Flight profile · American Airlines

AA2903

Washington (DCA) to Chicago (ORD) · scheduled 6:00 AM – 7:24 AM

This page describes how AA2903 has historically performed, which is a different and more useful question when you are deciding what to book. The live position below is a reading from receivers on the ground, not flight status. For today’s status, check with American Airlines.

How AA2903 performs

879 operations, July 2023 – July 2026.

On-time arrivals
84.1%
704 of 837 completed
Cancellation rate
4.66%
41 cancelled
Median arrival delay
-15 min
Average 5.7 min
Severe delays
8.0%
Arrived 60+ minutes late
Delay when late
132 min
Average, when 15+ min late
Departure punctuality
83.2%
Left the gate on time

AA2903 arrived on time 84.1% of the time, against 79.0% for all DCA → ORD flights, so this departure is 5.1 percentage points better than the route as a whole. Against American Airlines’ own flights on this route it is 5.7 points better.

Source: US DOT Bureau of Transportation Statistics · 879 operations, July 2023 – July 2026 · If AA2903 is cancelled or significantly changed, DOT’s automatic refund rule may entitle you to your money back. The refund checker walks through eligibility.

Where AA2903 is right now

A live reading, only if you ask for it. It is not part of the reliability record above.

Ask for the position of AA2903 right now, if receivers on the ground are hearing it. Nothing is requested until you press the button.

Against the rest of the route

Against the rest of the route
Flight or routeOn timeCancelledMedian delaySevereFlights
AA290384.1%4.66%-15 min8.0%879
American Airlines on DCA → ORD78.4%2.57%-11 min9.7%7,680
All DCA → ORD flights79.0%2.47%-8 min9.0%20,170

AA2903 has cancelled 4.66% of its scheduled departures (41 of 879), against 2.47% for all DCA → ORD flights — this specific departure cancels 2.2 percentage points more often than the route as a whole. Different flight numbers on the same route can carry different aircraft, crews and schedules, so this is a fact about AA2903 specifically, not about the route it flies.

Source: FlyerIntel analysis of US DOT BTS data · July 2023 – July 2026

Why AA2903 runs late

Where US DOT attributed the delay minutes on this flight.

Share of the delay minutes US DOT attributed to each cause for AA2903 on DCA → ORD. Only delays of 15 minutes or more are attributed.

Source: US DOT Bureau of Transportation Statistics · Data covering July 2023 – July 2026

Does a late aircraft delay this flight?

How often the plane that flies AA2903 comes in late, and what that does to the departure.

Inbound aircraft late
32%
254 of 790 departures came in 15+ minutes late
Scheduled time on the ground
6.8 hours
Median at DCA between the inbound and this departure
Left 15+ late after a late inbound
28%
254 departures
Left 15+ late after an on-time inbound
10%
536 departures

AA2903 is usually the aircraft’s first flight of the day: the schedule gives it a median of 6.8 hours on the ground at DCA, so its “inbound” is mostly the aircraft’s arrival the evening before.

When the aircraft came in 15 or more minutes late, AA2903 pushed back an average of 26 minutes late; when it came in on time or early, an average of 4 minutes late.

Each departure is matched to the flight the same aircraft flew into DCA just before it, using the tail number in the DOT file, 790 of 838 departures in all. A departure is left out when that flight is unknown, was cancelled or diverted, or the aircraft had sat for more than 12 hours. A flight the airline cancelled never left and is not in the count, and one flown by a substitute aircraft is matched to the aircraft that flew it, not to the one first planned.

Against the network at the same turn length

AA2903 against all flights with a scheduled turn of 6 hours or more, usually overnight
Share of departuresInbound landed 15+ lateLeft 15+ late after a late inboundLeft 15+ late after an on-time inbound
AA290332%28%10%
Network, same turn29%12%9%

A night on the ground absorbs most of a late arrival, so a late inbound matters far less on a turn this long than on a same-day one.

The network row is every flight with a scheduled turn of 6 hours or more, usually overnight. How often AA2903’s aircraft landed 15 or more minutes late is higher than the network’s and, once the aircraft was late, how often AA2903 left 15 or more minutes late is higher than the network’s.

More: Which Flight Numbers Pay for a Late Aircraft lists the flight numbers that pass a late aircraft on.

Source: US DOT Bureau of Transportation Statistics · Data covering July 2023 – July 2026

From pushback to gate

When it leaves DCA late, does that still matter by the time it reaches ORD?

Leaves the gate
16.8 minutes late
Average over 838 pushbacks
Taxi out at DCA
18.0 min
17.9 min for the airport in this hour
Taxi in at ORD
14.0 min
17.8 min for a typical ORD arrival
Reaches the gate at ORD
5.7 minutes late
Average over 837 arrivals

AA2903 pushes back an average of 16.8 minutes late and reaches the gate at ORD an average of 5.7 minutes late. The 11.1 minutes between those two figures are slack the published schedule already carries, which is why a late pushback on this flight does not reliably become a late arrival. Its 18.0 minutes on the taxiways at DCA is about what any departure from the airport in that hour spends.

On arrival at ORD, AA2903 spends 14.0 min taxiing from touchdown to the gate — ground time a connecting passenger sits through before they can even leave the aircraft, on top of however late the flight itself ran. That is 3.8 minutes shorter than a typical ORD arrival (17.8 min).

Departure delay is averaged over every pushback and arrival delay over every completed arrival, so the two counts differ by the flights that never reached ORD. Taxi-out is gate to wheels up and sits inside the scheduled block rather than on top of it; the comparison is 31,497 departures from DCA scheduled in the clock hour AA2903 most often leaves in. Taxi-in is landing to gate, compared against every arrival FlyerIntel has on file at ORD (952,666 flights), not just AA2903’s own. More on taxi time.

Source: FlyerIntel analysis of US DOT BTS data · July 2023 – July 2026

The hour it leaves DCA

Part of any flight’s record belongs to the clock rather than to the flight. AA2903 has one departure time, so this is what that time is worth.

AA290384.1%
All DCA departures, 6 AM–7 AM90.8%
All DCA departures, whole day76.5%

On-time arrival rate. 31,497 departures from DCA are scheduled in the 6 AM–7 AM hour over the period, against 435,076 across the whole day.

AA2903 is scheduled out of DCA at 6:00 AM, in the 6 AM–7 AM bank. Departures from DCA in that hour arrive on time 90.8% of the time against 76.5% across the airport’s whole day — the slot alone is worth 14.3 points better before anything about this particular flight is considered.

The two airport figures cover every departure from DCA to every destination, so they describe the slot rather than this journey — the like-for-like comparison against the route is above. Hours are scheduled local, and a number whose schedule moved between banks inside the period has no single hour and is not described here at all. Why the hour matters.

Source: FlyerIntel analysis of US DOT BTS data · July 2023 – July 2026 · Hours below 120 departures are not scored.

By day of week

Does AA2903 run worse on particular days?

Monday81.8%
Tuesday88.1%
Wednesday79.8%
Thursday87.3%
Friday78.6%
Saturday87.4%
Sunday85.8%

Tuesday is historically the most reliable day and Friday the least, a gap of 9.5 percentage points.

Source: FlyerIntel analysis of US DOT BTS data · July 2023 – July 2026 · Days with fewer than 120 operations are not scored.

Other routes AA2903 has flown

A flight number is a slot in a schedule, and carriers reassign it. These are the other city pairs this number served in the analysis period.

Other routes AA2903 has flown
RouteOperatedFlightsOn time
JAX→DCA Jacksonville to WashingtonDecember 18, 2025 – January 5, 202619—
LAX→CLT Los Angeles to CharlotteDecember 7–14, 20252—
ORD→DCA Chicago to WashingtonApril 4, 2024 – November 4, 202421086.5%

Because the same number has served different journeys, the statistics above describe only the route AA2903 currently operates. Blending them would produce an average of two different flights.

When it goes wrong

The average describes a usual day. This is the day that ruins a connection.

One arrival in ten is worse than

38 min late

against a median of 15 min early

Arrived an hour or more late8.0%
Typical delay when late132 min
Cancelled4.66%
Landed somewhere else0.11%

Percentiles over 879 operations; cancelled and diverted flights are counted separately.

Source: US DOT Bureau of Transportation Statistics · Data covering July 2023 – July 2026

Other flights on this route

Same airports, same journey, different odds of arriving on time.

Other flights on this route
FlightAirlineOn timeMedian arrivalOperations
UA1793United Airlines93.6%12 min early845
AA2304American Airlines89.4%15 min early952
AA2903 this flightAmerican Airlines84.1%15 min early879
AA1200American Airlines78.0%11 min early663
UA2148United Airlines56.0%9 min late787

Every departure listed carries at least 600 operations, so the comparison is between figures that mean something. AA2903 is placed in the same ordering rather than left out of it; the rows around it are the most punctual departures on the route, not all of them, so this is a comparison and not a rank. Schedules change; confirm the departure still operates before booking around this.

Data and methodology

Calculated by FlyerIntel from 879 records of AA2903 operating DCA → ORD, published by the US DOT Bureau of Transportation Statistics. A flight counts as delayed when it arrives 15 or more minutes after schedule. Cancelled and diverted flights are excluded from the on-time rate and reported separately. Flight numbers are resolved as carrier plus number plus route, because the same number serves different routes over time.

Full methodology · Data sources · Report a problem

Related research: Most and Least Reliable Flight Numbers · The Delay You Inherit · Making Up Time in the Air

Analysis last recomputed