The discount is a subtraction, not a mystery
A bundle discount is a percentage taken off the total of the two policies you buy together. The page applies it exactly that way: the separate premiums are added and rounded to the cent, the discount is applied to that sum, and the result is rounded again. A carrier quoting 10% on a $2,700 combined premium is selling $270 of savings a year, whatever the wording around it.
Everything about the deal sits inside that one subtraction - the quoted percentage and the two premiums - so three numbers are enough to price it exactly. The page prints the separate total, the bundled total, the dollars saved and the saving as a share of the separate total, which is the figure to hold against the effort of moving both policies.
The discount is clamped to the range 0% to 100%, so a mistyped 150% cannot turn a quote into a refund, and a negative entry cannot add a surcharge. A combined premium of $0 returns a warning instead of a table, because there is nothing to discount and the saving would be zero under every percentage.
The percentage is quoted, the dollars are earned
Carriers quote the discount in percentage points because the percentage is the same for every customer; you collect the saving in dollars, and the dollars depend on how large your premiums are. Ten percent of a $1,200 auto premium is $120 a year, and ten percent of a $2,400 premium is $240, so the same quoted rate is a different deal from one household to the next.
The saving rate printed on the page is recomputed from the rounded dollar figures rather than copied from the input. On a $2,700 combined premium the quoted 10% and the printed rate both come out at 10%; on any realistic pair of premiums the cent-level rounding is invisible in the printed rate, and the two numbers agree to the whole percent the page shows.
A ladder of fixed discounts - 5%, 10%, 15%, 20% and 25% - is run against your two premiums, so you can see what a better or worse quote would be worth before you ask for one. It is the same pair of premiums at each level, not a prediction of which discount your carrier will offer.
The real question is what you give up
The savings do not fall out of the air; they come from concentrating both policies with one company, and that trade-off is the part worth pricing. Your current premiums sit with carriers you already know, with your claims history and any credits already built in, while the bundled quote may quietly rest on different coverages, different deductibles or a different limit structure.
The honest comparison is the bundled quote against the same coverages at the same limits, not against a remembered premium. If a bundle raises a deductible or trims a coverage to reach its price, the arithmetic here still reports a saving, because it only sees the two premiums it was given.
That is also why the page never claims the bundle is cheaper than shopping each policy separately. It prices the quote you were offered; what a standalone specialist would charge for either line is a different comparison, and often the one that finds the bigger number.
What the arithmetic cannot see
The calculator takes three inputs and returns four numbers. It cannot weigh service quality, claims handling, how reachable a local agent is, or how the two policies would be rated if one of them were canceled later. Bundling also decides where renewal happens each year: one company re-rates both policies at once, so a poor claim experience on one line can raise the price of the other.
Not modeled either: multi-line credits beyond the bundle, loyalty discounts, and whether the combined premium in the quote was built on the same base as your current separate premiums. A discount quoted off a raised base can look generous and still cost more than two standalone policies.
The page prices the discount you were quoted and leaves the rest of the decision where it belongs, on your own weighing of the trade-offs - price, coverage, service and the years you expect to stay with both policies.
Danfio