Why Mortgage Calculators Disagree: Three Scoring Methods, One House
Put the same house into three mortgage calculators and you can get three different recommendations. Not because any of them has the arithmetic wrong — amortization is amortization — but because each one is quietly answering a different question, and none of them tells you which.
This piece runs one house through four common decisions and scores each decision three ways. The point is not that one method wins every time. It is that the disagreements fall into distinct patterns, and once you can recognize the pattern you can tell what a calculator is hiding without needing to see its code.
- There are three common ways to score a mortgage decision: monthly payment, cumulative interest, and total cost (cash at close + payments made + balance still owed).
- They disagree in three different ways: sometimes on direction, sometimes on magnitude, and sometimes on what the stakes even are.
- Payment-only scoring reverses the answer on term length. Cumulative-interest scoring gets points right but overstates the savings by more than double.
- The single most misleading number in mortgage marketing is "interest saved." It is real, and it is usually the smallest part of what actually changed.
The house, and the three ways to score it
Every example below uses the same purchase: $450,000, held for seven years, which is close to the median tenure for a first mortgage. Principal and interest only; taxes and insurance are the same across each pair and cancel out.
Monthly payment
Which option costs less per month. The number a lender quotes, the number in the listing, the number most people decide on.
Cumulative interest
How much interest each option accrues over the holding period. The number most "savings" calculators headline.
Total cost
Cash at close + every payment made + balance still owed at the end. Plus, where cash is tied up, what that cash would have earned elsewhere.
Method 3 is the one every comparison on this site uses, and the reasoning is on the methodology page. But the more useful thing here is to watch the other two fail in specific, repeatable ways.
Decision one: 30-year or 15-year
$360,000 loan after 20% down. 30-year at 6.75% versus 15-year at 6.05%, which is a typical spread between the two.
| Method | 30-year | 15-year | Says |
|---|---|---|---|
| Monthly payment | $2,335 | $3,048 | 30-year, by $713/mo |
| Cumulative interest, 7 yrs | $162,971 | $127,480 | 15-year, by $35,491 |
| Total cost, 7 yrs | $612,971 | $577,480 | 15-year, by $35,491 |
Total cost = $90,000 down + payments over 84 months + balance remaining at month 84.
The pattern: disagreement on direction. Payment-only says 30-year, and it is not a close call — $713 a month is a lot of money. The other two methods say 15-year by the same $35,491, and they agree with each other because with identical cash at close, the total-cost difference is the interest difference.
What payment-only is hiding: the 30-year borrower owes $326,835 after seven years. The 15-year borrower owes $231,480. That $95,000 gap in principal repaid is where the lower payment went. It was not saved. It was deferred, with interest.
This does not mean the 15-year is right for everyone. It means the payment comparison told you the wrong thing with total confidence. Choosing the 30-year is a legitimate decision about cash flow; choosing it because the calculator called it "cheaper" is a decision made on a number that does not measure cost.
Decision two: buy two points, or not
Same $360,000 loan. Two discount points cost $7,200 at closing and lower the rate from 6.75% to 6.25%, assuming a quarter-point of rate per point, which varies by day and lender.
| Method | No points | 2 points | Says |
|---|---|---|---|
| Monthly payment | $2,335 | $2,217 | Points, by $118/mo |
| Cumulative interest, 7 yrs | $162,971 | $150,314 | Points, by $12,657 |
| Total cost, 7 yrs | $612,971 | $607,514 | Points, by $5,457 |
The pattern: agreement on direction, disagreement on magnitude. All three say buy the points at a seven-year hold. But look at the size of the win. Cumulative interest says $12,657. Total cost says $5,457. The interest method is overstating the benefit by more than double, because it never subtracted the $7,200 you handed over to get it.
That gap matters at the decision boundary. The naive breakeven — $7,200 divided by $118 a month — is 61 months, so a borrower planning to hold five years would read this as a comfortable win. The total-cost breakeven is later than that, because the lower rate also slows principal paydown slightly. If your realistic hold is four to five years, the method you use decides whether points look like a clear yes or a coin flip.
Run your own numbers on the buydown calculator and set the hold honestly. The answer moves more with that input than with any other.
Decision three: 10% down or 20% down
This is the one where the third method has an extra term. Putting 10% down instead of 20% keeps $45,000 in your pocket. An honest comparison has to count what that $45,000 earns if invested, or it will always favor the bigger down payment.
10% down means a $405,000 loan, and mortgage insurance until the balance reaches 78% of the purchase price — modeled here at 0.5% of the loan per year, which is mid-range for good credit.
| Method | 20% down | 10% down | Says |
|---|---|---|---|
| Monthly payment (incl. PMI) | $2,335 | $2,796 | 20% down, by $461/mo |
| Cumulative interest + PMI, 7 yrs | $162,971 | $197,517 | 20% down, by $34,546 |
| Total cost, 7 yrs, no opportunity cost | $612,971 | $647,517 | 20% down, by $34,546 |
| Total cost, crediting 5% on the $45,000 | $612,971 | $629,198 | 20% down, by $16,227 |
The pattern: every method agrees, and three of them are still wrong about the size. 20% down wins here at any reasonable return assumption; at 6.75% mortgage rates plus PMI, the $45,000 has to earn well above 7% to break even, and seven years is not long enough for that to be a safe bet. But without the opportunity-cost term the margin looks like $34,546. With it, the margin is $16,227 — less than half.
A calculator that omits opportunity cost is not lying about the direction. It is inflating the case by exactly the amount the smaller-down-payment borrower's money would have earned, and it does that every time, on every comparison, in the same direction. Over enough decisions that is not noise; it is a thumb on the scale. The rent-vs-buy calculator includes the term by default for the same reason.
Decision four: pay $300 extra, or invest it
Same $360,000 loan at 6.75%. Option A sends $300 a month extra to principal. Option B invests $300 a month at 7%. Seven years.
| What you're told | Extra payments | Invest instead |
|---|---|---|
| "Interest saved" over 7 yrs | $6,900 | — |
| Principal actually paid down (extra equity) | $32,100 | — |
| Investment balance after 7 yrs | — | $32,589 |
The pattern: the headline number isn't measuring the stakes. "You'll save $6,900 in interest" is true and it is tiny. What actually happened is that $25,200 of extra payments became $32,100 of additional equity — the $6,900 is the interest that didn't accrue on principal you retired early. The real comparison is $32,100 of equity against $32,589 in the brokerage account. That is a wash, and it turns on the 7% assumption: at 6% the paydown wins, at 8% the investment does.
"Interest saved" is the most quoted number in this category and the least informative. It makes extra payments look weak (a few thousand dollars) when what they do is build equity on a par with investing at a solid return — guaranteed, at the mortgage rate, with no market exposure. Which is worth more depends on your reserves, your tax situation, and your appetite for a fixed return; it does not depend on the interest-saved figure at all. The extra-payment calculator reports the equity built for this reason.
Reading the pattern
Four decisions, three failure modes:
| Decision | How the methods disagree | What the weaker method hides |
|---|---|---|
| 30 vs 15 | Direction | Principal not repaid |
| Points | Magnitude, 2× | Cash paid at close |
| Down payment | Magnitude, 2× | What the cash would have earned |
| Extra payments | The metric itself | That equity built is the outcome, not interest avoided |
None of these is a bug. Each method is measuring something real. The trouble is that a calculator presents its metric as the answer, and the metric it chose usually happens to be the one that makes the decision look simpler than it is.
So the diagnostic is short. When a calculator gives you an answer, ask three things: does it count what I paid at closing, does it count what I still owe at the end, and does it count what my cash could have earned? A "no" to any of those tells you which pattern above you are looking at, and roughly how far off the number is.