Find an exoplanet
in real starlight.
A star dims. Then it dims again. Work with actual telescope measurements to measure the repeating pattern, estimate an orbital period and ask how much a brightness dip really tells us.
This is a guided reanalysis of TrES-2 b, an already known planet. It is not a planet-discovery submission or live telescope feed. The dots below are processed observations, not a simulated transit.
ACTUAL KEPLER OBSERVATIONS / TrES-2
Measure the dip. Test the rhythm.
1,410 observations · 14 recurring dips · about 29.4 minutes per cadence
Loading the controls. The observed dip and worked example remain readable below.
Inspect one observed dip
Amber uncertainty bars are enlarged 20× for visibility; brightness values are unchanged.
Mint: fixed ±0.5-hour measurement window. Gray: baseline 2.5–5.5 hours either side. 2 central / 11 baseline samples.
Bring recurring dips together
13 orbital intervalsYour timing estimate: 2.47075 days
Your goal: bring the repeated dips together into one clear V-shaped dip near the center.
Starts at 2.40000 days. Select two dips and press Stack selected dips to apply their timing estimate.
What is stacking, and why do the dots move?
Imagine a clock that resets after each proposed orbit. Instead of showing all 33 days end to end, the chart places every observation according to where it falls within one trial orbit. Observations from different orbits then appear together on the same horizontal scale. This is called phase folding; here, stacking means an overlay of the individual dots, not an average of their brightness.
The slider changes the clock’s reset interval. A wrong period places successive transits at different positions, spreading the low-brightness dots sideways. A suitable period brings those repeated events together near zero hours. Each dot keeps its measured brightness: the slider changes its horizontal phase position, not what Kepler recorded.
For this example, look for the V-shaped dip near the center. With a suitable period, this record’s low-brightness dots line up into one clear, roughly V-shaped dip, while most other dots remain near the normal brightness level. Compare that setting with the estimate from separate dip times and the published period. Transit shapes vary between systems because geometry, stellar brightness across the disk and sampling differ; the V is the useful visual guide for this particular record.
A matching pattern checks the same observations in a different arrangement. It does not create additional data or, by itself, rule out an eclipsing binary or another false positive. Open the published-interpretation reveal below the workspace to see the additional evidence for this known planet.
The 14 dips in this curated record are consecutive, so the interval count is B’s dip number minus A’s. Period = (time B − time A) / intervals. Folding rearranges the same observations; it is not independent confirmation.
Thirty-three days of starlight
Dots are retained Kepler observations from May–June 2009, not a simulation. Gold marks the inspected dip. Gaps are not filled. Brightness uses the retained quarter median; the axis begins at 98.3%, not zero.
Reveal the published interpretation
Yes—a planet is passing in front of its star.
If you predicted that a planet crosses its host star at regular orbital intervals, you got the main interpretation right. The system is TrES-2, and the transiting planet is TrES-2 b, a confirmed gas giant outside our Solar System. Each transit blocks a small fraction of the star’s light; it is a planet crossing a star, not one planet crossing another.
The discovery paper reported an orbital period of 2.47063 days—about 59.30 hours. Your estimate from this short Kepler record should be close, rather than exactly identical: the supplied dip centers are rough estimates and each cadence averages about 29.4 minutes of light.
Why astronomers accepted the planet interpretation: the repeating brightness dips were supported by spectroscopic checks and radial-velocity measurements of the star’s motion. Those measurements matched the timing expected from an orbiting planetary companion and helped test competing explanations. A regularly repeating dip alone would not establish that every unknown object is a planet.
TrES-2 b was discovered in 2006. The Kepler observations shown here were collected in 2009, so this activity revisits an established system; it does not discover or independently confirm the planet.
Read the published TrES-2 discovery and confirmation evidence →
Keep the limit: the radius ratio is only √(fractional light deficit). Stellar limb darkening, near-edge transit geometry and cadence averaging matter. This is a guided reanalysis of a known planet host, not a discovery or fitted planet radius. Read the data and method ↓
GUIDED INVESTIGATION / SUGGESTED 20–30 MINUTES
Make a claim you can check.
- Predict before measuring. Choose Full record and find several dips. Would a single chance fluctuation normally explain a series with similar spacing?
- Inspect the sampling. Choose Inspect dip, then Dip 01. Move the center through the dip and describe how the inferred depth changes as the fixed measurement window includes ingress, egress or out-of-transit light. Show uncertainties to inspect the enlarged pipeline-error bars.
- Measure a period. Choose Stack transits. Select Dip 01 as A and Dip 14 as B; the tool counts their 13 intervals automatically. Press Stack selected dips. Repeat with two consecutive dips and compare the period estimates. The supplied centers are rough navigation estimates, not fitted transit times.
- Test the same data a different way. Move the trial-period slider to compare a mismatched period with your timing estimate. The recurring dips should collect near the same phase at a suitable period. This is a consistency check on the same data, not independent confirmation.
- Keep the limit. Write a claim, two measured clues, a limitation and a useful additional observation. Consider a blended eclipsing binary, stellar variability or an instrumental problem before interpreting every dip as a planet. Then open Reveal the published interpretation beneath the chart to compare your prediction with the established result.
Compare with a worked example
The opening navigation center and final center give (164.76190 − 132.64218) / 13 ≈ 2.47075 days, close to the 2006 published 2.47063-day period. These centers are rough deficit-weighted navigation estimates, not fitted mid-transit times.
At the initial window, the local median deficit is about 1.34%, giving √0.0134 ≈ 0.116 as an area-based radius ratio. The discovery paper reported approximately 1.4% deep transits in its own bands; that is a comparison, not an exact prediction for Kepler or this estimator.
A defensible claim is that this record contains recurring transit-like dimming consistent with the already established TrES-2 b system. The activity alone does not establish planetary mass, atmospheric composition, habitability or a previously unknown planet.
For independent work, download the CSV, plot time against normalized flux and repeat the arithmetic in a spreadsheet. The initial graphs and worked example remain readable without JavaScript. No account, submission or purchase is required.
What turns a dip into a planet interpretation?
What you measured
Brightness versus time, a recurring spacing and a local depth estimate in one Kepler band. A light curve is not a resolved picture of the planet or a measurement of its surface color.
What needs more evidence
A stellar companion or blended eclipsing binary can mimic a transit. Astronomers examine imaging, image-centroid changes, odd/even events, spectra and radial velocities as appropriate. TrES-2’s discovery paper includes spectroscopy and radial-velocity evidence; this workbench does not reproduce those tests.
Why the size is approximate
A star is not uniformly bright from center to edge. Geometry, contaminating light and integration time affect the observed depth and shape. An absolute planetary radius also needs the stellar radius. We deliberately report an approximate ratio instead of assuming a star size.
DATA PROVENANCE / REPRODUCIBLE PROCESSING
From the archive file to these dots.
This page includes NASA Kepler mission data archived by MAST at the Space Telescope Science Institute, retrieved through its public data mirror on . The Last Transmission prepared the teaching subset and this interface; we did not operate the telescope or discover TrES-2 b. The public mirror lists the Kepler data as public domain.
- Original product:
kplr011446443-2009166043257_llc.fits, KIC 11446443, Quarter 1, long cadence, Data Release 25. Its header records May 13–June 15, 2009. - Use the archive’s corrected photometry: TIME, PDCSAP_FLUX and PDCSAP_FLUX_ERR. PDCSAP already applies mission pipeline processing, including systematic-error and aperture/crowding treatment. It is not raw detector output.
- Filter conservatively: of 1639 rows, omit 15 with missing/nonfinite or nonpositive required values, then 214 with nonzero SAP_QUALITY. Keep 1410 rows. Quality zero does not guarantee freedom from every artifact, and this mask may discard usable observations.
- Normalize without reshaping: divide flux and its supplied uncertainty by the retained quarter median, 403,696.25 electrons per second. No further detrending, smoothing, interpolation, gap filling or synthetic transit injection is applied. Display values are rounded to nine decimals.
- Navigate and measure: group retained samples below 99.6% of the quarter median, starting another group after a gap over 0.2 day; use each group’s deficit-weighted time as a rough center. The 14 groups are only navigation aids. Depth uses the median within a fixed half-width of 0.5 hour around the inspection center and the median 2.5–5.5 hours either side. If there are fewer than two central or six baseline samples, no estimate is reported.
- Keep units and uncertainty: BKJD is BJD on the TDB time system minus 2,454,833, in days; it is not a local calendar date. Error bars show pipeline flux uncertainties after normalization. At actual scale they are smaller than the plotted dots. Show uncertainties enlarges only the bars by 20×, explicitly labeled beside the plot; it does not change the numbers or dot positions. They omit baseline uncertainty, correlations and other systematics; the page supplies no formal fit errors or detection significance.
File integrity and reuse
Original FITS SHA-256: 10bfca91bdd928deda2e638de7a27cdc940b879452408a96b5be6146944639e9. All three original HDU checksums were verified. The local FITS download is byte-identical to the retrieved product.
The CSV retains the filtered time, original pipeline flux/error, normalized values, cadence number and quality flag. Original Kepler data are public-domain observations; acknowledge NASA Kepler and MAST/STScI when reusing them, and retain the processing notes when sharing this subset. This teaching page is not an official NASA or STScI product or endorsement.
Accessible observation table: selected inspection window
The table lists measurements within six hours either side of the selected center. The download contains all retained observations. Normalized uncertainties are the pipeline flux errors divided by the quarter median.
| BKJD / days | Normalized flux | Normalized uncertainty | Cadence |
|---|---|---|---|
| 132.411414293 | 1.000042575 | 0.000042171 | 1149 |
| 132.431848523 | 0.999942175 | 0.000042168 | 1150 |
| 132.472717084 | 1.000030267 | 0.000042180 | 1152 |
| 132.493151214 | 0.999921507 | 0.000042151 | 1153 |
| 132.513585444 | 1.000017959 | 0.000042159 | 1154 |
| 132.534019574 | 0.999985447 | 0.000042171 | 1155 |
| 132.554453904 | 0.999961063 | 0.000042148 | 1156 |
| 132.574888134 | 0.999971049 | 0.000042147 | 1157 |
| 132.615756594 | 0.995842326 | 0.000042109 | 1159 |
| 132.636190824 | 0.986391842 | 0.000041901 | 1160 |
| 132.656625053 | 0.986749087 | 0.000041913 | 1161 |
| 132.677059183 | 0.996610385 | 0.000042085 | 1162 |
| 132.697493512 | 1.000070985 | 0.000042177 | 1163 |
| 132.717927742 | 0.999979254 | 0.000042181 | 1164 |
| 132.738361871 | 1.000003870 | 0.000042169 | 1165 |
| 132.758796200 | 0.999953399 | 0.000042171 | 1166 |
| 132.779230430 | 1.000087860 | 0.000042152 | 1167 |
| 132.799664559 | 1.000022758 | 0.000042149 | 1168 |
| 132.820098788 | 0.999926306 | 0.000042154 | 1169 |
| 132.840533117 | 0.999986376 | 0.000042158 | 1170 |
| 132.860967246 | 0.999979099 | 0.000042163 | 1171 |
| 132.881401475 | 0.999887988 | 0.000042161 | 1172 |
CHECK THE RECORD
Follow the measurements to their sources.
Kepler archive and data products ↗
NASA Kepler observations archived at MAST/STScI. The original Quarter 1 long-cadence FITS file identifies KIC 11446443 and Data Release 25.
MAST: interpreting Kepler light curves ↗
Explains time units, long cadence, simple-aperture and pipeline-corrected PDCSAP flux. These observations have already been processed by the mission pipeline.
Kepler Archive Manual ↗
Documents light-curve columns, time system, quality flags and pipeline corrections. Our quality-zero subset is a conservative teaching choice, not a universal recommended mask.
The TrES-2 discovery paper ↗
O’Donovan and colleagues (2006) report a 2.47063-day orbit, approximately 1.4% deep transits in their observing bands, and supporting radial-velocity and spectroscopic tests. These are historical published values, not an up-to-date catalog fit.
NASA: what is a transit? ↗
Explains why a planet crossing a star can reduce the light received by an observer. A light curve records brightness over time, not an image of the planet.
NASA: Exoplanet Science Strategy ↗
Chapter 2 explains the area-ratio approximation and its assumptions: a complete transit and uniform stellar brightness. Real analysis accounts for limb darkening and geometry.
Public Kepler data mirror ↗
STScI-managed public Kepler files. This activity retrieved the original archive product from this mirror on October 24, 2026. It is stored locally here for reproducibility.
Another measurement. Another question.
A transit tells us about changing brightness. A spectrum can constrain a star’s properties or reveal radial motion; it does not make a transit interpretation automatic.