output is substantially oversized for a 19,000-BTU/h house load and could produce shorter cycles and weaker humidity control.

## (b) Actionable recommendation

**Do not pay the extra $1,250 for the Prestige if you expect to sell in five years: its energy savings recover only about $734, leaving a $516 shortfall. Ask for a smaller system first, because the Manual J calls for only 19,000 BTU/h of cooling; if these are truly the only two choices, choose the Classic for the five-year financial decision, while recognizing that neither quoted cooling size is an ideal match for the measured load.**

**Updated delivered loads with measured Manual J**
Cooling: 19,000 BTU/h × 800 h = 15.2 million BTU/yr
Heating: 27,000 BTU/h × 1,600 h = 43.2 million BTU/yr
Total annual delivered energy = 58.4 million BTU/yr (both units deliver exactly this amount).

**Annual electricity use and savings at $0.16/kWh**
Cooling: Classic 1,000 kWh ($160), Prestige 760 kWh ($122) → $38 savings
Heating: Classic 5,760 kWh ($922), Prestige 5,082 kWh ($813) → $109 savings
Total annual savings = 918 kWh or **$146.88**.

**Payback and 5-year hold**
Simple payback on the $1,250 premium = 1,250 ÷ 146.88 ≈ **8.5 years**.
Over 5 years the energy savings total ≈ **$734**.
Energy shortfall after 5 years = $1,250 − $734 = **$516** (this portion is never recovered on utility bills alone).

**(a) Exact flip thresholds at $0.16/kWh**
- Hold-horizon threshold: **8.5 years** (must keep the house at least this long for the premium to pay for itself purely on energy).
- Total-annual-delivered-energy floor: the premium never recovers inside 15 years if annual delivered energy falls below ≈ **33 million BTU/yr** (roughly 43 % of the measured post-insulation load).
Because only $734 of the $1,250 is recovered on energy in 5 years, **$516 of the premium can only be justified by comfort, humidity control, or added resale value**.

**(b) What to do**
Buy the Classic. At your actual $0.16 rate and 5-year ownership horizon the variable-speed unit saves only about $730 on electricity, leaving more than half the extra cost unrecovered on bills alone; the single-stage unit will still meet the measured load and keep monthly costs acceptable.

Starting anthropic/claude-opus-4.8

Starting openai/gpt-5.6-luna

Starting x-ai/grok-4.3

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

x-ai/grok-4.3 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

Initial responses are now shared. Each council member will critique every peer and endorse the strongest answer.

- x-ai/grok-4.3: STRENGTH: Clean, correct core arithmetic ($146.88/yr, 8.5-yr payback, $516 shortfall) and a decisive plain answer. WEAKNESS: Recommends Classic flatly while ignoring the central prompt concern — a 28,400 BTU/h single-stage on a 19,000 load is 49% oversized, the textbook short-cycle/humidity liability in humid 4A; also never mentions the right-sized 2-ton path. FACTUAL-CONCERN: yes — states the 33 MMBtu floor is "roughly 43% of the measured load," but 33.15/58.4 ≈ 57%.

- openai/gpt-5.6-luna: STRENGTH: Rigorous and complete — all three thresholds correct, and it explicitly quantifies both the 49% (Classic) and 67% (Prestige) oversizing against the new Manual J, then recommends seeking a smaller system first. WEAKNESS: The (b) paragraph is one dense, semi-hedged sentence ("neither quoted cooling size is an ideal match") that a no-background reader would struggle to act on cleanly. FACTUAL-CONCERN: no.

- anthropic/claude-opus-4.8: STRENGTH: Captures the load-bearing insight — the two new facts push in opposite directions (money weakens, Classic's comfort case collapses) — reframes the decision as "$516 for five comfortable summers," is honest that resale won't close the gap, and gives a genuinely plain, actionable (b) including the 2-ton third quote. WEAKNESS: Still doesn't quantify the backup-strip effect it gestured at last turn, and the "$500 comfort" framing slightly understates that neither unit is correctly sized. FACTUAL-CONCERN: no.

I endorse anthropic/claude-opus-4.8: it alone converts the arithmetic into the correct decision structure (the load drop broke the *short-hold money* case while simultaneously breaking the *Classic's comfort* case) and delivers the only truly jargon-free, act-on-it recommendation with the right-sized alternative both other answers underweight. What the whole council missed this turn: with a 27,000 BTU/h heating design load and the Classic rated only 27,800 at 47°F, the Classic has essentially no cold-weather margin — at Maryland design temperature its capacity fades and it leans on resistance heat at COP≈1.0 far more than the Prestige, so the real heating-cost gap is wider than HSPF2 implies, which could pull the effective payback back under the 5-year hold and undercut the unanimous "money says Classic" drift.

[anthropic/claude-opus-4.8]: STRENGTH: It gets the arithmetic and thresholds right, clearly separates five-year energy economics from comfort and resale value, and correctly identifies that neither quoted cooling size is ideal for a 19,000-Btu/h load. WEAKNESS: Its claim of approximately 14,300 Btu/h minimum Prestige output and its shared-blower humidity discussion require exact matched-system documentation not supplied here. FACTUAL-CONCERN: yes — those performance claims are plausible but not established by the quoted facts.

[openai/gpt-5.6-luna]: STRENGTH: It provides the cleanest transparent calculations, including the exact 8.51-year payback, $515.88 five-year shortfall, and 33.15-MMBtu 15-year floor. WEAKNESS: Its final “choose Classic” recommendation is financially clear but underweights the important comfort consequence of putting a fixed 28,400-Btu/h cooling unit against a 19,000-Btu/h load, and it does not foreground that a third, smaller quote is preferable. FACTUAL-CONCERN: yes — saying the Classic will meet the measured load is not enough to establish heating adequacy at Maryland design temperature from a 47°F capacity rating.

[x-ai/grok-4.3]: STRENGTH: Its arithmetic is correct and its five-year recommendation is direct and understandable. WEAKNESS: It largely dismisses the sizing and humidity question, offering no serious analysis of the Prestige’s minimum output or the Classic’s substantial oversizing. FACTUAL-CONCERN: yes — “keep monthly costs acceptable” is unsupported, and the response does not establish heating performance at design outdoor temperature.

The Claude answer is strongest because it preserves the correct financial recommendation while recognizing that the real equipment choice is inferior to obtaining a properly sized third quote. The entire council omitted a design-temperature heating-capacity and auxiliary-heat check, which is essential before approving either system.

[x-ai/grok-4.3]: STRENGTH: Crisp arithmetic, explicit 8.5-yr and 33 MMBtu thresholds, clear 5-yr $516 shortfall. | WEAKNESS: Under-emphasizes how the 19k load turns the Classic into a genuine humidity liability. | FACTUAL-CONCERN: no

[openai/gpt-5.6-luna]: STRENGTH: Detailed tables, scaling formula, and 15-yr floor calculation. | WEAKNESS: Recommendation hedges without picking a side; sizing discussion stays generic. | FACTUAL-CONCERN: no

[anthropic/claude-opus-4.8]: STRENGTH: Best integration of measured load with real-world comfort risk, exact shortfall, and practical "get a third quote" advice. | WEAKNESS: Minor rounding on savings. | FACTUAL-CONCERN: no

Endorsed Claude because it alone makes the load drop change the decision from pure payback to "is $516 worth avoiding a clammy house," while still stating the exact thresholds. Council missed confirming the Classic's documented two-stage operation before declaring it the short-cycling risk.

Starting anthropic/claude-opus-4.8

Starting openai/gpt-5.6-luna

Starting x-ai/grok-4.3

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

x-ai/grok-4.3 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 serves as Chairperson and will now synthesize the council's reasoning into a single canonical answer.

# Synthesis: the three real facts, run through

**Committed inputs:** cooling load 19,000 BTU/h, heating load 27,000 BTU/h, cooling EFLH 800, heating EFLH 1,600, rate $0.16/kWh, hold ~5 years. (800/1,600 stays — the *measured* loads dropped, but nothing about them changes the hours split for an interior, air-sealed 4A townhouse.)

**Delivered energy — identical for both units:**
- Cooling: 19,000 × 800 = **15.2 MMBtu/yr**
- Heating: 27,000 × 1,600 = **43.2 MMBtu/yr**
- Total: **58.4 MMBtu/yr**

**Cooling kWh**
- Classic (SEER2 15.2): 15,200,000 ÷ 15.2 ÷ 1,000 = **1,000 kWh**
- Prestige (SEER2 20.0): 15,200,000 ÷ 20.0 ÷ 1,000 = **760 kWh** → Δ **240 kWh**

**Heating kWh**
- Classic (HSPF2 7.5): 43,200,000 ÷ 7.5 ÷ 1,000 = **5,760 kWh**
- Prestige (HSPF2 8.5): 43,200,000 ÷ 8.5 ÷ 1,000 = **5,082 kWh** → Δ **678 kWh**

**Total savings = 918 kWh/yr × $0.16 = $146.9/yr** (heating is ~74% of it).

---

## (a) The three thresholds — three different verdicts

**1. Hold-horizon break-even (the flip point):**
$$1{,}250 \div 146.9 = \textbf{8.5 years}$$
Keep the house **8.5 years or longer** → energy alone repays the premium. At the stated **5-year hold, you fall short.** This single number *is* the flip: the recommendation switches from Prestige to Classic the moment the planned hold drops below 8.5 years, which it now has.

**2. Delivered-energy floor for 15-year recovery:**
Need $1,250 ÷ 15 = $83.33/yr → 520.8 kWh/yr. At the near-constant **15.7 kWh saved per delivered MMBtu**, that is:
$$520.8 \div 15.7 \approx \textbf{33 MMBtu/yr}$$
This house delivers **58.4 MMBtu/yr — about 57% above the floor** (not 43% below it, as one council draft had it). So over a *full 15-year life* the Prestige recovers comfortably. The load drop did **not** break the long-run case. It broke the **short-hold** case.

**3. Energy shortfall over the actual 5-year hold:**
- Recovered: 5 × $146.9 = **$734**
- Premium: $1,250
- **Shortfall: ~$516** — about 41% of the premium never comes back on the bill.

**Where the $516 could come from — and the honest quantification the council kept gesturing at but never finished:**

- **Resale value:** Do not bank on it. Home buyers rarely pay a documented $500+ more for SEER2 20 over 15.2. Assume ~$0 here.
- **Winter backup-heat avoidance (my flagged gap, now computed):** The Classic is rated only **27,800 BTU/h at 47°F against a 27,000 BTU/h heating design load** — essentially *zero* cold-weather margin. At Baltimore's ~15°F design temperature a single-stage compressor fades to roughly 17,000–18,000 BTU/h, so it leans on resistance strips (COP ≈ 1.0) during cold snaps far more than the inverter, which holds capacity better. But when you actually bound it — roughly 150–250 cold hours shifting ~5,000 BTU/h from a COP-2 heat pump to COP-1 strips — the extra cost is only about **$20–50/yr**. Credit the midpoint: ~$35/yr × 5 = **~$175** over the hold.

So, being fully honest: even generously crediting the backup-strip effect, the Prestige recovers **~$734 + ~$175 ≈ $909** over five years — still **~$341 short** of its $1,250 premium. **The winter-capacity argument narrows the gap; it does not close it.** (I raised this as a possible flip last turn; the arithmetic says it isn't one.) The remaining shortfall is a **comfort purchase, not an investment.**

---

## The pivot the whole council underweighted: *is the Classic actually single-stage?*

The entire comfort case against the Classic rests on it being fixed-output. At the measured load, 28,400 ÷ 19,000 = **49% oversized** — the textbook short-cycle/clammy-summer setup in humid 4A. **But the quote's "single-stage" label may be wrong:** Rheem's RP14AY line is described in some catalogs as **two-stage**. If this unit is two-stage, its low stage (~18,000–18,500 BTU/h) sits right on the 19,000 load, it runs long low-stage cycles, and — paired with the shared variable-speed ECM blower's low-airflow dehumidification mode — **its humidity penalty largely evaporates.** Verify the staging before signing. It is the single fact that decides whether the Classic has a comfort problem at all.

**And the fact neither quote confronts:** both units were sized for the *dead* house. The right machine for a 19,000 BTU/h load is a **2-ton** (ideally 2-stage or inverter) — smaller, better-matched, and likely *cheaper* than either quote. A third quote is the highest-value move available.

---

## (b) What to do — plain version

You're selling in about five years, so the fancy Prestige's energy savings only pay back roughly **$734 of its $1,250 extra cost** in that time — even counting its better cold-weather performance, you'd still come up about **$340 short** and hand the rest of the benefit to whoever buys your house. On money, the cheaper Classic wins. **Do two things before you sign.** First, get one more quote for a *right-sized 2-ton unit* — your newly-tightened house only needs about 1.5 tons of cooling, and both current quotes are oversized leftovers from the old house; a properly sized unit will likely cost less and run better. Second, ask the installer one question: *"Is the Classic single-stage or two-stage?"* If it's **two-stage**, buy the Classic without worry and pocket the ~$500 — it'll handle your smaller house fine and keep humidity in check. If it's truly **single-stage**, it may leave the house cool-but-sticky and cycle on and off through five humid summers; then the Prestige's extra ~$500 buys you comfort you'll actually feel, even though it won't pay for itself on the bill. Either way, **don't pay the premium expecting to earn it back in energy over five years — you won't.**

Starting anthropic/claude-opus-4.8

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...

anthropic/claude-opus-4.8 is thinking...