Blog

How to value a point, and why published valuations disagree

Why two honest publishers can price the same points differently, what each choice assumes, and a method for setting your own number from how you really redeem.

Published point valuations disagree because they answer different questions. One asks what a skilled redeemer could get on a good day. Another asks what the typical holder gets. A third asks what the points are worth if you only ever use the cash option. All three are legitimate, none is a price, and the number that matters for your balance is a fourth one that you can build yourself from the way you actually redeem.

This post explains where a published figure comes from, which choices move it, and how to turn those choices into a personal cents per point number you can use to decide whether an offer, a card or a redemption is worth your time. It assumes you already know the basic formula (dollars received divided by points spent, times 100); cents per point: why the same 10,000 points can be worth very different amounts works through it on one balance if you want the arithmetic first.

What a published number is trying to be

Several travel and personal-finance publishers post a monthly or seasonal figure for the large bank programs. As of mid to late 2026, The Points Guy's valuations put American Express Membership Rewards at about 2.0 cents per point, Chase Ultimate Rewards at about 2.05 cents and, in its October 2026 listing, Bilt at about 2.2 cents. Another publisher, ShopBack, gave ranges in mid-2026 of 1.8 to 2.1 cents for Chase and 1.8 to 2.2 for American Express, and one further guide showed American Express at 1.4 to 2.0. These are secondary, opinion-based estimates, and none of them is a guarantee of what a redemption will return.

The first thing to notice is that the figures are close to each other and far from one cent. The second is that they are described as estimates by the people who publish them. A cash redemption in most bank programs is typically worth one cent per point, so the gap between one cent and roughly two is the claimed premium for using the program's travel features. The disagreement among the published numbers is small compared with that premium, and the premium itself depends on choices you may not share with the publisher.

The programs these figures describe are Chase Ultimate Rewards and American Express Membership Rewards. Their profiles carry the dated, sourced facts about how each works; this post is about the arithmetic that sits on top.

Where a number comes from

Every valuation has the same skeleton: a set of redemptions, a dollar value for each, a points cost for each, and a rule for combining them into one figure. The rule is where publishers differ. A small made-up basket shows how far the same data can travel.

Say a publisher looks at four redemptions of a program, with these illustrative rates: cash at 1.0 cents, a travel-portal booking at 1.3 cents, an economy flight on a partner airline at 1.8 cents and a premium-cabin partner flight at 3.4 cents. The figures are invented to show the effect and are not any real program's rates.

Method Calculation Published figure
Best realistic redemption take the highest 3.4 cents
Plain average of the basket (1.0 + 1.3 + 1.8 + 3.4) ÷ 4 = 7.5 ÷ 4 1.875 cents
Typical good redemption average of the middle two: (1.3 + 1.8) ÷ 2 1.55 cents
Weighted by what holders really do 0.40 × 1.0 + 0.30 × 1.3 + 0.20 × 1.8 + 0.10 × 3.4 1.49 cents

Four defensible methods, one basket, and a spread from 1.49 to 3.4 cents. Nobody in the table has made an error. The first is a ceiling, useful for showing what the program can do at its most generous. The last describes a hypothetical typical holder who spends 40 percent of points on cash, 30 percent on portal bookings, 20 percent on economy partner flights and 10 percent on premium cabins. If you are that holder, 1.49 is your number. If you only fly premium cabins, it is far too low. If you only take cash, it is too high.

Real publishers sit between these poles, and they rarely publish the weights. That is the main reason two reputable lists differ by a tenth or two of a cent, and why a list that hovers near two cents says something about the redemptions its authors consider normal rather than about what any specific balance will return.

The five levers that move a valuation

Behind the method there are five assumptions, and each can be checked in a sentence. When you see a figure, ask which way each one leans.

1. The benchmark price. The top of the fraction is the cash price you would otherwise have paid. A publisher who benchmarks against the average fare for a route gets a higher number than one who benchmarks against the cheapest comparable fare. A last-minute or peak-date fare is higher still, and a redemption that looks spectacular against it may look ordinary against the fare you would really have booked three months out.

2. The cabin. Premium cabins have large cash prices and award prices that, in many programs, do not rise in proportion. A valuation that includes them has a fatter right tail. If you would never pay cash for the premium cabin, the cash price is not a saving to you, and counting it overstates your return.

3. Availability. An award price is only available if an award seat exists. A valuation that quotes the price of a seat without asking how often the seat can be found tells you what is possible, not what is likely. Value times probability is lower than value.

4. Access to partners. The highest figures usually come through a transfer partner arrangement, in which points move to an airline or hotel program and are spent there. That path is closed to a holder whose card does not allow transfers; why the card you hold decides how flexible your points are explains why. For that holder the realistic ceiling is the portal or cash, and a published valuation built on partner redemptions is simply not describing them.

5. The date. Program rules, award prices and partners change. A valuation is a snapshot of a program at one time, and devaluation is the name for the moment the snapshot stops being true. The cost of waiting to redeem is not in the number.

Pricing design changes what is being measured

Programs price awards in two broad ways, and the choice shapes what a valuation can say. In a chart-based design the points price of a given trip is set in a table, so the cents per point swings with the cash fare: it is high when the cash fare is high and low when the fare is cheap. In a dynamic award pricing design the points price follows the cash price, so the rate stays in a narrower band, and the valuation is steadier but the upside from a bargain fare disappears. award charts versus dynamic pricing for flights, and why it matters sets the two side by side.

For valuation this means a single figure for a chart-based program is partly a statement about which trips a publisher picked. A figure for a dynamic program is partly a statement about the typical ratio between points and fares. Neither can tell you what a particular Saturday in March will cost.

Building your own number: a worked example

The method is the weighted approach from the table, applied to your plans. Use only redemptions you would really make.

Say Priya holds 60,000 points and is not a premium-cabin flier. She expects to do three things, with all figures illustrative.

  1. A flight she takes every year. A route she would buy with cash at $380. The award costs 25,000 points plus $12 in taxes and fees. Value = $380 − $12 = $368. Rate = $368 ÷ 25,000 × 100 = 1.47 cents.
  2. A second, similar flight. The same arithmetic: 25,000 points, $368 of value.
  3. The leftover as cash. 10,000 points at one cent is $100.

Total value is $368 + $368 + $100 = $836 for 60,000 points. Her personal number is $836 ÷ 60,000 × 100 = 1.39 cents.

Now compare that with a published 2.0. At 2.0 cents her balance would be worth 60,000 × $0.02 = $1,200. Her own figure is $836, which is $364 lower, or about 30 percent below the headline. The difference does not mean the published number is wrong. It means Priya's redemptions are economy trips with a moderate cash price and a real taxes-and-fees charge, and the published figure assumes something richer.

Two additions make the personal number more honest. First, add the risk that a redemption does not happen: if she thinks there is a one in five chance a flight award cannot be found at the price she used, the expected value of that line is 0.8 × $368 = $294.40, and the total falls accordingly. Second, treat the cash option as a floor rather than a plan. If every other option fails, one cent is what she gets, and the personal number cannot fall below it as long as that option exists.

The points and miles value calculator runs the same arithmetic with your inputs, and what a point is worth to you, with a worksheet to find out extends it across a full year of planned redemptions.

A second holder: same program, different ceiling

The method gives a different answer when the options differ. Say Marcus holds 40,000 points in the same kind of bank program, but his card does not allow transfers to airline or hotel partners. His realistic menu is cash at one cent and a travel portal that prices bookings at a fixed illustrative rate of 1.25 cents per point.

Spending all 40,000 points in the portal buys 40,000 × $0.0125 = $500 of travel. That looks like 1.25 cents. Before he accepts it, he checks the portal against the open market for the same flight and finds the portal fare is 6 percent higher than the airline's own price. The same trip direct would cost $500 ÷ 1.06 = $471.70. His real value is the direct price he avoided paying, so the rate is $471.70 ÷ 40,000 × 100 = 1.18 cents. The portal premium took about seven hundredths of a cent off each point. Whether a portal price and a direct price really differ is something to check fare by fare, not to assume.

Compare that with a published 2.0. At 2.0 cents his balance would be worth $800. His own figure is $471.70, a gap of $328.30. Marcus is not doing anything wrong; the published number describes a path his card does not open. For him the useful comparison is between cash at one cent and the portal at 1.18, and the question becomes whether a trip he would otherwise buy is worth the 0.18-cent premium over cash. If the answer is yes, the personal number is 1.18. If he would not take the trip, his number is one cent.

The point of putting Priya and Marcus side by side is that one program produced two honest values, 1.39 and 1.18 cents, and a published 2.0 fits neither of them. That is not a flaw in the published figure. It is the reason for building your own.

How to read a published figure as an upper bracket

Treat a published valuation as the upper bracket of a range, with the cash rate as the lower bracket and your own weighted figure inside it. In practice that gives three questions to ask before you act on any number.

  • Is the redemption in the published basket? If the number comes from premium-cabin partner flights and you book economy, subtract your expectations.
  • Would I have paid cash for this? If not, the fare you are measuring against is not a saving.
  • Can I get the seat? If you cannot check availability before you transfer, the number is a hope.

Using a published valuation to decide whether to look harder is fine. Using it to claim that a balance is worth a specific amount, or to justify spending to earn points, is not.

Why valuations appear in decisions they should not

Because the figures are public, they show up in arguments about annual fees, welcome offers and card choices. Used correctly they help; used carelessly they do damage. Three habits cause most of the harm.

Valuing the earn side at the redeem rate. A card that earns 3 points per dollar at a published 2 cents is described as returning 6 percent. That is only true if you redeem at the published rate every time. At your personal 1.39 cents it returns 3 × 1.39 = 4.17 percent, and at one cent for cash it returns 3 percent. All three numbers are arithmetic, and only one of them is yours.

Netting an annual fee against rewards you will not redeem well. A fee is a cash cost today. A reward valued at 2 cents is an expectation about the future. The comparison is fair only when the reward figure is one you would be willing to bet on.

Hoarding for a number. Waiting until a balance can be exchanged at a high-value sweet spot has a cost: the program can change the sweet spot, and points can lose usefulness for reasons that have nothing to do with you. should you spend points now or save them for a bigger trip? weighs that trade.

When your number should change

A personal valuation is not permanent. Revisit it when any of these happens.

  • You change the way you travel: a first family trip, a move, a new home airport.
  • A program announces a change to award prices, partners or rules.
  • You change the card that holds your points, which can change which redemptions you can reach.
  • A year passes and you can compare plan to actual: what did you really redeem, and what was the real cents per point on each?

The last item is the most useful. A short log of redemptions, with points spent, cash value and taxes paid, replaces speculation with a record, and after a year or two it is a better guide than any publisher's table.

Key takeaways

  • Published valuations differ because publishers choose different baskets and combining rules: in the illustrative basket the same four redemptions gave 3.4, 1.875, 1.55 or 1.49 cents depending on the method.
  • As of 2026, The Points Guy and ShopBack put large bank programs at roughly 1.8 to 2.2 cents; these are opinion-based estimates, and cash redemption is typically one cent.
  • The levers behind any figure are the benchmark fare, the cabin, availability, access to transfer partners and the date, and each can raise or lower your result.
  • A personal number weights only redemptions you would make: in the example, $836 of value on 60,000 points is 1.39 cents, about 30 percent below a published 2.0.
  • Use the published figure as an upper bracket and one cent as the floor, and revisit your own number when your travel, your card or the program changes.

Terms used in this article

Related rewards programs

Related tools

Latest on this topic