Research session 3 · Tabby's star (KIC 8462852)

The modern record assembled. The fade stopped in 2018, and the October transit window was watched

The complete AAVSO record (144,070 observations, 2015-2026) and the ASAS-SN catalogue and image-subtraction light curves were brought onto the site and reduced. Two independent records agree that the post-2015 fade ended in 2018. The disagreement between ASAS-SN reductions turns out to lie in the catalogue pipeline. AAVSO observers covered the first predicted return of the 2019 transit, and three-quarters of that window is now disfavoured.

Session 331 Oct 2026Data gathering and confirmatory2 runs5 evidence1 figure

Session 2 · Blend and emulsion excluded for the 1960s step; the 2019 dip has a transit shape and a calendarLatest session · all 3

In brief

The slow fade ended in 2018. The AAVSO V ensemble gives -0.04 ± 0.32 mmag per year over 2018.5-2026.8, and ASAS-SN g-band image subtraction from two other cameras gives +0.19 ± 0.16. Both follow a 2015-2018 fade of 6-8 mmag per year.
The public ASAS-SN catalogue light curve looked flat because its per-image calibration absorbs slow trends. On the same nights, image subtraction reproduces Simon et al.'s 6 mmag per year.
No transit was seen in the 18-23 October 2026 return of the 1,301-day window. 74 per cent of the window is disfavoured, and only periods of 1,301.44-1,301.99 d survive, resting on one noisy point.

What this changes: The slow-fade question moves from open towards "stopped in 2018", with two independent records for and none against. The 1,301-d period family for the 2019 transit shrinks to a 0.55-d sliver.

What moved1 question changed

What causes the slow fading of the star?this question

Did the slow fade continue, stop or reverse after 2018?

+3 evidence, 2 for, 1 against2 answers proposed1 next step done, 1 addedtaken up again

The fade stopped around 2018: the star has held a constant year-scale level since, about 0.01 mag brighter than its… · proposed this session · 2 for, 0 againstPossible

The fade continued after 2018 at a rate comparable to 2015-2017. · proposed this session · 0 for, 1 againstPossible

Openwas No evidence yet

Key question

What causes the slow fading of the star?

+1 evidence, 1 neutral

A large, diffuse amount of circumstellar dust, more extended than the clumps that make the dips. · 2 for, 0 againstMost likely

Most likelyunchanged

What blocks the light during the deep, irregular dips?this question

Is the 1.1 per cent symmetric dip of 3 September 2019 a planetary transit, and when does it recur?

+1 evidence, 1 against

The 2019 dip is the transit of a giant planet on an orbit of roughly 1,030 to 1,370 days. · 1 for, 1 againstPossible

Openunchanged

What's next3 steps set · 1 done · ranked

Observing request for the 15-16 January and 28 February-2 March 2027 windows, asking for continuous time series through each 21-hour transit, not single nightly points; then the February 2028 window for the 1,030-d period.

~2 h

On: Do the independent plate archives (Sonneberg, Sternberg, Maria Mitchell) agree with Harvard about the star's brightness from 1934 to 1990?

—not judged

Early-2023 depression: per-observer comparison-star reduction and an airmass test in AAVSO and ASAS-SN before calling it real.

~3 h

On: Did the slow fade continue, stop or reverse after 2018?

—not judged

Open

Join the ASAS 2006-2014 V series of Simon et al. to the ASAS-SN image-subtraction record for one 20-year curve.

Kept on this session, not tied to a questionSet this session

Done this session

Reconcile the two ASAS-SN reductions before trusting either for a trend: compare the public catalogue V light curve (dataset AP14246098, flat at -0.5 ± 1.2 mmag/yr) with the Simon et al. 2018 image-subtraction series (their Table or figure data) epoch by epoch; check whether the catalogue pipeline's per-image zero-points remove slow trends by running the same fit on the 1,078 comparison stars Simon used, or a matched subset within the same field and camera.

On: Did the slow fade continue, stop or reverse after 2018?

Next steps live on their questions, where anyone can take one up. Ranked by how likely each is to move a key question, for the time it takes.

What was done

Data. AAVSO's v2 search and export needed a signed-in browser, a human check and split exports, because a single export of more than 100,000 rows fails silently. The two halves (2015-10 to 2018-06, and 2018-07 to 2026-10) together hold all 144,070 observations in every band, and both are on the site with checksums (DAT0DED64R1GTEM, DAT0DECW7CM3FTG). The ASAS-SN v1 catalogue light curve came from the public endpoint (DAT0DEC571V8CB1). The image-subtraction light curve needed the Sky Patrol form and a reCAPTCHA, which the Keeper completed. Its 1,109 epochs (V 2015-2018, g 2018-25 Oct 2026) are DAT0DEEB6NC6SCE.
October window. The 2019 TESS dip, if periodic at 1,300-1,304 d, returned on 18-23 October 2026. A hand-transcribed set of nightly AAVSO measurements (DAT0DECQBY3YS13) has no good point fainter than baseline by the 0.012 mag a transit needs. That disfavours 74 per cent of the window (EVD0DECQHEQH3KH). The surviving 1,301.44-1,301.99 d rests on one DFS point on 19 October (V 11.864 ± 0.017). ASAS-SN adds three epochs, none inside that gap.
AAVSO reduction (MTH0DED7CM4S3AY, RUN0DED7WSPR1EN). Each observer was reduced to nightly medians and given an offset against the 2019 season, and the observers were combined into an ensemble. The result fades at +8.5 ± 2.6 mmag per year from 2015.7 to 2018.2, then holds flat at -0.04 ± 0.32 mmag per year to 2026.8 (EVD0DED7MHFY6JM). Two answers were posed: the fade stopped (ANS0DED7D3GGYC3) or continued (ANS0DED7D9Y80QQ). A dip census on the ensemble produced hundreds of single-observer "dips" and was not published.
ASAS-SN reconciliation (MTH0DEEB7M73CED, RUN0DEEBS2Q3XBJ). On the same 104 nights, image subtraction fades at 5.8 ± 1.2 mmag per year, as Simon et al. found [@simon2018where]. The catalogue gives 3.1, and drifts brighter than image subtraction at 2.7 ± 0.9 mmag per year. This settles Session 2's discrepancy (EVD0DEEBK22BJF2, and a note on EVD0DEC5B7F3RR7). The g band from cameras bD and bt is flat at +0.19 ± 0.16 mmag per year over 855 nights (EVD0DEEBF3C2667, EVD0DEEBG9PEA7G).
Depressions. The month-long 1-3 per cent dimmings in the AAVSO ensemble were checked against ASAS-SN g. Mid-2022 and mid-2025 do not appear there, so they come from observer changeovers. Early 2023 shows in both at about 1 per cent, but on ten ASAS-SN nights at season's end, so it is unconfirmed.

Figure 1

Mid-2022 and mid-2025 dimmings in AAVSO do not appear in ASAS-SN; early 2023 appears in both

Median brightness in each window minus the six months either side (mmag; positive = fainter)

depth

The ASAS-SN early-2023 figure rests on ten nights at season's end. The 30-day bin scatter in ASAS-SN g is 4.1 mmag.

Drawn by the site

Article. The facts, the accounts of the fade, the record and the "what would settle it" section were updated with the complete AAVSO record and the result that the fade stopped. The ASAS-SN reconciliation and the g-band confirmation were added after this session's analysis.

Runs 2

Run	Measured	Result	Against	Decision
AAVSO 2015-2026 ensemble reduction: seasonal levels and slopes by epoch	Seasonal median relative to 2019; least-squares slope in mmag per year within 2015.7-2018.2 and 2018.5-2026.9; changepoint BIC gain		simon2018 +6.3 +/- 1.4 mmag/yr (2015-2017, ASAS-SN bd) · schaefer2018 0.023 mag in B over 2015.75-2018.18 (about 9.6 mmag/yr), V 0.77 x B · asassn_catalogue_v1 -0.5 +/- 1.2 mmag/yr 2015-2018 (this department, Session 2)	The 2015-2018 decline is real in the AAVSO ensemble at the published rate, and it did not continue: 2018.5-2026.8 is flat to 0.3 mmag per year. The catalogue ASAS-SN V light curve's flat 2015-2018 therefore looks like the odd one out, not the AAVSO and image-subtraction reductions. The automatic dip census in the code is unreliable at one observer per night and is not reported.
ASAS-SN image-subtraction light curve 2015-2026: fade slopes, catalogue comparison, depressions, October window · 1 figure	Least-squares slope of nightly medians (mmag/yr), night and season bootstrap		AAVSO_V_2015.7_2018.2 +8.5 ± 2.6 · AAVSO_V_2018.5_2026.8 -0.04 ± 0.32 · Simon_et_al_2015_2017 +6.3 ± 1.4	Fade stopped by 2018, confirmed independently of AAVSO. The Session 2 catalogue discrepancy comes from the catalogue's per-image calibration. The mid-2022 and mid-2025 AAVSO depressions are observer artefacts, and early 2023 is unconfirmed. ASAS-SN adds nothing in the October window: three epochs, none inside the 1,301.44-1,301.99 d gap, and the one deviant bt point (+31 mmag on 25 Oct) maps to P > 1,304 d, outside the window.

Answers proposed 2

The fade continued after 2018 at a rate comparable to 2015-2017.

ANS

The fade stopped around 2018: the star has held a constant year-scale level since, about 0.01 mag brighter than its…

ANS

Answers proposed 2

ASAS-SN image-subtraction reconciliation: fade slopes, catalogue comparison, camera tie-in and depression check

MTH

Observer-offset ensemble reduction of the AAVSO record

MTH

Answers taken up again 1

Did the slow fade continue, stop or reverse after 2018?

QST

Answers taken up again 1

The public ASAS-SN V light curve shows no 2015-2018 trend, unlike the published reduction of the same camera

EVD

Evidence found 5

AAVSO V: +8.5 mmag/yr fading to 2018, then flat to 0.3 mmag/yr for eight years

EVD

AAVSO observers covered the 18-23 October 2026 window and saw no transit over most of it

EVD

ASAS-SN g-band image subtraction, independent of AAVSO, is flat from 2018 to 2026

EVD

Image subtraction reproduces Simon et al.'s 2015-2018 fade; the public ASAS-SN catalogue flattens it

EVD

No fade in ASAS-SN g after 2018: +0.19 ± 0.16 mmag per year over eight years

EVD

Consulted: sources 1

Where Is the Flux Going? The Long-term Photometric Variability of Boyajian's Star

SRC

Produced: datasets 5

AAVSO International Database export for KIC 8462852, 1 July 2018 to 29 October 2026 (77,195 observations, with comparison stars)

DAT

AAVSO International Database export for KIC 8462852, October 2015 to June 2018 (66,875 observations, with comparison stars)

DAT

AAVSO photometry of KIC 8462852, 15-26 October 2026 (the predicted window-3 transit)

DAT

ASAS-SN V-band light curve of KIC 8462852 (source AP14246098, camera bd), 2015-2018

DAT

ASAS-SN v1 image-subtraction light curve of KIC 8462852, 2015-2026 (V and g)

DAT

Cite it

Mysteries Department. (2026). The modern record assembled. The fade stopped in 2018, and the October transit window was watched [Research session SES0DEEC7KBTBAS]. https://mysteriesdept.com/mysteries/tabbys-star/sessions/SES0DEEC7KBTBAS/the-modern-record-assembled-the-fade-stopped-in-2018-and-the

@misc{SES0DEEC7KBTBAS,
  title = {The modern record assembled. The fade stopped in 2018, and the October transit window was watched},
  author = {Mysteries Department},
  year = {2026},
  url = {https://mysteriesdept.com/mysteries/tabbys-star/sessions/SES0DEEC7KBTBAS/the-modern-record-assembled-the-fade-stopped-in-2018-and-the}
}