Running the honest math…
We use a conservative baseline for points math and publish a higher achievable estimate separately. These are planning values, so compare them with the redemption you can actually book before deciding whether a card is worth it.
Conservative by default
Point values below our achievable estimate
100% auditable
Every baseline published in full
We say skip cards
Pay cash when the math demands it
Cash paid is money leaving your account. Points spent are a balance you give up. The baseline assigns that balance a planning value; the higher achievable estimate describes a different possible redemption. Neither is a cash refund or guaranteed savings.
Example: New York to London, round trip, one traveler. A $650 fare divided by an estimated 40,000 points gives 1.6¢ per point. At a 1¢ baseline those points represent $400. The gated engine answer is Use points. Award cash fees are omitted; add the actual fees before deciding.
Follow the complete worked receipt or check our operating rules.
| Program | Baseline | Achievable | Upside |
|---|---|---|---|
| Chase Ultimate Rewards | 1c | 2c | |
| Amex Membership Rewards | 1.1c | 2c | |
| Capital One miles | 1c | 1.7c | |
| Citi ThankYou Points | 1c | 1.8c | |
| Bilt Rewards | 1.1c | 2c | |
| Wells Fargo Rewards | 1c | 1.5c | |
| United MileagePlus | 1.3c | 1.5c | |
| Delta SkyMiles | 1.1c | 1.3c | |
| American AAdvantage | 1.3c | 1.6c | |
| Southwest Rapid Rewards | 1.3c | 1.4c | |
| Atmos Rewards (Alaska) | 1.4c | 1.8c | |
| JetBlue TrueBlue | 1.1c | 1.3c | |
| Marriott Bonvoy | 0.7c | 0.85c | |
| Hilton Honors | 0.4c | 0.5c | |
| World of Hyatt | 1.7c | 2.3c | |
| IHG One Rewards | 0.5c | 0.6c |
We use the baseline column for all card and trip math. The achievable column is what a motivated person can hit with transfer partners: the bar shows how much upside each program leaves on the table.
When you redeem points, we compute cents per point = (cash price ÷ points used) × 100, then grade it against the program's baseline. This single check is the most common way people quietly lose value, and almost no tool flags it.
Below ~80% of baseline: pay cash, save the points.
Just under baseline: roughly a wash, your call.
At baseline : a solid redemption, book it.
Above achievable : an outright steal.
Worked example
Chase baseline checked 2026-09-07: standard value and selected Points Boost offers. Boosts and legacy eligibility are not assumed. Other dataset review dates remain unchanged.
What we DON'T count
The optimizer makes a few judgment calls with fixed cutoffs. Rather than hide them, here they are: the same constants the live engine uses, so you can check our work and disagree if your situation differs.
$200
Trips under this total cash cost are too cheap to optimize : we tell you to just pay cash rather than burn points or chase a card's minimum spend for pennies.
70%
We only suggest a new card when your existing points plus its signup bonus reach at least this share of what the trip needs. Below it, the bonus is a token contribution and pushing the card would be selling, not helping.
0.5×
A credit you mark "maybe" counts at half its face value: you'll catch some of it, not all. Credits you'll definitely use count fully; ones you won't count zero.
A live cash price is the single biggest input: without it we can grade the points play only as an estimate, so a card verdict stays unconfirmed until you enter the fare you actually see.