What points were worth, August 2026
Over eight days we paired 99 award prices with the cash price of the same room or the same seat, bought on the same screen on the same day, across 25 loyalty programs. This is what that showed, including the parts that did not work.
This is a dated study, not the live benchmark. The figures below were measured in August 2026 and have not been updated since. They are a record of what we found, on those dates, with those observations behind them.
For how Wanderpoints benchmarks a redemption today, see how we value points. This page is deliberately frozen: a measurement that gets quietly edited afterwards is not a measurement.
How a measurement was made
An award chart is not a valuation. A chart says a room is 15,000 points. It does not say what 15,000 points is worth. The value is whatever the points displaced in cash, so every observation pairs an award price with the cash price of the same thing, bought on the same screen on the same day.
Same room or same cabin, same dates, looked up in the same sitting. A cash price pulled a week later is a different product, and there is no shortcut here that leaves the honesty claim standing.
Three kinds of number
measured Ten programs
Five or six paired observations each, spread deliberately across cheap and expensive, midweek and weekend, ordinary dates and holiday crushes, domestic and international. We published the median, never the average, so one spectacular find could not carry a program.
Under five observations we published nothing at all and left the program on an estimate. A thinly supported number is worse than an admitted guess.
derived Six transferable currencies
Amex, Chase and the rest are not spent directly. They move into an airline or hotel program, so what they are worth depends on where they can go. We took the best partner we had actually measured, applied the transfer ratio, discounted it, and floored it at whatever the bank will pay you in cash without transferring at all.
The discount was a judgement call and we would rather argue about it in the open than bury it. Taking the best partner at face value assumes the award you want is always there at the price you want, which is exactly the optimism the exercise existed to remove.
estimate Nine programs
Conservative guesses, set in May 2026 and labelled as guesses everywhere they appeared. They were not a copy of anybody else's published table. When a program was measured it moved out of this group and the estimate was discarded.
Where every program stood
As measured on 28 August 2026. Not maintained since.
| Program | Cents | Basis |
|---|---|---|
| Transferable currencies | ||
| Bilt Rewards | 2.00 | derived |
| Chase Ultimate Rewards | 2.00 | derived |
| American Express Membership Rewards | 1.45 | derived |
| Citi ThankYou Points | 1.30 | derived |
| Wells Fargo Rewards | 1.30 | derived |
| Capital One Miles | 0.80 | derived |
| Airlines | ||
| Alaska Mileage Plan | 1.50 | measured 5 obs |
| Southwest Rapid Rewards | 1.50 | measured 6 obs |
| United MileagePlus | 1.50 | measured 6 obs |
| JetBlue TrueBlue | 1.40 | measured 5 obs |
| American Airlines AAdvantage | 1.20 | measured 6 obs |
| Delta SkyMiles | 1.15 | measured 6 obs |
| Air Canada Aeroplan | 1.50 | estimate |
| Avianca LifeMiles | 1.40 | estimate |
| Turkish Airlines Miles&Smiles | 1.40 | estimate |
| Virgin Atlantic Flying Club | 1.40 | estimate |
| Flying Blue | 1.30 | estimate |
| The British Airways Club | 1.30 | estimate |
| Emirates Skywards | 1.20 | estimate |
| Hotels | ||
| World of Hyatt | 2.20 | measured 6 obs |
| Marriott Bonvoy | 1.05 | measured 6 obs |
| Hilton Honors | 0.80 | measured 6 obs |
| IHG One Rewards | 0.70 | measured 5 obs |
| Wyndham Rewards | 0.90 | estimate |
| Choice Privileges | 0.60 | estimate |
The rules that made this slow
Most of these were learned by getting them wrong first. Each one is a way of accidentally measuring something other than the currency, while still producing a number that looks fine.
- The entry level room, whatever the brand calls it. A premium room award is a different product at a worse rate. At one Conrad the bay view cost 4.3 times the points and only 1.66 times the cash, which reads as 0.30 cents against the standard room's 0.76.
- Cash means all in, taxes included. A hotel award charges you no tax, so the tax is part of what the points saved. Leaving it out understates every hotel program by fifteen to twenty percent.
- The standard flexible member rate. Not prepaid, not a promotion, not a rate that needs status or a co-brand card in your wallet. The test is whether any member could book it.
- Main cabin, never basic economy. The cheapest fare an airline sells is a worse product than the award, so pricing against it makes every program look better than it is.
- Same trip shape on both sides, and seven nights. International fares still carry minimum stay rules. The same London flight and the same 60,000 mile award priced $6,167 as a one night turnaround and $1,265 with a week's stay.
- The booking page, not the search results. One search screen claimed to include taxes and understated the room by 16.7 percent against its own checkout.
What the measurements showed
These four are the reason the exercise was worth doing. They are claims about how award pricing behaved in August 2026, on the observations described above, and they are the part of this page most likely to still be true a year from now.
On a fixed date, a premium award usually returns less per point
Business or first class returned less per point than economy on five of the six programs where both were captured, from five percent worse on Southwest to fifty two percent worse on American. The cleanest case held everything else fixed: identical flights, identical dates, only the cabin changed, and first class still lost eleven percent.
Corrected 28 August 2026. This finding was published as "premium cabins are never a good use of miles", with "seven airlines, seven penalties, no exceptions". That was an overstatement and the data behind it does not carry it.
Three reasons. Measured against its own program's economy median rather than against a single paired observation, Southwest's premium cabin came out two percent better, so there was an exception. The observation sheet also contains a business redemption at 4.01 cents per point, roughly three times that program's economy figure, set aside under the seven night rule for good reasons but fatal to the word "never". And each program's premium figure rests on one observation: the three transatlantic ones priced comparable business at 400,000, 745,000 and 905,000 points on the same dates, a spread that says those were dynamic prices caught on one day rather than what the cabin usually costs.
What stands is narrower and still worth knowing. If you need a particular week, the premium award on that week is usually poor value, and the penalty is small domestically and large across the Atlantic. That is a finding about fixed dates, not about cabins. Flexible dates were never tested here, and a source that values redemptions people actually chose to book reports premium doing considerably better. Working in reference/RECEIPTS.md.
Holiday dates take value out of your points
On five of seven programs, Thanksgiving week returned less per point than an ordinary October week, by between six and sixty seven percent. Cash goes up and the award price goes up faster. The exceptions were the two programs that price awards from a published chart rather than off the fare.
Partner awards are usually the worst row, not the best
Three of the four programs where we measured one put the partner award at or near the bottom. The reason sits in the fees. To the same city on the same dates, one airline charged $232 flying its own aircraft and $489 flying a partner's. That difference is the operating carrier's surcharge, and it follows the aeroplane rather than the program.
Hotel programs have opposite shapes
Hyatt publishes a category chart, and its value barely moves: 2.01 to 2.48 cents across everything from an airport hotel to Park Hyatt Tokyo. Hilton does not, and its value runs 0.42 to 1.82. Cash at the expensive end rose almost fourteen times while the points cost rose barely three. Hyatt points are worth the same wherever you spend them. Hilton points are worth four times more at expensive properties than cheap ones.
What could not be measured, and why
Four programs will not show you what an award costs until you already hold their points. Not a rounding of the price, not a range: no number at all until there is a balance in the account.
That is the main reason eight programs were still estimates. Six of the eight are also programs Americans transfer into for one specific trip rather than earn in, so award space on a partner's aircraft is thin and unpredictable even once you are through the door.
One program was held rather than blocked. Air Canada Aeroplan was measured, but from Toronto, Montreal and Vancouver. Every currency in the table is a US card product, so the redemption that matters departs a US airport, and a program measured in its own home market is measuring the wrong thing. Those five observations still inform the findings above, and the figure was left on an estimate rather than published.
We would rather say that plainly than publish a number we could not stand behind. It is also worth knowing on its own: a currency you cannot price before committing to it is a currency worth more scepticism.
Where our own estimates were wrong
The nine remaining estimates were the survivors of a set of twenty five written in May 2026. Measuring ten of them showed how far off those guesses were, and in which direction.
Eight of the ten came in above the estimate, and all four hotel programs did. Hyatt went from 1.7 to 2.20, Marriott from 0.7 to 1.05, Hilton from 0.5 to 0.80. The reason is the taxes rule: a room rate hides fifteen to twenty percent that an award does not charge you, and the guesses had quietly inherited that gap.
Two came in lower, both airlines: American at 1.5 against a measured 1.20, and Delta at 1.2 against 1.15. Aeroplan was further out than either, at 1.5 against 0.90, but that measurement came from the wrong origin and was held rather than published.
Which implies the eight estimates left on the table were probably low too, particularly the two hotel ones. We would rather tell you that than let you assume otherwise.
Why this stopped at sixteen programs
Sixteen of twenty five carried a number by the end, from 99 observations gathered over eight days. That works out at roughly six observations per program, which is enough to be worth publishing with its sample size attached and not enough to maintain as a living table across twenty five programs alongside building the app.
So this became a study rather than a product. What Wanderpoints benchmarks against now is described on how we value points, and the reasoning for the change is there too.
Every observation recorded the property or flight number, the date, the room or fare type, and the total actually charged at checkout, and the published figures were rebuilt from those observations rather than edited by hand. The measurements were made between 25 and 28 August 2026. The estimates were set on 1 May 2026. No number on this page has been changed since and none will be. Where a conclusion drawn from those numbers turns out to overreach, it is corrected in place with a date and with what it previously said, as the premium cabin finding was on 28 August 2026. A measurement quietly revised afterwards is not a measurement, and a claim left standing after its own data contradicts it is not honest either.
How we value points today · Back to Wanderpoints
