Why Points Valuations Disagree, and How to Read the Spread

Three publishers, three different questions

The publishers are not competing to estimate one quantity and missing by different amounts. They are answering different questions, and the answers happen to share a unit. Each states its own method, and it is worth reading them next to each other:

CardRatings — what people actually got

Real-world award bookings, comparing the cash price against the points required and calculating the value received on a cents-per-point basis — explicitly not best-case awards. That is a backward-looking, empirical question: across bookings that really happened, what did a point return? Its weakness is auditability. Sample size and composition are undisclosed, which makes it the least auditable of the three. That omission is itself something we disclose.

The Points Guy — what an informed buyer should pay

A blend of the price at which they would buy the currency, award costs in the program (factoring in availability and fees), and “our expertise in the program’s inner workings,” over extensive data for nine US airline and six hotel programs. Note the third component. It is an explicitly editorial input, and stating it openly is to their credit — but it means the figure is a considered opinion about a program, not a measurement of one. Most current, and explicitly part-editorial. Refreshed monthly.

Upgraded Points — how useful the currency is

A six-factor assessment — point value, flexibility, transfer-partner availability, redemption options, purchase pricing, and the average redemption value from three separate bookings — weighted against award availability. Price is one of six inputs here, so a currency can score well for flexibility and transfer options even when its redemption rates are ordinary. Widest program coverage, and the only one of the three that scores factors beyond price. A three-booking average is a small sample, stated.

Averaging those three into one benchmark — which is what a single-number valuation page effectively does, whether or not it says so — combines a documented sample median, a part-editorial estimate and a six-factor utility score. The result carries a decimal point it has not earned. That is why our valuation reference exposes a range and no mean, and why the tools report three states rather than a pass or a fail.

How to read a spread

A spread is information, not noise. Roughly, it tells you how much of a currency's value depends on how you personally use it.

• A narrow spread means the methods converge. When an empirical sample, an editorial estimate and a utility score all land in the same place, the currency behaves consistently — it is worth about that much however you redeem it. Treat the figure as reasonably firm.
• A wide spread usually means the currency rewards specificity. One publisher is capturing typical bookings and another is capturing what the program can do in the right hands. Hilton Honors sits at 0.40¢–0.50¢ — small numbers, but a 25% spread, which on a hundred-thousand-point stay is real money.
• Sometimes the spread is not about value at all. British Airways Avios runs 1.25¢–1.40¢. The Points Guy publishes one figure for Avios as a shared currency; Upgraded Points values each Avios program separately and puts British Airways, Iberia, Aer Lingus and Qatar Privilege Club all at 1.25¢. Part of this spread is therefore a difference in how the currency is defined, not only in what it is worth.

There is also a mundane caveat that no amount of methodology fixes: the figures are not contemporaneous. Our three sources currently carry 2 different as-of dates (1 July 2026; 10 June 2026). Part of any spread is therefore just the calendar — one publisher may have already priced in a devaluation another has not seen yet. Every figure on this site is shown with its date for exactly that reason.

Which publisher should you use? The one that describes you

This is the practical resolution, and it is not a cop-out. Since the three are answering different questions, the right figure to hold yourself to is the one whose question matches your behaviour.

• If you book what is available on the dates you need — main-cabin seats, ordinary hotel nights, no chasing — the bookings-based figure is the honest bar. Anything higher describes somebody else's trips.
• If you plan trips around award space — flexible dates, sweet-spot partner awards, premium cabins you would not buy with cash — the higher figures are reachable, but only because you are doing the work that produces them.
• If you are deciding whether to hold a currency at all, the utility-weighted view is the relevant one, because flexibility and transfer partners are exactly what you are buying.

Then use the range as a range. A redemption above every published figure is good by any of the three standards. Below every one, it is poor by all of them. In between — which is where most real bookings land — the money is genuinely arguable, and the decision comes down to availability, refundability, expiry and whether you wanted the trip.

Informational only, not professional advice. This guide summarises and compares third-party published estimates; it does not endorse any of them, and it recommends no credit card or loyalty program. Figures are reproduced with their publishers' as-of dates and change without notice.

What we left out, and why

A word on our own method, since this guide is asking you to weigh other people's. We require at least two of the tracked publishers to value a program before it gets a row — one figure is a quotation, not a range, and a range is the point. 14 of the 28 programs carry all 3.

We also looked at NerdWallet and excluded it, which is worth stating plainly because on method it is arguably the most reproducible of any we reviewed — a median of a documented sample of routes at set booking windows. Its per-program figures were not available to us as published text, and the program pages that do state a figure in prose carry no as-of date and disagreed with each other. A figure with no date cannot be compared against dated ones, so transcribing it would have quietly degraded every row it touched. That is a limitation of our table, not a criticism of theirs, and it is listed on the reference page itself.

Sources

• The Points Guy — points and miles valuations. Figures as of 1 July 2026. Refreshed monthly.
• CardRatings — points and miles valuations. Figures as of 10 June 2026. Updated periodically; no published schedule.
• Upgraded Points — points and miles valuations. Figures as of 1 July 2026. Refreshed monthly.
• Figures in this guide are reproduced from the three publishers above as published, and are compared in our own points & miles valuation reference. The spread arithmetic, the 20% threshold and the reading of each methodology are ours, not any publisher's.

Last reviewed: July 2026