What a clay shooter's eyes are doing, measured rather than assumed.
Open, browser-based psychophysics on the questions clay shooters argue about: how early you can
call a target's line, whether a lens tint helps your eyes separate a clay from its background, which eye
does your sighting, how much lead you see, and where in your field you pick up a target late.
Each study calibrates your screen, states in advance what result would count, and gives you your own data.
It is also a lab notebook. Dr. Satwant Kumar, a physician and vision neuroscientist, took up
Olympic-style trap in 2026 as a complete beginner and is documenting the attempt with every rep measured
and every miss logged.
The trap occlusion task: a clay leaves the trench, is seen for 33 to 67 milliseconds, then hidden.
The threshold is the smallest launch angle you call correctly three times in four.
2,862
calibrated trials, every one logged
0.09°
the smallest launch angle called so far, at 67 ms of visible flight
≈ 7 cm
what the median 50 ms threshold amounts to as sideways drift at 17 m
01Five studies, one instrument
Open to anyone with a laptop, a keyboard and a plastic card the size of a driving licence.
The studies share the same approach to screen calibration and frame timing and differ in the question. Each runs in a
normal browser and needs a plastic card (a driving licence or any card that size) to calibrate the screen. Each shows you a
first result about three minutes in. The trap occlusion and lens tint studies run under a Clay & Cortex
account and then offer the full session, about 25 minutes. The other three need no account: the eye dominance profiler is
about five minutes in total, the lead perception study offers a full profile in about ten, and the pick-up map a full map in about six.
Study 1 · ISSF trap · open
Trap occlusion study
How early can you call the line of a trap target?
A clay is launched from the trench at a small angle left or right of straightaway, is visible for between 33
and 67 milliseconds, and is then occluded. You report left or right. An adaptive procedure finds the
smallest angle you can call at each viewing time, so the result is a curve of angular threshold against
visible flight, not a score.
Three viewing durations
First result in about 3 minutes
Full session about 25 minutes
60 Hz display or better
Open to anyone with a Clay & Cortex account: an email and a password, and the same
account runs every study. Trap shooters at any level, and non-shooters as a comparison group.
Which lens tint separates a clay from its background, for your eyes, on your field?
A tinted lens cannot add light. It can only remove light differently from the clay and from what is behind
it, and whether that helps depends on the clay's pigment, the background, where the sun is, and how sensitive
your own eyes are to the brightness and colour differences that result. This study measures the last of
those and tests whether a physical model predicts which lens classes work for you.
Your isoluminant point
Three cone-contrast axes
Six lenses, presented blind
First result in about 3 minutes, full session about 25
Open to anyone with a Clay & Cortex account, the same one as the other studies. You get
your profile at the end.
Study 3 · all disciplines · open · no account needed
Eye dominance profiler
Which eye does the sighting, and is it always the same one?
A five-minute sighting-dominance test: frame a target through a small gap between your hands, close one
eye, and see which eye kept it. Thirteen trials, across the screen and at two distances, give the eye, how
consistent it was, and whether it switched when the target was off to one side, which is what matters on a
trap layout and what the classic one-shot tests miss.
About five minutes
No account needed
Thirteen trials, five positions
Contribute anonymously, or not
Your result is yours either way. Contributing it, with an optional questionnaire, is a separate consented step:
anonymous, or linked to your account if you are signed in.
Study 4 · all disciplines · open · no account needed
Lead perception study
How much lead do you see?
A crossing clay has to be shot at where it will be when the pellets arrive. How much that is, is arithmetic; how
much of it a shooter sees as right is the judgement every rule of thumb in the sport is about, and nobody
measures it. A clay crosses the screen at a known speed, over and over; you set a ring where you would point the gun
at the instant you fire; the shot's time of flight says where the pattern needed to be. No feedback, by design: an
assessment, not a trainer.
Three speeds, two directions, two geometries
First result in about 3 minutes, full profile about 10
Bias and precision against the physics
No account needed
You get a lead profile a coach can read in one sentence. Contributing it, with an optional questionnaire
including how you shoot a crosser, is a separate consented step.
Study 5 · all disciplines · open · no account needed
Pick-up map
Where in your field do you pick up a target late?
A target is seen some tenths of a second after it appears, and where it appears relative to where you are looking
decides how late. You hold a look point above a cluttered horizon; a small target appears at one of five places
across and one of three heights; you press the instant you see it. The map of your latency by side and height says
which part of the field you serve late, and what your look point might do about it: the thing a coach actually changes.
Five positions, three heights, realistic clutter
Viewing distance measured, not asked
First map in about 3 minutes, full map about 6
No account needed
The look-point suggestion is a starting point for a session with a coach, not an instruction. Contributing the
record, with an optional questionnaire, is a separate consented step.
Published as they come in, anonymised, whether or not they flatter.
All five studies publish their data on the results page, regenerated from the study
database. This is the headline figure from the trap occlusion study: the smallest launch angle each person
could call, against how long the target was visible.
Across everyone measured so far, the smallest launch angle a shooter can call falls from a median of 0.48° when the target is visible for 33 ms to 0.17° at 67 ms. 2,862 calibrated trials so far, updated 2026-10-28.
Threshold against visible flight, one line per session, coloured by self-reported experience. Log scales. Hollow points sit at the floor of what the task can present, so the true value is lower.
From beginner to the Olympic trap line, documented.
At 38, I'm a physician and neuroscientist. Eight weeks ago I decided to find out how far
science, data, and 25,000 targets a year can take a beginner in Olympic-style trap shooting.
Dr. Kumar's stated target is the 2032 Olympic Games. The probability, in his own words, is low. The point
is the record: every rep measured, every miss logged, every method tested, so that whatever the outcome, the
attempt is documented properly. The studies on this site are part of that notebook. The rest of it, week by
week, is on Instagram and YouTube.
Screen scale from matching a standard plastic card (ISO/IEC 7810 ID-1, the driving-licence size) in every study;
viewing distance by the blind-spot method (Li, Joo, Yeatman & Reinecke 2020), cross-checked against what you
report, in the trap occlusion, lens tint and pick-up map studies, and stated by you in the other two; refresh rate
measured. Every stimulus is specified in degrees of visual angle and shown for a whole number of frames, and in
the two threshold studies a dropped frame rejects the trial.
Thresholds, not scores
In the trap occlusion and lens tint studies, adaptive Bayesian placement (ZEST; King-Smith et al. 1994) puts each
trial where it tells the most; thresholds are fitted by maximum likelihood with bootstrap confidence intervals,
and a threshold outside the range actually tested is reported as not measurable rather than as a number. The
other three report what they measured, an eye, a bias with its precision, a latency map, with the trial counts
behind it, and report "undetermined" when those are too few.
Gates stated in advance
Each study says, before the data come in, what result would justify the next step: test–retest
reliability before any individual claim, model-against-measurement agreement before a personal
recommendation, and a field study with real equipment before any claim about performance.
Your data are yours
At the end of a session you can download the complete record: the configuration, the random seed, your
calibration, and every stimulus and response. The server keeps the same file under your account in the two
account studies, and in the other three only if you choose to contribute it. No name is asked for; where an
account is used, the email is used to sign in and for nothing else.
Screening, not diagnosis
The colour-vision task in the tint study runs on a display whose colour is assumed, not measured. A flag
from it means "see an optometrist", nothing more, and no result here is a clinical finding.
Independent
No lens or equipment maker is involved in any of the studies. Lens names appear because people shop by name;
the ranking is computed from the physics of the lens, the clay and the light, and inclusion is not
endorsement. Any commercial link on this site is labelled as one. The rules.
05Tools, and the rules they will follow
Research on one side, products on the other.
In development
Training tools and a simulator
The studies exist to answer questions; the tools will be what those answers make possible: a personal
visual profile, a lens recommendation that uses it, and a simulator for the parts of trap that can be
trained off the range. Nothing is for sale yet. When something is, it will appear here with
the evidence behind it.
The line, written down before there is anything to sell
The person running these studies will also sell tools built on them, and the site says so. The rules that
keep that honest are written down in advance: every product carries an evidence label (model, measured, or
field-validated), no ranking is ever for sale, every affiliate link is marked, study participants are never
a mailing list, and negative results are published.
MBBS, PhD. Vision and cognitive neuroscientist, trained as a medical doctor.
He studies how the brain sees; now he is applying the laboratory's methods to clay shooting,
starting with his own.
Dr. Kumar is a physician-scientist. He trained as a physician (MBBS, the equivalent of an MD), earned his PhD
in neuroscience at KU Leuven in Belgium, and spent five years as a postdoctoral fellow at The University of
Texas at Austin's Center for Perceptual Systems, where he developed optical tools to record and control
neural activity in the visual cortex. His work on how the brain identifies bodies and separates objects from
their backgrounds has appeared in PNAS, eLife, Cerebral Cortex and the
Journal of Cognitive Neuroscience.
Today he is the founder of Asperity Industries, an AI robotics company in
Tyler, Texas, and co-director of the Dementia Neuroscience
Lab, where he works on brain imaging and clinical AI for Alzheimer's disease. He previously founded
NeuroReef Labs, a healthcare AI startup backed by Antler, and serves as volunteer Chief Scientific Officer of
the Alzheimer's Alliance of Smith County and as volunteer faculty in neurology at The University of Texas at
Tyler.
Since 2026 he has also been a trap shooter in training. Clay & Cortex is where the two
meet.
Yilmaz, H., Shah, A. D., Letrou, A., Kumar, S., Vogels, R. & Yildirim, I. (2025). Multiarea processing in
body patches of the primate inferotemporal cortex implements inverse graphics. Proceedings of the National
Academy of Sciences, 122(28), e2420287122. doi:10.1073/pnas.2420287122
Chen, C. Y., Chen, Y. Y., Benvenuti, G., Kumar, S., Ramakrishnan, C., Deisseroth, K., Geisler, W. S. &
Seidemann, E. (2022). Similar neural and perceptual masking effects of low-power optogenetic stimulation in
primate V1. eLife, 11, e68393. doi:10.7554/eLife.68393
Kumar, S., Mergen, E. & Vogels, R. (2022). It is not just the category: behavioral effects of fMRI-guided
electrical microstimulation result from a complex interplay of factors. Cerebral Cortex Communications,
3(1), tgac010. doi:10.1093/texcom/tgac010
Kumar, S. & Vogels, R. (2019). Body patches in inferior temporal cortex encode categories with different
temporal dynamics. Journal of Cognitive Neuroscience, 31(11), 1699–1709.
doi:10.1162/jocn_a_01444
Kumar, S., Popivanov, I. D. & Vogels, R. (2019). Transformation of visual representations across ventral
stream body-selective patches. Cerebral Cortex, 29(1), 215–229.
doi:10.1093/cercor/bhx320
Kalfas, I., Kumar, S. & Vogels, R. (2017). Shape selectivity of middle superior temporal sulcus body patch
neurons. eNeuro, 4(3). doi:10.1523/ENEURO.0113-17.2017
07Taking part, and contact
Voluntary, anonymous in what is published, and yours to withdraw.
What taking part means
Participation is voluntary and you can stop at any time by closing the page; nothing is uploaded until a
session ends. Each study records your responses, the stimulus that produced each one, your questionnaire
answers and your screen calibration: under your account in the trap occlusion and lens tint studies, and in
the other three only if you choose to contribute the record, anonymously or linked to your account. The results
are used for research on visual performance in clay shooting and may be published in aggregate; no individual is
identifiable in what is published.
One account covers every study; what runs under it is linked by an internal id, never by
the email. To withdraw a session or have the account removed, write to the address here from the email you
registered with.
Contact
Questions about a study, a result, or a collaboration:
Clay & Cortex is an independent research project of Dr. Satwant Kumar. It is not affiliated
with, sponsored by, or endorsed by any shooting federation, club, or equipment maker.
