Production proposal. Every number on this page is a
genuine output of the shipped models on real ZTF objects — including one the
detector misses. What a live deployment would add is throughput, not different
numbers.
ZTF19acbzgog
AT2019rsi · 00:42:23.53 +41:11:52.2 · 30 detections
over 76 d · peak 17.91
Anomaly
Anomaly. Energy clears the 1% false-alarm budget and the prediction set excludes the argmax class.
Light curve
Classification 6 classes · 190 features
SN Ia50%p 0.024
SN CC44%p 0.143
SLSN1%p 0.125
AGN0%p 0.053
TDE1%p 0.083
CV4%p 0.085
{SN CC, SLSN}
90% ± 10%coverage the set is calibrated for — it may
miss the true class one time in ten, by construction
measured out of fold, SN Ia
90.3% ± 0.3%
n = 7159
credibility does it match any class at all
0.143
confidence is the runner-up excluded
0.875
The most probable class is not in the set: its p-value is below the 10% error level, so the classifier is contradicting itself.
Anomaly own 181-column space
30.6%(24–42%)
probability this is outside the six-class taxonomy, at the
benchmark's 3.1% base rate
likelihood ratio carries no prior — rescale
for your stream
13.6×
novelty p-value 1 known-class object in
315 is this extreme
0.0032
energy
+2.06
percentile against the known-class stream99.7
Evidence context at alert time, no spectrum
Gaia point source
none
nothing persistent at the position above the Gaia limit
WISE W1−W2
-0.55
no mid-infrared AGN excess
Pre-outburst counterpart
none
nothing at this position in Pan-STARRS before the transient
Host offset
2.13″
displaced from its host
Photometric redshift
none
no distance proxy available
ZTF21abrqria
SN2021vuw · 06:40:47.45 +27:47:01.8 · 70 detections
over 98 d · peak 18.72
Ambiguous
Ambiguous. 2 classes survive at 10% error — hand over as a candidate set, not a label.
Light curve
Classification 6 classes · 190 features
SN Ia0%p 0.002
SN CC48%p 0.158
SLSN52%p 0.583
AGN0%p 0.053
TDE0%p 0.083
CV0%p 0.021
{SN CC, SLSN}
90% ± 10%coverage the set is calibrated for — it may
miss the true class one time in ten, by construction
measured out of fold, SLSN
94.9% ± 1.8%
n = 156
credibility does it match any class at all
0.583
confidence is the runner-up excluded
0.842
Anomaly own 181-column space
5.5%(5–6%)
probability this is outside the six-class taxonomy, at the
benchmark's 3.2% base rate
likelihood ratio carries no prior — rescale
for your stream
1.8×
novelty p-value 1 known-class object in
6 is this extreme
0.1556
energy
-0.08
percentile against the known-class stream84.5
Evidence context at alert time, no spectrum
Gaia point source
none
nothing persistent at the position above the Gaia limit
WISE W1−W2
no match
below the AllWISE limit
Pre-outburst counterpart
none
nothing at this position in Pan-STARRS before the transient
Host offset
no host
no extended source within 30″
Photometric redshift
none
no distance proxy available
ZTF19acbwccr
SN2019sms · 10:58:05.50 +40:11:19.3 · 8 detections
over 33 d · peak 18.96
Routine
Routine. Singleton set, energy in the bulk of the known-class stream.
Light curve
Classification 6 classes · 190 features
SN Ia8%p 0.005
SN CC91%p 0.382
SLSN0%p 0.042
AGN0%p 0.053
TDE0%p 0.083
CV0%p 0.021
{SN CC}
90% ± 10%coverage the set is calibrated for — it may
miss the true class one time in ten, by construction
measured out of fold, SN CC
91.9% ± 0.5%
n = 2559
credibility does it match any class at all
0.382
confidence is the runner-up excluded
0.917
Anomaly own 181-column space
3.9%(4–4%)
probability this is outside the six-class taxonomy, at the
benchmark's 3.2% base rate
likelihood ratio carries no prior — rescale
for your stream
1.2×
novelty p-value 1 known-class object in
4 is this extreme
0.2235
energy
-0.46
percentile against the known-class stream77.7
Evidence context at alert time, no spectrum
Gaia point source
none
nothing persistent at the position above the Gaia limit
WISE W1−W2
no match
below the AllWISE limit
Pre-outburst counterpart
none
nothing at this position in Pan-STARRS before the transient
Host offset
8.99″
displaced from its host
Photometric redshift
none
no distance proxy available
ZTF25abhoijc
SN2025thc · 01:15:44.09 +25:27:16.8 · 17 detections
over 39 d · peak 18.75
Routine
Routine. Singleton set, energy in the bulk of the known-class stream.
Light curve
Classification 6 classes · 190 features
SN Ia99%p 0.457
SN CC1%p 0.018
SLSN0%p 0.042
AGN0%p 0.053
TDE0%p 0.083
CV0%p 0.021
{SN Ia}
90% ± 10%coverage the set is calibrated for — it may
miss the true class one time in ten, by construction
measured out of fold, SN Ia
90.3% ± 0.3%
n = 7159
credibility does it match any class at all
0.457
confidence is the runner-up excluded
0.917
Anomaly own 181-column space
2.0%(2–2%)
probability this is outside the six-class taxonomy, at the
benchmark's 3.2% base rate
likelihood ratio carries no prior — rescale
for your stream
0.6×
novelty p-value 1 known-class object in
2 is this extreme
0.4057
energy
-1.14
percentile against the known-class stream59.5
Evidence context at alert time, no spectrum
Gaia point source
none
nothing persistent at the position above the Gaia limit
WISE W1−W2
0.12
no mid-infrared AGN excess
Pre-outburst counterpart
none
nothing at this position in Pan-STARRS before the transient
Host offset
2.19″
displaced from its host
Photometric redshift
0.081
gives Mpseudo = -19.1
ZTF22abkfhua
AT2022wtn · 23:23:23.78 +10:41:07.8 · 48 detections
over 97 d · peak 18.08
Follow-up
Follow-up. Singleton set for a class whose defining property only spectroscopy resolves.
Light curve
Classification 6 classes · 190 features
SN Ia0%p 0.001
SN CC0%p 0.016
SLSN0%p 0.042
AGN0%p 0.053
TDE99%p 0.750
CV0%p 0.021
{TDE}
90% ± 10%coverage the set is calibrated for — it may
miss the true class one time in ten, by construction
measured out of fold, TDE
92.6% ± 3.2%
n = 68
credibility does it match any class at all
0.750
confidence is the runner-up excluded
0.947
Anomaly own 181-column space
1.5%(1–2%)
probability this is outside the six-class taxonomy, at the
benchmark's 3.2% base rate
likelihood ratio carries no prior — rescale
for your stream
0.5×
novelty p-value 1 known-class object in
2 is this extreme
0.5111
energy
-1.48
percentile against the known-class stream48.9
Evidence context at alert time, no spectrum
Gaia point source
0.15″ away
a persistent source sits at the position — nuclear or stellar
WISE W1−W2
0.02
no mid-infrared AGN excess
Pre-outburst counterpart
0.18″ · r = 18.6
Pan-STARRS saw something here years before the outburst
Host offset
0.18″
on the nucleus
Photometric redshift
0.073
gives Mpseudo = -19.5
ZTF19abgcnzi
AT2019lbk · 20:48:21.69 -05:47:13.9 · 10 detections
over 34 d · peak 17.32
Routine
Routine. Singleton set, energy in the bulk of the known-class stream.
Light curve
Classification 6 classes · 190 features
SN Ia4%p 0.004
SN CC2%p 0.018
SLSN0%p 0.042
AGN0%p 0.053
TDE0%p 0.083
CV94%p 0.340
{CV}
90% ± 10%coverage the set is calibrated for — it may
miss the true class one time in ten, by construction
measured out of fold, CV
93.4% ± 1.4%
n = 304
credibility does it match any class at all
0.340
confidence is the runner-up excluded
0.917
Anomaly own 181-column space
0.4%(0–1%)
probability this is outside the six-class taxonomy, at the
benchmark's 3.2% base rate
likelihood ratio carries no prior — rescale
for your stream
0.1×
novelty p-value 1 known-class object in
1 is this extreme
0.9346
energy
-2.94
percentile against the known-class stream6.5
Evidence context at alert time, no spectrum
Gaia point source
none
nothing persistent at the position above the Gaia limit
WISE W1−W2
no match
below the AllWISE limit
Pre-outburst counterpart
none
nothing at this position in Pan-STARRS before the transient
Host offset
8.31″
displaced from its host
Photometric redshift
0.310
gives Mpseudo = -23.7
ZTF18abuwcul
AT2018hum · 02:05:39.97 -09:07:50.7 · 13 detections
over 85 d · peak 18.44
Routine
Routine. Singleton set, energy in the bulk of the known-class stream.
Light curve
Classification 6 classes · 190 features
SN Ia0%p 0.001
SN CC0%p 0.003
SLSN0%p 0.042
AGN100%p 0.605
TDE0%p 0.083
CV0%p 0.021
{AGN}
90% ± 10%coverage the set is calibrated for — it may
miss the true class one time in ten, by construction
measured out of fold, AGN
94.8% ± 1.4%
n = 251
credibility does it match any class at all
0.605
confidence is the runner-up excluded
0.917
Anomaly own 181-column space
0.2%(0–0%)
probability this is outside the six-class taxonomy, at the
benchmark's 3.2% base rate
likelihood ratio carries no prior — rescale
for your stream
0.1×
novelty p-value 1 known-class object in
1 is this extreme
0.9943
energy
-3.49
percentile against the known-class stream0.6
Evidence context at alert time, no spectrum
Gaia point source
0.11″ away
a persistent source sits at the position — nuclear or stellar
WISE W1−W2
0.99
mid-infrared colour of an accreting nucleus
Pre-outburst counterpart
0.11″ · r = 17.4
Pan-STARRS saw something here years before the outburst
Host offset
0.11″
on the nucleus
Photometric redshift
0.193
gives Mpseudo = -21.4
ZTF23aabvpgb
AT2023bji · 05:43:48.25 +62:58:41.6 · 9 detections
over 14 d · peak 16.33
Routine
Routine. Singleton set, energy in the bulk of the known-class stream.
Light curve
Classification 6 classes · 190 features
SN Ia0%p 0.001
SN CC0%p 0.005
SLSN0%p 0.042
AGN0%p 0.026
TDE0%p 0.083
CV100%p 0.936
{CV}
90% ± 10%coverage the set is calibrated for — it may
miss the true class one time in ten, by construction
measured out of fold, CV
93.4% ± 1.4%
n = 304
credibility does it match any class at all
0.936
confidence is the runner-up excluded
0.917
Anomaly own 181-column space
0.2%(0–0%)
probability this is outside the six-class taxonomy, at the
benchmark's 3.1% base rate
likelihood ratio carries no prior — rescale
for your stream
0.1×
novelty p-value 1 known-class object in
1 is this extreme
0.9841
energy
-3.34
percentile against the known-class stream1.6
Evidence context at alert time, no spectrum
Gaia point source
none
nothing persistent at the position above the Gaia limit
WISE W1−W2
no match
below the AllWISE limit
Pre-outburst counterpart
none
nothing at this position in Pan-STARRS before the transient
Host offset
21.31″
displaced from its host
Photometric redshift
none
no distance proxy available
How to read it. The prediction set is the operational
output, not the argmax: it is calibrated so that it contains the true class
90% of the time, and the 10% it admits is stated rather than hidden. A set
with two classes is the model declining to choose, which is the correct
hand-off to spectroscopy. The anomaly probability is separate machinery —
scored in its own representation, calibrated against the benchmark's
out-of-taxonomy objects — and it is quoted with the base rate it assumes,
because the same likelihood ratio gives a very different probability in a raw
alert stream than in a spectroscopically confirmed sample.