How to read these numbers

Version 1 · this build is for evaluation only. Nothing here is bindable and nothing is submitted to a carrier.

What the tool actually does

You enter a risk once. Every carrier is priced by applying that carrier's own filed rates — the rate manual and class plan approved by the California Department of Insurance — in the order that filing prints them. Each number traces back to a page of a filing, not to a guess or an average.

Where a carrier cannot price a risk, the card says which table is missing. A refusal is information. A silent estimate would be worse than no answer.

The four things you can compare

Number	What it is	What to trust it for
MAT price	A computation from the carrier's filed rates	Reproducing the filing, to the cent, where we have proven it
Carrier portal price	The carrier's own system, quoted live	The answer key. If you have this, it beats everything else
ITC price	What TurboRater shows	A reference point — not a reference standard. See below
Refusal	A named reason a carrier cannot price	Knowing the carrier would not write the risk, not that we failed

Why ITC is a reference point, not a standard

ITC mostly passes the risk through to the carrier's own engine, so where it does that, its number is the carrier's number. But it also does things we can measure and disagree with:

It rewrites the annual mileage on submission (we have seen 10,000 become 11,500) without telling the agent.
It applies its own ZIP-to-territory table, which measurably differs from the table printed in the carrier's filing.
On one programme it prices from a level the filing does not print.

So when MAT and ITC differ, the useful question is "which one matches the carrier's own portal?" — and that is what column 3 of the capture sheet is for.

What to expect from each carrier today

Carrier	Standing	What you will see
Coast National	A	Reproduces the carrier's published worked example to the cent ($2,243.48). Expect close agreement.
Triton General – Diam	B	Within 2 cents of a live portal breakdown, and the fee stack decomposes exactly. Expect close agreement.
Aspire General	A	Reproduces its filed example exactly — but declines some risks (see below).
Nations Insurance	C	Will read low against ITC on many risks — typically in the region of 15–20%. This is expected and the cause is measured. Do not read it as a broken number.
Integon National	C	Computed from the filing, not yet cross-checked against a live portal. Unknown territory; report what you see.
Orion Indemnity	C	Same as Integon. Orion's manual prices a 6-month term by default; the annual term is derived from its printed factor.
Bluefire – Nations NIC	Not quoted	Not in this build. Its filed programme sells a single liability limit, so a current-limit quote needs different modelling. It is queued, not hidden.

Aspire declines risks. On at least one benchmark risk, all four Aspire products returned "Rate not provided" in ITC. When that happens the Aspire figure on this page is a faithful computation of Aspire's filed rates — not an offer. Always check whether Aspire actually quoted the risk in the portal before treating its number as reachable.

Keep the risks inside this envelope for the first twenty

1–2 drivers, 1–2 vehicles
Violations entered as points (the form takes a points count)
Comp/collision on or off, medical payments optional
Terms: 1, 3, 6 or 12 months (which carriers offer which is shown in the term menu)
California only. Owner policies. No excluded drivers, no non-owner policies, no 3+ vehicle households — those are outside what this build models, and a surprising number from one of them would be noise, not a finding.

No customer data

This is a public test build. Use realistic but synthetic names, addresses and plate/VIN details. Do not put a real applicant's personal information into it.

How to record a risk

Enter the risk once; note the six carrier prices (or the refusal reason).
Quote the same risk in the carrier's own portal where you have access — that is the number that matters.
Note the ITC figure for the same risk.
Put all four on one row of the capture sheet, with anything odd in the notes.

Twenty rows is enough to tell us where the real gaps are. A risk that disagrees is useful; a risk that is not recorded is wasted.

What we do with it

Every disagreement becomes one of two things: a named cause with a controlled test behind it, or an open item we publish. What we will not do is adjust a number until it matches — a rater tuned to look right is worth less than one that says what it does not know.

← Back to the rater · How each number is verified