Event tickets with seat maps
Choose seats on seat maps under constraints.
20 project-authored proposals · taxonomy v1.1 · 0 new benchmark runs. Verifiers and partial-credit rules below are designs, not executed results.
- Typical applications
- Ticketmaster · Damai
- Class GUI-need label
G3· GUI-only
Task proposals · 20
Expand to inspect the complete original wording and proposed checks. Original proposal text remains English in both interfaces.
R12.04.01.03-T01 Pick three adjacent cinema seats in the center third of rows F to H
Original instruction
On the cinema-booking clone, book 3 adjacent seats in the center third of rows F-H for the 19:15 IMAX showing of 'Orbit' on Saturday. Seats are drawn on a canvas seat map. Choose the triple closest to the screen center line. Stop at payment.
- Platforms
web· Any OS (web)- Applications
- cinema booking clone
- Interfaces
brow· Browser- GUI need · task setting
G3· GUI-only- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
click· Point & clickvis· Visual judgmentgeom· Precise geometry- Work product
WP08· App/system state- Conditions
risk· Irreversible/risky- Proposed difficulty
- 2 / 5
Proposed verifier
Held seats: 3 seat IDs in one row F-H, consecutive per seat graph (aisles break adjacency), all within the center third of the row, minimal distance to center among available triples; status=held. The clone issues no API token and accepts seat holds only from browser sessions (CSRF + session id), so scripted holds are rejected.
Proposed partial credit
30% adjacency; 30% row and center-third; 30% closest-to-center; 10% stopped at payment.
R12.04.01.03-T02 Select two concert seats in a section under a price cap with one aisle seat
Original instruction
On the Ticketmaster-style clone, select 2 adjacent seats in section 104 for the 'Northern Lights Tour' on 2026-11-21, each <= $120 including fees (fees shown on hover), with one seat on the aisle. Proceed to checkout and stop at payment.
- Platforms
web· Any OS (web)- Applications
- Ticketmaster-style clone
- Interfaces
brow· Browser- GUI need · task setting
G3· GUI-only- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
hover· Hover & contextclick· Point & clickvis· Visual judgment- Work product
WP08· App/system state- Conditions
risk· Irreversible/risky- Proposed difficulty
- 2 / 5
Proposed verifier
Cart: 2 seats in sec 104, adjacent, one with aisle flag, each all-in price <= 120 (fees only in hover tooltip); not purchased. The clone issues no API token and accepts seat holds only from browser sessions (CSRF + session id), so scripted holds are rejected.
Proposed partial credit
30% section and adjacency; 30% aisle; 30% price cap incl. fees; 10% stop.
R12.04.01.03-T03 Book a wheelchair space and companion seat for a theater show
Original instruction
On the theater-booking clone, book 1 wheelchair space plus 1 adjacent companion seat for 'Wicked' at the Saturday 14:00 matinee, preferring the orchestra level, price tier B or cheaper. Accessible spaces are marked with an icon on the seat map. Stop at payment.
- Platforms
web· Any OS (web)- Applications
- theater booking clone
- Interfaces
brow· Browser- GUI need · task setting
G3· GUI-only- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
vis· Visual judgmentclick· Point & clicknav· Pan/zoom/scroll- Work product
WP08· App/system state- Conditions
risk· Irreversible/risky- Proposed difficulty
- 3 / 5
Proposed verifier
Cart: one seat with accessible=wheelchair, one companion seat adjacent (seat graph), orchestra if any available, tier <= B; held, not purchased. The clone issues no API token and accepts seat holds only from browser sessions (CSRF + session id), so scripted holds are rejected.
Proposed partial credit
40% wheelchair space; 30% adjacent companion; 20% level/tier; 10% stop.
R12.04.01.03-T04 Exchange tickets to an earlier showtime keeping the same seats if possible
Original instruction
On the cinema clone, the user's order for 'Orbit' at 21:30 (seats G7-G8) should move to the 18:45 showing the same day. Keep G7-G8 if free; otherwise choose the nearest adjacent pair in row G or the rows next to it. Complete the exchange (no extra charge expected).
- Platforms
web· Any OS (web)- Applications
- cinema booking clone
- Interfaces
brow· Browser- GUI need · task setting
G3· GUI-only- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
click· Point & clickvis· Visual judgment- Work product
WP08· App/system state- Conditions
risk· Irreversible/risky- Proposed difficulty
- 3 / 5
Proposed verifier
Order exchanged to 18:45; seeded G7 taken → seats = nearest adjacent pair per seat-distance metric; original showtime seats released; one order id. The clone issues no API token and accepts exchanges only from the browser session (CSRF + session id), so scripted requests are rejected.
Proposed partial credit
40% showtime exchanged; 40% seat choice; 20% original released.
R12.04.01.03-T05 Buy discounted museum tickets with mixed ticket types and a timed entry
Original instruction
On the aquarium ticketing clone, buy tickets for Sunday: 1 adult, 1 veteran (requires ID verification checkbox), 2 children (ages 5 and 11; under-6 is free but still needs a ticket), timed entry 10:30. Apply the membership discount code from fixtures/member_card.png. Pay with sandbox card.
- Platforms
web· Any OS (web)- Applications
- aquarium ticketing clone
- Interfaces
brow· Browser- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
vis· Visual judgmentclick· Point & clicktype· Text & forms- Work product
WP08· App/system state- Conditions
risk· Irreversible/risky- Proposed difficulty
- 2 / 5
Proposed verifier
Order: ticket types ADULT 1, VETERAN 1 (+ID ack), CHILD 1, CHILD_FREE 1; timed slot 10:30 Sunday; discount code applied (read from image); total equals backend. The clone issues no API token and accepts orders only from the browser session (CSRF + session id), so scripted checkouts are rejected.
Proposed partial credit
40% ticket types; 20% timed slot; 20% discount; 20% order placed.
R12.04.01.03-T06 Choose four seats together including an aisle at the lowest tier on a stadium map
Original instruction
On the stadium ticketing clone, select 4 seats together (same row, consecutive) with at least one aisle seat, in the lowest price tier that has such a block for the 2026-12-06 game, preferring sideline sections over end zones within that tier. Stop at payment.
- Platforms
web· Any OS (web)- Applications
- stadium ticketing clone
- Interfaces
brow· Browser- GUI need · task setting
G3· GUI-only- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
nav· Pan/zoom/scrollvis· Visual judgmentclick· Point & click- Work product
WP08· App/system state- Conditions
risk· Irreversible/risky- Proposed difficulty
- 3 / 5
Proposed verifier
Predicate on held seats: same row, consecutive, includes aisle-flagged seat, tier = lowest tier with a feasible block, sideline if feasible in that tier. The clone issues no API token and accepts seat holds only from the browser session (CSRF + session id).
Proposed partial credit
30% 4 together; 20% aisle; 30% lowest tier; 20% sideline preference.
R12.04.01.03-T07 Select seats on a Damai-style mobile app with real-name ticket holders
Original instruction
On the Damai-style app clone (Android emulator), buy 2 tickets for the 2026-12-31 concert in the 680-yuan tier, seats adjacent, in the floor area left block (A区), and bind the two real-name attendees from the user's saved list ('Wang Fang', 'Zhou Lei'). Stop at the payment-password screen.
- Platforms
and· Android- Applications
- Damai clone
- Interfaces
mob· Mobile touch- GUI need · task setting
G3· GUI-only- GUI need · ordinary access
G3· GUI-only- Interaction primitives
touch· Touch gesturesvis· Visual judgmentnav· Pan/zoom/scroll- Work product
WP08· App/system state- Conditions
risk· Irreversible/risky- Proposed difficulty
- 3 / 5
Proposed verifier
Draft order: tier 680, block A, 2 adjacent seats (pinch-zoom seat map), attendees bound to the two saved IDs; status=awaiting_payment; no password input. No adb shell or emulator console is exposed, and the clone backend accepts seat holds only from the app session (no API token).
Proposed partial credit
25% tier; 35% block + adjacency; 30% attendees; 10% stop at payment.
R12.04.01.03-T08 Find the best-view seats using a 'view from seat' preview
Original instruction
On the arena ticketing clone, the user wants 2 adjacent seats under $90 each whose 'view from seat' preview image shows the full stage without a pillar obstruction. Several cheap sections have obstructed views not labeled in text. Pick the cheapest qualifying pair and stop at payment.
- Platforms
web· Any OS (web)- Applications
- arena ticketing clone
- Interfaces
brow· Browser- GUI need · task setting
G3· GUI-only- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
vis· Visual judgmenthover· Hover & contextclick· Point & click- Work product
WP08· App/system state- Conditions
risk· Irreversible/risky- Proposed difficulty
- 4 / 5
Proposed verifier
Held seats: price <= 90, adjacent, section's ground-truth obstruction flag = false (determined only from preview images), cheapest such pair. The clone issues no API token and accepts seat holds only from browser sessions (CSRF + session id), so scripted holds are rejected.
Proposed partial credit
50% unobstructed; 30% cheapest; 20% adjacency + stop.
R12.04.01.03-T09 Buy resale tickets checking that seats in a listing are actually adjacent
Original instruction
On the resale marketplace clone, buy 2 seats for the Dec 12 hockey game where both seats are adjacent (some listings say 'together' but seat numbers are non-consecutive across an aisle), total <= $260 including fees, in the lower bowl. Use the map highlight to check. Stop at payment.
- Platforms
web· Any OS (web)- Applications
- resale marketplace clone
- Interfaces
brow· Browser- GUI need · task setting
G3· GUI-only- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
vis· Visual judgmentclick· Point & clicknav· Pan/zoom/scroll- Work product
WP08· App/system state- Conditions
risk· Irreversible/risky- Proposed difficulty
- 3 / 5
Proposed verifier
Selected listing's seats adjacent per seat graph (not just label), lower bowl, all-in <= 260, cheapest such listing; held. The clone issues no API token and accepts seat holds only from browser sessions (CSRF + session id), so scripted holds are rejected.
Proposed partial credit
40% true adjacency; 30% price; 20% section; 10% stop.
R12.04.01.03-T10 Reserve train seats facing direction of travel with a table on a seat map
Original instruction
On the European rail booking clone, reserve 2 second-class seats on ICE 578 (2026-11-20, Frankfurt to Berlin) facing the direction of travel, at a table, adjacent, in a quiet zone car. Seat map icons indicate direction and table. Stop at payment.
- Platforms
web· Any OS (web)- Applications
- rail booking clone
- Interfaces
brow· Browser- GUI need · task setting
G3· GUI-only- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
vis· Visual judgmentclick· Point & clicknav· Pan/zoom/scroll- Work product
WP08· App/system state- Conditions
risk· Irreversible/risky- Proposed difficulty
- 3 / 5
Proposed verifier
Reservation seats: car with quiet-zone flag, seat attributes facing_forward=true and table=true (icons only), adjacent; second class; held. The clone issues no API token and accepts seat holds only from browser sessions (CSRF + session id), so scripted holds are rejected.
Proposed partial credit
30% quiet car; 30% forward-facing; 20% table; 20% adjacency + stop.
R12.04.01.03-T11 Book an airline-style seat map for a ferry cabin with bunk constraints
Original instruction
On the ferry booking clone, book an overnight crossing for 3 (2 adults, 1 child 8) in a single inside cabin with 4 berths, lower berths for the adults, child in an upper berth (allowed age >= 6), and a vehicle space for one car (plate from fixtures/car.txt). Stop at payment.
- Platforms
web· Any OS (web)- Applications
- ferry booking clone
- Interfaces
brow· Browser- GUI need · task setting
G2· GUI-essential- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
click· Point & clickvis· Visual judgmenttype· Text & forms- Work product
WP08· App/system state- Conditions
risk· Irreversible/risky- Proposed difficulty
- 3 / 5
Proposed verifier
Booking: one cabin type inside-4, berth assignments per passenger as constrained (berth diagram), vehicle with plate; status=held. The clone issues no API token and accepts bookings only from the browser session (CSRF + session id), so scripted requests are rejected.
Proposed partial credit
30% cabin; 40% berth assignment; 20% vehicle; 10% stop.
R12.04.01.03-T12 Choose festival camping plus tickets with a group of five on a zone map
Original instruction
On the festival ticketing clone, buy 5 weekend passes and one group camping plot in zone C on the camping map that is within 150 m of water taps (map legend) and adjacent to the plot already reserved by the user's friend group (plot C-44). Stop at payment.
- Platforms
web· Any OS (web)- Applications
- festival ticketing clone
- Interfaces
brow· Browser- GUI need · task setting
G3· GUI-only- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
nav· Pan/zoom/scrollvis· Visual judgmentclick· Point & click- Work product
WP08· App/system state- Conditions
risk· Irreversible/risky- Proposed difficulty
- 3 / 5
Proposed verifier
Cart: 5 passes; plot in zone C, adjacent to C-44 per plot graph, distance to nearest tap <= 150 m (map coordinates); held. The clone issues no API token and accepts seat holds only from browser sessions (CSRF + session id), so scripted holds are rejected.
Proposed partial credit
30% passes; 40% plot adjacency; 20% water distance; 10% stop.
R12.04.01.03-T13 Hold the best seats available at release time in a virtual queue
Original instruction
On the ticketing clone, tickets for 'Aurora Live' go on sale at a scripted time. Join the virtual queue (no automation of the queue page), and when admitted, within the 8-minute cart timer select 2 adjacent seats in the best available price tier <= $200 each, closest to the stage. Stop at payment. Do not open multiple sessions.
- Platforms
web· Any OS (web)- Applications
- ticketing clone
- Interfaces
brow· Browser- GUI need · task setting
G3· GUI-only- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
click· Point & clickvis· Visual judgmentrt· Real-time control- Work product
WP08· App/system state- Conditions
risk· Irreversible/riskyinterrupt· Interruptions- Proposed difficulty
- 4 / 5
Proposed verifier
Single session id in queue log; seats held before timer expiry; adjacency; tier constraint; distance-to-stage rank among available pairs at selection time (inventory changes live) within top 3. The clone issues no API token and queue/cart requests must carry the browser session's per-page nonce, so scripted requests are rejected and logged.
Proposed partial credit
20% queue compliance; 30% held in time; 30% constraints; 20% near-best seats.
R12.04.01.03-T14 Split a party of six into two adjacent rows when no row has six seats
Original instruction
On the theater clone, book 6 seats for 'Hamlet' Friday 19:30 in the stalls: no single row has 6 consecutive seats, so pick 3+3 directly front-back aligned in consecutive rows (seat numbers matching), as central as possible. Stop at payment.
- Platforms
web· Any OS (web)- Applications
- theater booking clone
- Interfaces
brow· Browser- GUI need · task setting
G3· GUI-only- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
vis· Visual judgmentclick· Point & clickgeom· Precise geometry- Work product
WP08· App/system state- Conditions
risk· Irreversible/risky- Proposed difficulty
- 4 / 5
Proposed verifier
Held seats: two blocks of 3 consecutive in consecutive rows with aligned seat numbers (seat geometry, staggered rows count as aligned via map x-coordinates ±1 seat), centrality score among top 2 feasible. The clone issues no API token and accepts seat holds only from the browser session (CSRF + session id).
Proposed partial credit
40% 3+3 consecutive; 30% aligned rows; 20% centrality; 10% stop.
R12.04.01.03-T15 Use a season-ticket holder's presale code and transfer one ticket to a friend
Original instruction
On the team ticketing clone (logged in test season-ticket account), buy 2 extra seats for the playoff game using presale code in fixtures/presale_email.eml, seats adjacent to the account's season seats (Sec 210, Row 8, Seats 5-6). Then transfer one of the new tickets to 'friend@example.test' via the transfer flow, stopping before 'Send transfer'.
- Platforms
mac· macOSwin· Windows- Applications
- team ticketing clone · Mail
- Interfaces
brow· Browserdesk· Desktop GUI- GUI need · task setting
G3· GUI-only- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
multi· Cross-appvis· Visual judgmentclick· Point & click- Work product
WP08· App/system state- Conditions
multiapp· Multi-apprisk· Irreversible/risky- Proposed difficulty
- 4 / 5
Proposed verifier
Order with presale code: seats 210/8/3-4 or 7-8 (adjacent to 5-6) purchased with sandbox card; transfer draft for one new seat to the email; status=draft. The clone issues no API token and accepts purchases and transfers only from the browser session (CSRF + session id), so scripted requests are rejected.
Proposed partial credit
20% presale code; 40% adjacent seats; 20% purchase; 20% transfer draft stopped.
R12.04.01.03-T16 Book a planetarium show with recliner seats avoiding the projector shadow area
Original instruction
On the planetarium booking clone, the dome seat map is circular. Book 2 adjacent recliner seats for the 16:00 show, avoiding the seats flagged in the FAQ diagram (image) as having partial projector shadow, and not in the front ring. Stop at payment.
- Platforms
web· Any OS (web)- Applications
- planetarium booking clone
- Interfaces
brow· Browser- GUI need · task setting
G3· GUI-only- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
vis· Visual judgmentclick· Point & clickgeom· Precise geometry- Work product
WP08· App/system state- Conditions
risk· Irreversible/risky- Proposed difficulty
- 4 / 5
Proposed verifier
Held seats: adjacent on the circular seat graph, recliner type, not in shadow set (from FAQ image) and not ring 1. The clone issues no API token and accepts seat holds only from browser sessions (CSRF + session id), so scripted holds are rejected.
Proposed partial credit
30% adjacency on circular map; 40% shadow area avoided; 20% ring/type; 10% stop.
R12.04.01.03-T17 Upgrade existing tickets to better seats paying only the difference
Original instruction
On the Ticketmaster-style clone, the user holds 2 seats in Sec 312 for the Dec 6 concert. Use the 'Upgrade seats' flow to move to 2 adjacent seats in a 100-level section with the smallest price difference, max $150 total difference. Complete with the sandbox card.
- Platforms
web· Any OS (web)- Applications
- Ticketmaster-style clone
- Interfaces
brow· Browser- GUI need · task setting
G3· GUI-only- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
click· Point & clickvis· Visual judgmentnav· Pan/zoom/scroll- Work product
WP08· App/system state- Conditions
risk· Irreversible/risky- Proposed difficulty
- 3 / 5
Proposed verifier
Order upgraded (same order id): 2 adjacent seats in 1xx section, difference = min available and <= 150; old seats released. The clone issues no API token and accepts upgrades only from the browser session (CSRF + session id), so scripted requests are rejected.
Proposed partial credit
40% 100-level adjacent; 40% minimal difference; 20% old seats released.
R12.04.01.03-T18 Plan a concert night: tickets, parking pass and nearby dinner reservation
Original instruction
For the 2026-11-28 concert at Riverside Arena: buy 2 adjacent seats in sections 101-105 under $150 each on the ticketing clone, add a parking pass for the lot closest to the gate nearest those sections (venue map), and reserve dinner for 2 at 17:30 within a 10-minute walk of that gate on the OpenDining clone. Stop at payment for tickets.
- Platforms
web· Any OS (web)- Applications
- ticketing clone · OpenDining clone · venue map
- Interfaces
brow· Browser- GUI need · task setting
G3· GUI-only- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
multi· Cross-appnav· Pan/zoom/scrollvis· Visual judgment- Work product
WP08· App/system state- Conditions
multiapp· Multi-apprisk· Irreversible/risky- Proposed difficulty
- 5 / 5
Proposed verifier
Cart: seats meeting constraints + parking lot id nearest the gate serving those sections (seeded venue graph); reservation at a restaurant within 800 m walk of that gate at 17:30. Neither clone issues an API token; holds and reservations are accepted only from browser sessions (CSRF + session id), so scripted requests are rejected.
Proposed partial credit
35% seats; 30% parking lot; 30% dinner; 5% stop at payment.
R12.04.01.03-T19 Reserve opera seats with surtitle visibility using a sightline diagram
Original instruction
On the opera-house booking clone, book 2 adjacent seats in the balcony for 'La Bohème' on 2026-12-12 from which the surtitle screen is visible (seats with surtitle visibility are shaded in the 'Sightlines' PDF diagram linked on the page, not on the seat map), cheapest such pair. Stop at payment.
- Platforms
mac· macOSwin· Windowsweb· Any OS (web)- Applications
- opera booking clone · PDF viewer
- Interfaces
brow· Browser- GUI need · task setting
G3· GUI-only- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
vis· Visual judgmentmulti· Cross-appclick· Point & click- Work product
WP08· App/system state- Conditions
risk· Irreversible/riskymultiapp· Multi-app- Proposed difficulty
- 4 / 5
Proposed verifier
Held seats in balcony, adjacent, within the shaded visibility set (ground-truth list derived from the diagram), cheapest such pair. The clone issues no API token and accepts seat holds only from browser sessions (CSRF + session id), so scripted holds are rejected.
Proposed partial credit
50% visibility constraint; 30% cheapest; 20% adjacency + stop.
R12.04.01.03-T20 Book a school group's 28 seats across a cinema with chaperone placement rules
Original instruction
On the cinema group-booking clone, book 28 seats for 24 students and 4 chaperones for the 10:00 Wednesday showing: students in contiguous blocks of at most 6 per row, every student block must have a chaperone seated in the same row adjacent to it, no seats in the first 3 rows. Stop at payment.
- Platforms
web· Any OS (web)- Applications
- cinema group booking clone
- Interfaces
brow· Browser- GUI need · task setting
G3· GUI-only- GUI need · ordinary access
G2· GUI-essential- Interaction primitives
vis· Visual judgmentclick· Point & clickgeom· Precise geometry- Work product
WP08· App/system state- Conditions
risk· Irreversible/riskylong· Long-horizon- Proposed difficulty
- 5 / 5
Proposed verifier
Predicate checker on 28 held seats with role labels: block sizes <= 6, each block adjacent to a chaperone in the same row, rows >= 4; all seats held in one order. The clone issues no API token and accepts seat holds only from the browser session (CSRF + session id).
Proposed partial credit
30% seat count/rows; 40% block rule; 30% chaperone adjacency.
Mapped research references
Existing benchmark-to-class links from the research catalog. They are not record-level annotations, executions or product endorsements.