The question
OptimalCardSetup's free tier lets you build a wallet of up to 2 cards. We wanted to answer two questions with data instead of opinion. First, when a constraint solver picks the highest-value 2-card wallet for many different households, which cards actually show up? Second, how much modeled value does the 2-card limit cost compared with a 5-card wallet?
To find out, we built 252 synthetic household spending profiles and ran each one through the same optimization code that powers the optimizer. This report covers what came back. It describes the cards that won for these profiles under stated assumptions. It is not a ranking of the best cards and not financial advice.
Share of optimal 2-card wallets that include Citi Double Cash
48.4%
Method
- Card data: the catalog verified as of September 22, 2026, with 53 cards. A default request leaves 45 of them eligible. Business cards, the one debit card, and three cards closed to new applicants are excluded, as they are for any anonymous user. See the full card list.
- Profiles: 252 synthetic households, 36 for each of seven archetypes. Each archetype has a monthly dollar range for each spend category. A seeded random generator drew each household's amounts from those ranges and rounded them to the nearest $5. We chose the ranges ourselves so the households would look plausible. They are illustrative. They do not come from survey data, and they do not represent the U.S. population.
- Solve: each profile went through the same request handling, constraint solver, and result calculation as the live optimizer. It was solved once at the free tier's 2-card limit and once at a 5-card limit. Results are in cash mode, which values every point at its cash redemption value. Most bank points count at 1 cent, Amex Membership Rewards points at 0.6 cent, and Hilton Honors points at 0.5 cent. Card credits count at the catalog's conservative values, as they do for any free-tier user. The Bilt rent switch stayed at its default (off), so rent earned nothing on any card.
- Proof: the solver proved all 1,260 main solves optimal: 252 profiles times five runs (cash at 2 cards, cash at 5 cards, $0-fee cards only at 2, travel mode at 2, travel mode at 5). No result was dropped for lack of proof.
- Hand checks: we recalculated 11 of the 2-card wallets and 4 of the 5-card wallets by hand from the cards' earn rates and caps. All 15 matched the solver's figure to the cent.
The seven archetypes
| Archetype | Profiles | Median monthly spend (excl. rent) | Range |
|---|---|---|---|
| Student | 36 | $715 | $545 to $1,030 |
| Retiree / low spend | 36 | $1,635 | $1,185 to $2,010 |
| Urban renter | 36 | $2,202.50 | $1,625 to $2,765 |
| Remote worker | 36 | $2,640 | $2,085 to $3,305 |
| Young family | 36 | $3,260 | $2,715 to $3,770 |
| Frequent traveler | 36 | $3,920 | $2,880 to $4,450 |
| Suburban family | 36 | $5,167.50 | $4,525 to $5,880 |
Urban renters also pay rent, with a median of $2,082.50 a month. With the Bilt switch off, rent earns nothing on any card, so it doesn't change their results. Frequent travelers book $400 to $1,200 of flights and $300 to $900 of hotels directly every month. The per-category ranges for every archetype are published with the method notes (see Reproduce, below).
Finding 1: twelve cards cover every 2-card wallet
Of the 45 eligible cards, only 12 appeared in any optimal 2-card wallet, in 21 different pairs. These are the ten that appeared most often:
| Card | Annual fee | Share of 2-card wallets | Where it appeared most |
|---|---|---|---|
| Citi Double Cash | $0 | 48.4% (122) | Young family 86.1%, suburban family 83.3%, frequent traveler 80.6% |
| Amex Blue Cash Preferred | $95 | 40.1% (101) | Young family 88.9%, retiree 63.9%, remote worker 61.1% |
| US Bank Cash+ | $0 | 36.9% (93) | Retiree 91.7%, student 55.6% |
| Hilton Honors Aspire | $550 | 20.2% (51) | Urban renter 83.3% |
| Capital One Savor | $0 | 20.2% (51) | Suburban family 72.2%, student 41.7% |
| Amex Blue Cash Everyday | $0 | 14.7% (37) | Student 61.1% |
| Wells Fargo Autograph Journey | $95 | 14.3% (36) | Frequent traveler 100% |
| US Bank Altitude Go | $0 | 2.0% (5) | Student 11.1% |
| Prime Visa | $0 (requires Prime) | 1.2% (3) | Student 5.6% |
| Wells Fargo Autograph | $0 | 1.2% (3) | Urban renter 5.6% |
The most common pair was Amex Blue Cash Preferred with Citi Double Cash, which was optimal for 58 profiles (23.0%). The pattern behind most of the table is one flat-rate card for everyday spending plus one category specialist. Blue Cash Preferred covers groceries, Cash+ covers bills and other chosen categories, and Savor covers dining and entertainment. That is the trade-off we explain in flat-rate vs. category cards, here backed by 252 solves. Double Cash won where everyday spending was high. It appeared in 1 of 36 retiree wallets and none of the student wallets, because a category card beat a flat 2% at those spending levels.
Finding 2: what the 2-card limit costs
Median extra modeled value from a 5-card limit instead of 2
$239.60/yr
| Archetype | Median net, 2 cards | Median net, 5 cards | Median gain | Median gain, % |
|---|---|---|---|---|
| Student | $236.40 | $289.80 | $58.80 | 23.8% |
| Retiree / low spend | $547.30 | $699.70 | $146.70 | 27.2% |
| Urban renter | $685.00 | $866.75 | $184.05 | 27.1% |
| Frequent traveler | $1,337.40 | $1,600.20 | $255.60 | 18.9% |
| Remote worker | $807.60 | $1,094.20 | $283.25 | 35.3% |
| Young family | $978.40 | $1,292.50 | $316.50 | 32.3% |
| Suburban family | $1,492.35 | $1,931.35 | $458.50 | 30.0% |
Net value is modeled annual rewards minus annual fees, after the catalog's conservative credit values. Every profile gained something from the extra slots. The smallest gain was $39.00 and the largest was $506.20, for a suburban family. In dollars, the gain grew with spending. In percentage terms, remote workers gained the most (35.3%). Frequent travelers gained the least (18.9%), because their biggest categories were already covered within two slots.
Every 5-card wallet used all five slots. US Bank Cash+ was in all 252 of them. Citi Double Cash was in 90.1%, Blue Cash Preferred in 72.6%, US Bank Altitude Go in 56.7%, and Prime Visa in 54.4%. Prime Visa's 5% rates assume an Amazon Prime membership, and the Costco Anywhere Visa (15.5% of 5-card wallets) requires a Costco membership. The model does not subtract either membership fee.
Valuing points at the optimizer's default travel valuations instead of cash lowers the median 2-to-5 gain to $169.22 (16.2%). Our earlier Q3 two-card report averaged $190.80 across four profiles on the June catalog. This report's mean is $242.02 across 252 different profiles on the September catalog. Because the profiles and catalogs differ, the two figures are not a trend.
Finding 3: most 2-card wallets carry one modest fee
In 166 of the 252 optimal 2-card wallets (65.9%), at least one card had an annual fee. Only three fee cards appeared at all: Blue Cash Preferred ($95, 101 wallets), Hilton Honors Aspire ($550, 51), and Autograph Journey ($95, 36). Among wallets with a fee card, the median sticker fee was $95, and the median effective fee after the modeled credits was also $95. Thirty-four wallets had a modeled effective fee of $0. All of them paired Hilton Honors Aspire with a $0-fee card (see Finding 5).
How often a fee card appeared varied by archetype: every frequent traveler wallet, 88.9% of young-family wallets, 86.1% of urban-renter wallets, but only 22.2% of student wallets.
Finding 4: the best $0-fee pair is usually close
Share of profiles whose best $0-fee pair is within $25/yr of their best pair overall
65.1%
We re-solved every profile at 2 cards with every annual-fee card excluded. The median shortfall against the overall best pair was $15.35 a year, and the largest was $215.20. Frequent travelers were the exception. Only 5 of 36 had a $0-fee pair within $25, and their median shortfall was $121.60. Autograph Journey's hotel and flight rates are worth more than its $95 fee at those travel levels. Removing it cost a median $113.95, and the $0-fee Wells Fargo Autograph replaced it in 33 of 36 wallets. Removing Blue Cash Preferred from the wallets it was in cost a median $25.00.
For a fee-by-fee way to run the same test on your own cards, see when annual fee cards are worth it and our annual-fee wallet report.
Finding 5: a $550 card that wins on credits
The surprise in the table is Hilton Honors Aspire. A $550-fee card was optimal in 51 of the 2-card wallets (20.2%), including 30 of the 36 urban renters. The reason is credits. The catalog counts three of Aspire's credits at conservative values: $200 for the airline credit, $200 of the $400 resort credit, and $200 for the annual free night. That is $600 a year against a $550 fee, so the modeled effective fee is $0. In cash mode, Hilton points are worth 0.5 cent each, so Aspire earns 3.5% back on dining and directly booked flights and 1.5% on most other purchases.
We checked whether those credits drive the result by re-solving all 51 profiles with every Aspire credit valued at $0 (a Premium setting). Aspire dropped out of every one of the 51 wallets. The median value of the next-best wallet was $16.80 a year lower, and the largest difference was $74.70. Citi Double Cash was the most common replacement (30 wallets), followed by Wells Fargo Autograph (13). In other words, Aspire wins these wallets by a small margin, and only if you would actually use its airline credit, resort credit, and free night.
The same pattern appeared in travel mode. There, Capital One Venture X was in 69.8% of the optimal 2-card wallets, Amex Gold in 36.9%, and Bilt Palladium in 10.3%. With each card's credits valued at $0, it dropped out of every wallet it had been in. The next-best wallets were a median $23.61 (Venture X), $26.58 (Amex Gold), and $19.97 (Palladium) a year lower. Each of the four cards whose modeled credits we set to zero won these wallets mainly on those credits, by a median margin of about $17 to $27 a year over the next-best wallet, not by a clear lead.
What about rent?
We also turned on the Bilt rent switch for the 36 urban renters. Those solves are much harder. Within the optimizer's 60-second limit, only 4 of the 36 were proven optimal, so we are not publishing a Bilt finding from this run. For the tier mechanics, see how Bilt's rent tiers work.
What this doesn't say
- Synthetic profiles. The 252 households are illustrative. They are not a survey, a sample of our users, or a picture of the U.S. population. The results apply only to these inputs.
- Modeled, cash-mode values. Results depend on how points are valued. Travel valuations change which cards win, as the travel-mode results above show.
- Welcome offers are not modeled. The figures are ongoing yearly value. A card's first-year value can differ a lot.
- Approval is not modeled. Issuer eligibility rules and credit decisions are outside the model. See issuer application rules.
- Rotating categories are not modeled. Cards whose top rate depends on a quarterly category you have to activate are valued at their fixed rates only.
- Credits are conservative estimates. If you would not use a card's credits, Finding 5 shows how much that can change the answer.
- Memberships are not charged. Prime Visa and Costco Anywhere Visa assume a membership the model does not price.
- Not a ranking and not advice. Check each issuer's current terms before you apply.
Reproduce it
Everything here comes from one seeded script run against the September 22, 2026 catalog. The script is backend/scripts/data_report_q4_2026.py (seed 20261007), and it writes every profile's spend and result to docs/reports/data-report-q4-2026.json. The method notes, per-category ranges, hand checks, and exact command are in docs/reports/data-report-q4-2026.md. These files live in our private repository. If you want them for your own analysis, ask through support. You can also re-enter any profile's spending in the optimizer yourself.
Your wallet is not an archetype
These 252 households are stand-ins. Your own spending will put you somewhere in these tables, or outside them. Run the optimizer with your real numbers at 2 cards, then compare against more cards or against $0-fee cards only. To see two specific cards side by side, use card comparisons.
Catalog verified September 22, 2026. Solver runs October 7, 2026, in-process, cash mode unless stated; all 1,260 main solves proven optimal.