Survivorship Bias in the Detection and Characterization of Interstellar Objects
by Omer Eldadi (1), Gershon Tenenbaum (1) and Avi Loeb (2)
- Department of Psychology, Reichman University, Herzliya, Israel
- Department of Astronomy, Harvard University, Cambridge, MA, USA
(Submitted for publication as a perspective article in a peer-reviewed journal)
Abstract
The three known interstellar objects (ISOs): 1I/’Oumuamua, 2I/Borisov, and 3I/ATLAS, were each detected serendipitously by surveys designed for other purposes. Hubble Space Telescope observations confirm that even 3I/ATLAS, the brightest of the three, would have escaped detection without bright enhancement from its dust coma, while corrected nucleus measurements revise the inferred number density upward by an order of magnitude. We argue that the known sample is shaped not by a selection bias in which a researcher chooses a non-representative subset from an accessible population, but by survivorship bias: objects too small, too dark, too fast, or insufficiently active are structurally excluded from detection and leave no observational trace. We estimate that existing instruments prove less than ~0.1% of the plausible ISO parameter space across four independent axes of invisibility: size, albedo, velocity, and activity. The apparent diversity of the known three objects further activates the representativeness heuristic, creating an illusion of population coverage that compounds the statistical distortion. We conclude by outlining a multi-modal detection architecture designed to find that the object’s current infrastructure is structurally incapable of detecting.
Introduction
During the Second World War, the Statistical Research Group at Columbia University was tasked with a problem of immediate operational importance: where should armor be added to allied bombers to improve their survivability? Engineers had catalogued the distribution of bullet holes on aircraft returning from combat missions and proposed reinforcing the most heavily damaged areas: the fuselage, wings, and fuel systems [1]. The mathematician Abraham Wald recognized the critical flaw in this reasoning [2]. The damage distribution they observed came exclusively from aircraft that had survived; the areas showing no damage like the engines and cockpit, were precisely those where hits were fatal, because aircraft struck have never returned. The military was studying survivors, not the full population, and thus, their observations were only partially reliable.
This insight, now known as survivorship bias [3], has become a foundational concept in statistics, epidemiology, finance, and experimental design. We argue here that interstellar object (ISO) science faces an analogous and equally consequential form of this bias. Since the first confirmed ISO, 1I/‘Oumuamua, was detected in 2017 [4,5], only two additional interstellar visitors have been identified: 2I/Borisov in 2019 [6] and 3I/ATLAS in 2025 [7]. Yet all three were detected because they were large enough, bright enough, and close enough to the Sun, and sufficiently well-placed geometrically to be captured by surveys designed primarily for near-Earth object (NEO) detection. They are the “detectable bombers”. The objects that were too small, too dark, too fast, on unfavorable trajectories, or lacking outgassing activity, remain undetected. Individual authors have noted aspects of this detection incompleteness, including evidence that previous wide-field surveys missed numerous ISOs of comparable size to those already detected [8], and that dark, non-reflective ISOs require entirely new detection modalities [9,32]. This Perspective proposes survivorship bias as a unifying framework for these detection limitations.
Survivorship Bias — Not Selection Bias
It is imperative to distinguish the bias we describe from conventional selection bias. Selection bias arises when a researcher, consciously or unconsciously, chooses a non-representative subset from an available population. For example, by preferentially studying bright galaxies in a catalog that also contains faint ones, or by recruiting only college-affiliated volunteers for a clinical trial. In such cases, the full population is in principle accessible; the distortion is introduced by the act of selection. The ISO detection problem is fundamentally different. We are not selecting a biased subset from a larger accessible catalog, but rather analyzing the entire available sample, every interstellar object ever detected, and that sample consists of three objects as of this writing. There is no drawer of neglected ISOs waiting to be included. The bias is not in our analysis of the data; it is in the data itself. Objects that were too dark, too small, too fast, or too poorly positioned were never registered by any instrument, never assigned a designation, and never entered any database. This is the defining structure of survivorship bias: the absence of the non-survivors is invisible precisely because they leave no trace. Just as Wald’s analysts could not study the bombers that never returned, because those aircraft and their crews were lost over enemy territory, we cannot study the ISOs that transited the Solar System without detection, because they left no observational residue.
This distinction carries methodological consequences. When facing selection bias, one corrects by improving sampling from a known population. When facing survivorship bias, the population itself is unknown, and correction requires expanding the conditions under which survival (here, detection), is possible. One cannot resample from a pool that does not exist, thus must build new instruments, open new wavelength windows, and develop new detection architectures so that objects which previously could not “survive” into our catalogs are finally able to do so. The solution is not better statistics applied to three objects. It is the creation of detection conditions under which the next three hundred objects include those that the current infrastructure is structurally incapable of finding.
The Three Survivors: What We Know and How We Found Them
The known ISO sample, while small, is already diverse and the circumstances of each detection are as revealing as the objects themselves. 1I/‘Oumuamua was detected by the Pan-STARRS1 survey on 19 October 2017, already past perihelion and outbound [4]. It displayed no detectable coma or outgassing yet exhibited an anomalous non-gravitational acceleration that remains without consensus explanation [10]. Its extreme aspect ratio exceeding 6:1, unlike any known Solar System body [⁴]. 1I/ʻOumuamua was discovered only because it passed within 0.16 AU of Earth. Had its trajectory differed by a small margin, it would have gone unnoticed entirely.
2I/Borisov was discovered on 30 August 2019 by amateur astronomer Gennadiy Borisov6. With a classic cometary coma and CO abundance exceeding 170% relative to H₂O [11,12], it was the most compositionally familiar of the three ISOs, yet its volatile inventory pointed to formation in a carbon-rich environment unlike our own protoplanetary disk. Borisov survived into our catalogs precisely because it behaved like a comet; its activity made it bright enough for a 0.65-m amateur telescope to find. If the next five hundred interstellar detections resemble 1I/’Oumuamua, then it is 2I/Borisov (not ‘Oumuamua), that is the anomaly. The assumption that cometary activity constitutes the default state of interstellar matter may itself be an artifact of survivorship bias: we classify Borisov as ‘normal’ precisely because it resembles the objects our instruments were built to find.
3I/ATLAS which was detected on 1 July 2025 by the ATLAS survey at approximately 4.5 AU from the Sun7, proved the most massive and complex. Hubble Space Telescope (HST) observations constrained the nucleus radius to rn ≤ 2.8 km8 (an initial upper bound from early imaging); subsequent analysis extracted a refined estimate of rn = 1.3 ± 0.2 km15, yet revealed that the surrounding dust coma contributed the vast majority of observed brightness. The dust coma dominated the optical cross-section of 3I/ATLAS. Without it, the object would have gone undetected by ATLAS8. The object displayed sunward dust emission indicating anisotropic ejection from the dayside of the nucleus rather than a conventional radiation-pressure-shaped dust tail8. The James Webb Space Telescope (JWST) spectroscopy revealed a CO₂-dominated volatile inventory with anomalous nickel-to-iron ratios [¹³], while polarimetric observations showed extreme negative polarization reaching −2.7% at 7° phase angle, unprecedented among known comets [¹⁴].
A critical observation unites these three detections: all were found serendipitously by wide-field surveys designed for other purposes. None was detected by a dedicated ISO search program. Moreover, all three exhibited properties that made them conspicuous: large effective cross-sections, relatively moderate interstellar velocities (26–68 km/s), and heliocentric distances within approximately 5 AU at discovery. The number density of objects with the scattering cross-section of 3I/ATLAS has been estimated as ~3 × 10⁻⁴ au⁻³ at first [⁷], but this figure was based on a coma-contaminated absolute magnitude that substantially overestimated the nucleus size. Subsequent HST nucleus extraction yielded a true radius of rn = 1.3 ± 0.2 km six months later [15], revising the number density upward by an order of magnitude to ~3 × 10⁻³ au⁻³ and implying that approximately one comparable object resides within 4.5 AU of the Sun at any given time [15]. Yet, HST observations confirm that 3I/ATLAS itself would have escaped detection without the brightness enhancement provided by its dust coma [⁸]. Previous surveys missed such objects because inactive nuclei lack the coma brightness that made 3I/ATLAS detectable [⁸]. This is survivorship bias in action: the properties of the detected ISOs reflect the detection threshold, not the intrinsic object population.
The Representativeness Trap: How Cognitive Biases Compound Statistical Distortion
A further cognitive dimension compounds the statistical problem, and it is this dimension that distinguishes our conceptual framework from purely astronomical discussions of completeness. The three known ISOs exhibit strikingly different properties: 1I/’Oumuamua was anomalously shaped and inert, 2I/Borisov was conventionally cometary, and 3I/ATLAS was chemically complex and massive. Such diversity activates what Kahneman and Tversky termed the representativeness heuristic — the tendency to judge a small sample as representative of the parent population when it displays internal variety [16].
The representativeness heuristic leads individuals to evaluate the probability that a sample belongs to (or represents) a population based on the degree to which it resembles the population’s expected features, rather than on the actual statistical properties of the sampling process [16,17]. Critically, Tversky and Kahneman demonstrated that people are systematically insensitive to sample size when evaluating the reliability of statistical results, expecting small samples to reproduce the properties of the parent population [18]. The related tendency to judge representativeness by surface resemblance rather than sampling logic [16] suggests that a small sample displaying apparent variety may be treated as though it were a large, representative one. In the ISO context, the apparent coverage of “anomalous”, “normal”, and “complex” archetypes creates a compelling but illusory sense that the population has been adequately sampled. This heuristic has been documented extensively in scientific reasoning itself. Nickerson [19] reviewed confirmation bias in science — the tendency to interpret new evidence as consistent with existing beliefs and noted that people are particularly susceptible to drawing premature conclusions, which are then reinforced by selective attention to supportive evidence.
Greenwald [20] demonstrated the consequences of prejudice against null results, showing how positive findings in small samples are outweighed relative to their evidential value. More recently, Ioannidis [21] formalized conditions under which published research findings are likely to be false, with small sample size as a primary risk factor. Research on team cognition in expert groups has revealed that shared mental models, while facilitating coordination [22], can also promote premature consensus and resistance to disconfirming evidence when teams operate under time pressure with limited data [23].
Three data points drawn exclusively from the detectable fraction of parameter space cannot constrain the properties of the undetectable majority, regardless of how different those three points appear from one another. The analogy is direct: if a marine biologist caught only three fish, all near the surface, all attracted to bait — the fact that they belonged to three different species would not justify conclusions about the deep ocean. Recognizing this heuristic trap is essential if the astronomical community is to resist premature closure on the nature of the interstellar population.
Premature closure is not merely a theoretical concern. The rapid proliferation of formation and ejection models tailored to the properties of three objects [24,25,26] suggests that the field may already be anchoring on a biased sample. Anchoring, the tendency to rely excessively on the first available information, is another well-documented cognitive bias [27] that compounds representativeness. Once researchers invest intellectual effort in explaining the three known ISOs, the psychological cost of acknowledging that these objects may be atypical of the broader population rises, creating resistance to revision even in the face of null results from future surveys. People must be aware of this dynamic and actively guard against it.
Quantifying the Invisible Majority
The survivorship-bias framework carries immediate quantitative implications. Population estimates derived from the known sample are necessarily lower bounds on the true ISO number density. To estimate the fraction of parameter space currently accessible to detection, we consider four independent axes of observational sensitivity.
(i) Size: current surveys require effective diameters of at least 100 m at distances of 1 AU in their sensitivity to reflected sunlight. Power-law extrapolation of Solar System size distributions suggests that objects below this threshold outnumber those above it by orders of magnitude [15], so optical surveys sample a small fraction of the actual size distribution. Indeed, Peña-Asensio and Seligman [34] argued that a power-law extrapolation from spacecraft-detected interstellar dust to kilometer-scale ISOs overpredicts the number of intermediate-sized interstellar meteoroids by 2–7 orders of magnitude relative to meteor survey constraints, revealing a flux-gap across the very size range where current instruments are blind.
(ii) Albedo: optical surveys are sensitive to reflected sunlight and therefore preferentially detect objects with moderate-to-high albedo. Bodies with geometric albedo below ~0.02 — analogous to the darkest known asteroids, would fall below detection thresholds at distances beyond ~0.5 AU; we estimate roughly 30% of the albedo distribution is currently accessible.
(iii) Velocity: the requirement for multi-night arc detections to confirm an object and compute an orbit imposes an effective velocity ceiling of approximately 200 km/s; faster objects produce a faint, long streak or single-frame detections that current pipelines discard. Based on theoretical velocity distributions for stars in the solar neighborhood, approximately 40% of ISOs are expected below this threshold.
(iv) Activity: Two of the three detected ISOs (2I/Borisov and 3I/ATLAS) exhibited outgassing and dust production that amplified their apparent brightness by factors of 10–1000 relative to their bare nuclei; inactive bodies of equivalent nucleus size would be 2.5–7.5 magnitudes fainter [8,15]. The sole inactive detection, 1I/’Oumuamua, required a closest approach of just 0.25 AU to Earth, an exceptionally rare geometry that underscores how difficult it is to detect ISOs without coma enhancement. We estimate that roughly 10% of ISOs display sufficient activity to produce coma enhancement at heliocentric distances where current surveys operate.
Treating these axes as approximately independent, the combined detection fraction is ~0.10 × 0.30 × 0.40 × 0.10 ≈ 0.001, equivalent to about 0.1% of the total ISO parameter space. This estimate is necessarily approximate as the axes are not perfectly independent, and each factor carries uncertainty of at least a factor of two, but it establishes that current surveys are sensitive to a very small fraction of the interstellar population. We emphasize that this is a conceptual estimate intended to illustrate the scale of the problem, not a rigorous statistical bound.
Combined with the arrival rate of approximately one 3I/ATLAS-like object per year within 4.5 AU, this implies that multiple interstellar visitors have been transiting the inner Solar System undetected throughout the era of modern sky surveys. Such a revised estimate accounts only for objects that resemble the three survivors. The population of dark, inactive, or fast ISOs remains entirely unconstrained. 2I/Borisov and 3I/ATLAS both exhibited outgassing driven by water and carbon-bearing volatiles, while 1I/’Oumuamua appeared entirely inactive — itself a datum that current models struggle to explain. This is not a property of the interstellar population; it is a property of the detection threshold. Just as Malmquist bias [28] distorts flux-limited stellar samples, ISO detection inherits compounding selection effects across all four axes simultaneously, and no forward-model detection function yet exists to correct for them. We are studying the distribution of damage on returning bombers and concluding that engines are rarely hit.
Armoring the Engines: A Multi-Modal Detection Architecture
Wald’s recommendation was to armor where damage was absent, not where it was present. The analogous prescription for ISO science is to invest in detection capabilities for the classes of objects we are currently not finding. No single instrument can overcome a multi-dimensional survivorship bias. What is required is a complementary architecture in which each modality addresses a specific axis of invisibility. The Vera C. Rubin Observatory’s Legacy Survey of Space and Time (LSST), with its 8.4-meter aperture and four-night cadence, will push the optical detection threshold to smaller diameters and increase the ISO detection rate from ~1 per decade to potentially several per year [29,30]. However, Rubin observes only the southern sky. Full temporal coverage requires a northern-hemisphere counterpart such as the planned Argus Array (https://argus.unc.edu/); the proposed Comprehensive ISO Network (CISON) architecture [31] would close the geometric gap by combining dual-hemisphere wide-field discovery with rapid high-resolution characterization and selective escalation to interceptor missions. Current pipelines demand multi-night detections to confirm an object and compute an orbit. For ISOs transiting the inner Solar System above ~200 km/s, this window collapses. Real-time machine-learning pipelines operating on single exposures are necessary to capture the fastest visitors. By coupling discovery architecture to predictive classification frameworks, ISO assessment can shift from reactive to anticipatory, identifying objects likely to escape detection before they do.
The most fundamental long-term solution lies in gravitational detection. Thoss and Loeb [32] showed that proposed space-based gravitational-wave experiments, particularly DECIGO, could detect the perturbation of detector test masses by dark objects streaming through the Solar System, with detection volumes reaching several million kilometers for sufficiently massive perturbers. Although their analysis targets dark matter clumps and primordial black holes, the method applies generically to any unbound massive body and is entirely independent of electromagnetic radiation, albedo, or outgassing. For ISO-scale masses, current projections require extremely close approaches (sub-AU for LISA). This capability therefore remains contingent on future detector sensitivities and is included here to illustrate the complete detection architecture rather than as a near-term solution. Finally, ESA’s Comet Interceptor [33] and proposed rapid-response platforms ensure that characterization is not biased toward properties measurable only by remote photometry. In-situ measurements can determine whether an intercepted object is representative or anomalous in ways that remote observation cannot.
Discussion and Conclusions
We have argued that interstellar object science is subject to a form of survivorship bias that is multi-dimensional, severe, and structurally analogous to the problem Abraham Wald identified in 1943. This conclusion is supported by independent flux analyses [34] which demonstrated that spacecraft dust measurements and kilometer-scale ISO detections cannot be connected by a single size-frequency distribution, implying that the detected populations may represent distinct source reservoirs rather than endpoints of a continuous spectrum. In both cases, the sample available for study has been filtered by a process that preferentially removes the most informative cases, and in both cases, the correct response is to direct resources toward the unobserved region of parameter space. The contribution of this Perspective is twofold. First, we propose survivorship bias, as distinct from selection bias, as the appropriate conceptual framework for understanding ISO detection incompleteness, and we distinguish this from the more familiar (and less severe) selection biases that are routinely corrected in other astronomical contexts. Second, we identify the cognitive dimension of the problem: the representativeness heuristic, anchoring, and premature closure operate on small, diverse samples in ways that are well-documented but have not previously been discussed in the ISO literature. Indeed, HST observations have confirmed that 3I/ATLAS itself would have escaped detection without the brightness enhancement provided by its dust coma [8], demonstrating that even the detected sample includes objects that nearly failed to ‘survive’ into our catalogs.
We emphasize that this perspective is entirely agnostic to the composition and origin of undetected ISOs. The survivorship-bias argument applies equally to icy comets, rocky asteroids, metallic fragments, and any other hypothetical objects. What it requires is a recognition that three objects, however scientifically valuable, cannot be treated as representative without explicit correction for the detection function, and that the path to correction runs not through better statistics applied to three data points, but through the construction of instruments capable of finding the next three hundred.
The history of astronomy is, in many ways, a history of overcoming survivorship bias. Every advance in instrumentation, from the optical telescope to the radio dish to the X-ray satellite to a gravitational wave interferometer, revealed populations that were invisible to previous technology. Each time, the newly visible objects were not merely more of the same; they were qualitatively different, populating regions of parameter space that had been structurally inaccessible. There is every reason to expect that the same will hold for interstellar objects. The “detectable bombers” have informed us that the interstellar medium delivers material to our doorstep. The next generation of observatories will inform us about the true statistical and physical nature of these packages.
References
(1) Mangel, M., & Samaniego, F. J. (1984). Abraham Wald’s work on aircraft survivability. Journal of the American Statistical Association, 79(386), 259–267. https://doi.org/10.1080/01621459.1984.10478038
(2) Wald, A. (1980). A method of estimating plane vulnerability based on damage of survivors (CRC 432). Center for Naval Analyses, Operations Evaluation Group. (Original work published 1943).
(3) Ellenberg, J. (2014). How not to be wrong: The power of mathematical thinking. The Penguin Press.
(4) Meech, K., Weryk, R., Micheli, M., Kleyna, J. T., Hainaut, O. R., Jedicke, R., … & Keane, J. V. (2017). A brief visit from a red and extremely elongated interstellar asteroid. Nature, 552, 378–381. https://doi.org/10.1038/nature25020
(5) Bialy, S., & Loeb, A. (2018). Could solar radiation pressure explain ‘Oumuamua’s peculiar acceleration? The Astrophysical Journal Letters, 868(1), L1. https://doi.org/10.3847/2041-8213/aaeda8
(6) Guzik, P., Drahus, M., Rusek, K. et al. (2020). Initial characterization of interstellar comet 2I/Borisov. Nature Astronomy 4, 53–57. https://doi.org/10.1038/s41550-019-0931-8
(7) Seligman, D. Z., Micheli, M., Farnocchia, D., Denneau, L., Noonan, J. W., Hsieh, H. H., Santana-Ros, T., Tonry, J., Auchettl, K., Conversi, L., Devogèle, M., Faggioli, L., Feinstein, A. D., Fenucci, M., Ferrais, M., Frincke, T., Gillon, M., Hainaut, O. R., Hart, K., . . . Zhang, Q. (2025). Discovery and preliminary characterization of a third interstellar object: 3I/ATLAS. The Astrophysical Journal Letters, 989(2), L36. https://doi.org/10.3847/2041-8213/adf49a
(8) Jewitt, D., Hui, M.-T., Mutchler, M., Kim, Y., & Agarwal, J. (2025). Hubble Space Telescope observations of the interstellar interloper 3I/ATLAS. The Astrophysical Journal Letters, 990(1), L2. https://doi.org/10.3847/2041-8213/adf8d8
(9) Eldadi, O., Tenenbaum, G. & Loeb, A. (2025). Advancing interstellar science: a global framework for comprehensive study of interstellar objects. arXiv preprint. https://arxiv.org/abs/2510.01405.
(10) Loeb A. (2022). On the possibility of an artificial origin for ‘Oumuamua. Astrobiology, 22(12), 1392–1399. https://doi.org/10.1089/ast.2021.0193
(11) Bodewits, D., Noonan, J. W., Feldman, P. D., Bannister, M. T., Farnocchia, D., Harris, W. M., Li, J.-Y., Mandt, K. E., Parker, J. W., & Xing, Z. (2020). The carbon monoxide-rich interstellar comet 2I/Borisov. Nature Astronomy, 4(9), 867–871. https://doi.org/10.1038/s41550-020-1095-2
(12) Cordiner, M. A., Milam, S. N., Biver, N., Bockelée-Morvan, D., Roth, N. I., Bergin, E. A., Jehin, E., Remijan, A. J., Charnley, S. B., Mumma, M. J., Boissier, J., Crovisier, J., Paganini, L., Kuan, Y.-J., & Lis, D. C. (2020). Unusually high CO abundance of the first active interstellar comet. Nature Astronomy, 4, 861–866. https://doi.org/10.1038/s41550-020-1087-2
(13) Cordiner, M. A., Roth, N. X., Kelley, M. S. P., Bodewits, D., Charnley, S. B., Drozdovskaya, M. N., Farnocchia, D., Micheli, M., Milam, S. N., Opitom, C., Schwamb, M. E., Thomas, C. A., & Bagnulo, S. (2025). JWST detection of a carbon-dioxide-dominated gas coma surrounding interstellar object 3I/ATLAS. The Astrophysical Journal Letters, 991, Article L43. https://doi.org/10.3847/2041-8213/ae0647
(14) Gray, Z., Bagnulo, S., Borisov, G., Kwon, Y. G., Cellino, A., Kolokolova, L., Dorsey, R. C., Fedorets, G., Granvik, M., & MacLennan, E. (2025). Extreme negative polarization of new interstellar comet 3I/ATLAS. The Astrophysical Journal Letters, 992(2), Article L29. https://doi.org/10.3847/2041-8213/ae0c08
(15) Hui, M.-T., Jewitt, D., Mutchler, M. J., Agarwal, J., & Kim, Y. (2026). Nucleus and postperihelion activity of interstellar object 3I/ATLAS observed by Hubble Space Telescope. arXiv preprint. arXiv:2601.21569.
(16) Kahneman, D., & Tversky, A. (1972). Subjective probability: A judgment of representativeness. Cognitive Psychology, 3(3), 430–454. https://doi.org/10.1016/0010-0285(72)90016-3
(17) Kahneman, D. (2011). Thinking, Fast and Slow. Farrar, Straus and Giroux.
(18) Tversky, A., & Kahneman, D. (1971). Belief in the law of small numbers. Psychological Bulletin, 76(2), 105–110. https://doi.org/10.1037/h0031322
(19) Nickerson, R. S. (1998). Confirmation bias: A ubiquitous phenomenon in many guises. Review of General Psychology, 2(2), 175–220. https://doi.org/10.1037/1089-2680.2.2.175
(20) Greenwald, A. G. (1975). Consequences of prejudice against the null hypothesis. Psychological Bulletin, 82(1), 1–20. https://doi.org/10.1037/h0076157
(21) Ioannidis, J. P. A. (2005). Why most published research findings are false. PLoS Medicine. 2(8), e124. https://doi.org/10.1371/journal.pmed.0020124
(22) Eldadi, O., & Tenenbaum, G. (2025). Team cognition (TC) in sport: Foundations, development, and performance implications. Psychology of Sport and Exercise, 80, 102927. https://doi.org/10.1016/j.psychsport.2025.102927
(23) Janis, I. L. (1982). Groupthink: Psychological studies of policy decisions and fiascoes (2nd ed.). Boston: Houghton Mifflin.
(24) Raymond, S. N., Armitage, P. J., Veras, D., Quintana, E. V., & Barclay, T. (2018). Implications of the interstellar object 1I/’Oumuamua for planetary dynamics and planetesimal formation. Monthly Notices of the Royal Astronomical Society, 476(3), 3031–3038. https://doi.org/10.1093/mnras/sty468
(25) Desch, S. J., & Jackson, A. P. (2021). 1I/’Oumuamua as an N₂ ice fragment of an exo-Pluto surface II: Generation of N₂ ice fragments and the origin of ‘Oumuamua. Journal of Geophysical Research: Planets, 126(5), e2020JE006807. https://doi.org/10.1029/2020JE006807
(26) Bergner, J. B., & Seligman, D. Z. (2023). Acceleration of 1I/’Oumuamua from radiolytically produced H₂ in H₂O ice. Nature, 615, 610–613. https://doi.org/10.1038/s41586-022-05687-w
(27) Tversky, A. & Kahneman, D. (1974). Judgment under uncertainty: Heuristics and biases. Science 185, 1124–1131. https://doi.org/10.1126/science.185.4157.1124
(28) Malmquist, K. G. (1922). On some relations in stellar statistics. Meddelanden fran Lunds Astronomiska Observatorium, Serie I, 100, 1–52.
(29) Dorsey, R. C., Hopkins, M. J., Bannister, M. T., Lawler, S. M., Lintott, C., Parker, A. H., & Forbes, J. C. (2025). The visibility of the Ötautahi-Oxford interstellar object population model in LSST. arXiv preprint. https://doi.org/10.48550/arXiv.2502.16741
(30) Hoover, D. J., Seligman, D. Z., & Payne, M. J. (2022). The population of interstellar objects detectable with the LSST and accessible for in situ rendezvous with various mission designs. The Planetary Science Journal, 3(71). https://doi.org/10.3847/PSJ/ac58fe
(31) Trivedi, O., & Loeb, A. (2026). A comprehensive network for the discovery and characterization of interstellar objects. arXiv preprint. https://doi.org/10.48550/arXiv.2601.21184
(32) Thoss, V., & Loeb, A. (2025). Detecting dark objects in the Solar System with gravitational wave observatories. Physical Review D, 112(8), 083050. https://doi.org/10.1103/g3wh-hg4x
(33) Snodgrass, C., Jones, G.H. (2019). The European Space Agency’s Comet Interceptor lies in wait. Nature Communications, 10, 5418 (2019). https://doi.org/10.1038/s41467-019-13470-1
(34) Peña-Asensio, E., & Seligman, D. Z. (2025). The interstellar flux gap: From dust to kilometer-scale objects. Astronomy & Astrophysics, 704, L1. https://doi.org/10.1051/0004-6361/202557337
ABOUT THE POSTING CO-AUTHOR
Avi Loeb is the head of the Galileo Project, founding director of Harvard University’s — Black Hole Initiative, director of the Institute for Theory and Computation at the Harvard-Smithsonian Center for Astrophysics, and the former chair of the astronomy department at Harvard University (2011–2020). He is a former member of the President’s Council of Advisors on Science and Technology and a former chair of the Board on Physics and Astronomy of the National Academies. He is the bestselling author of “Extraterrestrial: The First Sign of Intelligent Life Beyond Earth” and a co-author of the textbook “Life in the Cosmos”, both published in 2021. The paperback edition of his new book, titled “Interstellar”, was published in August 2024.
Professional website:
https://lweb.cfa.harvard.edu/~loeb/
Social media:
https://avi-loeb.medium.com/
https://www.youtube.com/@ProfessorAviLoeb
https://open.spotify.com/show/1zhndXkvSY2b8FdjspFpCd
https://x.com/ProfAviLoeb
