{"id":"589328c2-2ac8-480c-b10a-3fbd44dcf1f7","arxiv_id":"1908.05291","paper_version":2,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"A Fisher forecast shows that order-of-magnitude improvements in helium-4 abundance measurements could yield light-relic constraints near CMB-S4 sensitivity, but deuterium measurements alone will not be competitive.","lead":"This paper forecasts how precisely future measurements of primordial helium and deuterium abundances would need to be to match next-generation cosmic microwave background experiments in constraining the density of light relics. It finds helium improvements are useful while deuterium is unlikely to keep pace unless nuclear reaction rates and the baryon density are also substantially improved.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Fisher Gaussianity/linearity is the fragile premise for the headline sigma(Neff)=0.0246; the qualitative Yp-vs-D/H ranking likely survives, and an MCMC validation would settle the numerical impact.","rationale":"The paper's central qualitative claim—that improved Yp measurements help and improved D/H measurements are unlikely to catch up to CMB-S4—is well supported by the degeneracy analysis and is robust. The quantitative headline number, however, rests on the Fisher approximation in Eq. (3). I considered other candidate concerns: the CMB-S4 Neff prior already assumes BBN consistency (noted in footnote 44), the current Yp measurement is systematics-limited (noted in the Discussion), and the D/H conclusion is conservative because marginalizing over more rates or using non-Gaussian rate priors would push D/H further from competitiveness. None of these change the verdict. A full MCMC check would either validate the Fisher numbers or show how much they are biased, so the reader's CONDITIONAL verdict remains appropriate; no additional adjustment is needed.","tokens_in":12977,"tokens_out":11997,"duration_ms":133023,"concrete_test":"Run a full MCMC (e.g., with emcee or Stan) over {Neff, eta, 100 nuclear rates} using the paper's modified AlterBBN code for the Rank 1 configuration in Table II (sigma(Yp)/Yp=0.16%, sigma(D/H)=1.2%, CMB-S4 priors on eta and Neff), with log-normal priors on the rates by default and a Gaussian-prior variant as a control. Compare the marginalized 68% interval and the shape of the Neff posterior to the Fisher sigma=0.0246; also check the derivative of Yp and D with respect to Neff and rate 20 at the fiducial and at the edges of the prior. If the MCMC interval differs from 0.0246 by more than ~10%, or the posterior is visibly skewed, the Fisher-based Table II numbers need an explicit non-Gaussianity caveat; if the interval matches, the concern is resolved.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Equation (3) is a pure Fisher forecast: the abundance likelihood and all priors are treated as Gaussian and only first derivatives of the abundances enter. The paper does not validate this with a full posterior. This matters most for the headline number in Table II (Rank 1, sigma(Neff)=0.0246), because at the optimistic Yp precision (0.16% fractional) the posterior is formed by a very informative abundance together with 100 poorly constrained nuclear rates. Several rates have prior errors of 15% or larger (up to 83.7% for rate 87), where a Gaussian 1-sigma prior is a poor description and the abundance response to these rates is not guaranteed to be linear over the prior volume. There is also an unaddressed parameterization subtlety: nuclear rates are positive definite, but if pi in Eq. (3) is the rate itself, a Gaussian prior with fractional error f permits negative rates for entries with f>~0.4; using log-normal priors would remove this pathology and could shift the marginalized sigma(Neff). The D/H conclusion is the most robust part of the paper, since it is driven by the near-linear degeneracy between deuterium and the rate of 2H(p,gamma)3He (rate 20) together with eta; additional non-linearities would only strengthen the statement that D/H alone is not competitive. Thus the concern is about the quantitative forecasts, not the qualitative ranking.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper performs a Fisher-matrix forecast of how well future measurements of primordial light-element abundances, combined with Planck or CMB-S4 priors, can constrain the effective number of relativistic species Neff. The authors use AlterBBN with a 100-reaction network, treat eta and nuclear rates as nuisance parameters, and forecast sigma(Neff) for various assumed improvements in Yp, D/H, and neutron-lifetime measurements. They find that an order-of-magnitude improvement in Yp could yield sigma(Neff)=0.0246 when combined with CMB-S4 priors, whereas improved D/H measurements are not competitive because of degeneracies with eta and the 2H(p,gamma)3He rate. They also identify which nuclear-rate uncertainties limit the Yp-based constraint.","tokens_in":13222,"tokens_out":5124,"duration_ms":52935,"significance":"If the quantitative forecasts survive validation, the paper provides a useful mapping of where BBN abundance measurements need to go to keep pace with CMB-S4 and identifies specific nuclear rates that limit the inference. The qualitative conclusion that Yp is more promising than D/H for constraining Neff is well supported by the degeneracy argument and is likely robust. The paper's use of a full BBN network with 100 rates, external inputs from NACRE II, and the explicit ranking in Table II are strengths. The central quantitative claim, however, rests on a Fisher approximation that is not validated, which tempers the significance of the headline numbers.","major_comments":[{"comment":"The Fisher matrix forecast assumes that the abundance likelihood and all priors are Gaussian and that abundance derivatives are linear over the relevant parameter volume. For the headline case in Table II (rank 1, sigma(Neff)=0.0246) the helium measurement is very informative and the posterior is marginalized over 100 nuclear rates, several with assigned uncertainties of 15% or more; under these conditions Gaussianity and linearity are not guaranteed. I request an MCMC validation of at least the rank-1 and rank-5 cases, or a quantitative test of the linearity assumption (e.g., second-derivative terms), before the numerical constraints are quoted as forecasts.","section":"Section II, Eq. (3) and Table II"},{"comment":"The text does not state how the Planck and CMB-S4 priors on eta and Neff are combined with the Fisher matrix of Eq. (3). If the priors are added as inverse-variance diagonal terms, this should be written explicitly; if not, the procedure is ambiguous. Table II and Figures 1-2 depend on this step, and the current description makes the forecasts difficult to reproduce.","section":"Section II, Eq. (3)"},{"comment":"The nuclear rates are treated as Gaussian random parameters with fractional uncertainties. For rates with large uncertainties, most notably rate 87 with 83.7%, a Gaussian prior assigns substantial probability to negative reaction rates, which is unphysical. I recommend redoing the forecast with log-normal priors on the rates (or otherwise restricting the support) and checking whether sigma(Neff) in Table II shifts; this is also relevant to the Gaussianity assumption in Eq. (3).","section":"Section II and Table III"}],"minor_comments":[{"comment":"Equation (3) contains a duplicated 'of' in the phrase 'abundances of of 2H'.","section":"Section II, Eq. (3)"},{"comment":"The axis labels appear to read '10 3', '10 2', and '10 1' without superscript minus signs; they should be 10^{-3}, 10^{-2}, and 10^{-1}.","section":"Figures 1 and 2"},{"comment":"The phrase 'the relevant unclear rates' should read 'the relevant nuclear rates'.","section":"Section III, discussion of Figure 4"},{"comment":"The caption uses square placeholders for the nuclear reaction rate; these should be replaced with the actual rate symbols so that the four panels are interpretable.","section":"Figure 3 caption"}],"recommendation":"major_revision","confidential_remarks":"The paper is well within scope and the qualitative message is likely correct. The main risk is that the headline Fisher numbers are over-interpreted; validation with an MCMC is the key missing ingredient. I do not see a circularity or novelty problem. If the authors add the requested checks and clarifications, I would support publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Punchline: this is the quantitative forecast the BBN community has been missing: how much better do Yp and D/H measurements need to get to keep pace with CMB-S4 on Neff? The qualitative answer—helium yes, deuterium no—is credible and matches what people suspected; the numerical sigma(Neff)=0.0246 headline is plausible but rests on a Fisher approximation that is not stress-tested.\n\nThe paper does a lot right. It uses a full BBN network with 100 nuclear rates, assigns priors from NACRE II where available, and marginalizes over all of them within the Fisher framework. The ranking in Table II is genuinely useful: an order-of-magnitude Yp improvement plus CMB-S4 priors gets sigma(Neff)=0.0246, while D/H improvements barely move the needle unless you also shrink the uncertainties on rate 20 (2H(p,gamma)3He) and rate 28 (2H(d,n)3He). That is a concrete, actionable statement for experimentalists. The paper also correctly notes that current Yp measurements are systematics-limited, and that the neutron lifetime is not currently the bottleneck.\n\nThe soft spots are the quantitative ones. Equation (3) is pure Fisher: Gaussian likelihood, first derivatives only. With 100 nuclear rates, several with 15% or larger prior uncertainties (rate 87 at 83.7%), a Gaussian prior on a positive-definite rate is a real pathology—it allows negative rates at beyond about 1.4 sigma. A log-normal prior would remove that and could shift the marginalized sigma(Neff). The paper also assigns a flat 15% uncertainty to any rate without a tabulated value, which is ad hoc, though they argue none of those are limiting. And the CMB-S4 priors on Neff and eta are treated as independent; the real forecast has correlation that could change the combination. None of this kills the qualitative ranking: the D/H conclusion is driven by a near-linear degeneracy with rate 20 and eta, and nonlinearities would only make D/H look worse. But the headline number should be viewed as a Fisher estimate, not a prediction; an MCMC validation would settle whether it is biased.\n\nOne more thing: the modified AlterBBN code is not released. For a forecasting paper, shipping the code would make the numbers reproducible. That is a reasonable condition.\n\nBottom line: this is a solid, useful forecasting paper. The qualitative conclusions are robust and the quantitative claims are clearly flagged as Fisher forecasts, though the approximation is not validated. A serious referee should ask for an MCMC check and the code, but this is not a desk reject. I would bring it to reading group as an example of careful Fisher forecasting with a clear message.","headline":"A useful Fisher forecast showing Yp matters and D/H doesn't for Neff, but the headline numbers need an MCMC robustness check.","tokens_in":13785,"tokens_out":4237,"would_cite":true,"duration_ms":36790,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["26.35.+c","98.80.-k"],"model":"deepseek-v4-flash","headline":"A Fisher forecast over a 100-reaction BBN network shows that an order-of-magnitude helium-4 improvement could match or beat projected CMB limits on light relics, while deuterium improvements are blocked by degeneracies.","keywords":["Big Bang nucleosynthesis","light relics","effective number of neutrino species","primordial helium-4 abundance","deuterium abundance","Fisher matrix forecast","nuclear reaction rates","cosmic microwave background"],"falsifier":"Compute the full posterior (for example with a Markov-chain Monte Carlo) for the best-ranked scenario — a tenfold helium-4 improvement with projected next-generation CMB priors — using the same BBN network and uncertainties. If the posterior sigma(Neff) differs from the Fisher forecast of 0.0246 by more than about 10 percent, the Gaussian-linear assumption fails and the quantitative conclusions are unreliable. A separate test: if an improved deuterium measurement at 0.12 percent fractional uncertainty, with current nuclear-rate uncertainties and no CMB eta prior, sharpens sigma(Neff) far below the forecast value near 0.22, then the paper's claim that deuterium is non-competitive is contradicted.","tokens_in":12745,"feed_emoji":"🔭","tokens_out":16079,"duration_ms":140401,"temperature":0.7,"pith_summary":"The paper asks whether measurements of primordial light-element abundances can keep pace with coming cosmic microwave background (CMB) constraints on light relics — extra radiation beyond standard neutrinos, which the paper tracks through the effective number of neutrino species $N_{\\mathrm{eff}}$. Using a Fisher-matrix forecast over a Big Bang nucleosynthesis (BBN) network with 100 adjustable nuclear rates, it finds that an order-of-magnitude improvement in the primordial helium-4 abundance ($Y_p$) brings $\\sigma(N_{\\mathrm{eff}})$ close to what next-generation CMB surveys will achieve, and that combining such a helium measurement with projected CMB priors reaches $\\sigma(N_{\\mathrm{eff}}) = 0.0246$. The paper finds the opposite for deuterium: because deuterium is degenerate with the baryon-to-photon ratio $\\eta$ and with the rate of the reaction $^2\\mathrm{H}(p,\\gamma)^3\\mathrm{He}$, even a tenfold better deuterium measurement gives little improvement unless those auxiliary parameters are pinned down. The authors conclude that improved helium-4 measurements are a useful complement to CMB light-relic constraints, while improved deuterium measurements are unlikely to be competitive.","feed_headline":"Tenfold better helium-4 tightens light-relic limits; deuterium cannot","feed_subtitle":"Better helium-4 could rival CMB surveys for extra-light-relic detection; deuterium cannot.","key_machinery":"The machinery is a Fisher information matrix built from finite-difference derivatives of five computed primordial abundances ($^2\\mathrm{H}$, $^3\\mathrm{He}$, $^4\\mathrm{He}$ as $Y_p$, $^6\\mathrm{Li}$, $^7\\mathrm{Li}$) with respect to $N_{\\mathrm{eff}}$, the baryon-to-photon ratio $\\eta$, and the rates of the 100 reactions in the BBN network. A publicly available BBN code is modified so each nuclear rate can be varied; the current abundance measurements and tabulated nuclear-rate uncertainties enter as Gaussian errors, and CMB priors on $\\eta$ and $N_{\\mathrm{eff}}$ add to the same matrix. The key diagnostic is the correlation of $N_{\\mathrm{eff}}$ with each other parameter for a given abundance: for helium the dominant degeneracies are the neutron decay rate and the reaction $\\mathrm{H}(n,\\gamma)^2\\mathrm{H}$, while for deuterium the rate $^2\\mathrm{H}(p,\\gamma)^3\\mathrm{He}$ is the limiting degeneracy.","core_discovery":"The central discovery is an asymmetry in the information content of the two most precisely measured primordial abundances. The primordial helium-4 mass fraction $Y_p$ responds to $N_{\\mathrm{eff}}$ through the neutron-to-proton ratio at weak freeze-out, and this response survives marginalization over $\\eta$ and the nuclear rates; therefore better $Y_p$ measurements consistently tighten the forecast on $N_{\\mathrm{eff}}$. The deuterium abundance, in contrast, is so strongly degenerate with the reaction rate $^2\\mathrm{H}(p,\\gamma)^3\\mathrm{He}$ and with the baryon-to-photon ratio that the current deuterium constraint already sits near a floor: without order-of-magnitude improvements in those auxiliary parameters, improved deuterium data add almost nothing for $N_{\\mathrm{eff}}$. With projected next-generation CMB priors and a tenfold helium improvement, the forecast reaches $\\sigma(N_{\\mathrm{eff}}) = 0.0246$, enough to see a single new thermal fermion ($\\Delta N_{\\mathrm{eff}} \\ge 0.047$) at just under 2$\\sigma$.","pith_inferences":["One extension the paper leaves implicit: if helium-4 systematics improve by more than an order of magnitude, the next bottlenecks are the neutron-lifetime uncertainty and the $\\mathrm{H}(n,\\gamma)^2\\mathrm{H}$ rate, so nuclear-reaction experiments and cosmological abundance surveys would need to be planned together.","The same Fisher formalism could be inverted to ask how well a hypothetical light relic's spin and decoupling temperature could be distinguished from a pure density shift, using the small differences in BBN versus CMB sensitivity that the paper notes rather than only the shared parameter $N_{\\mathrm{eff}}$.","A non-Gaussian or nonlinear test of the forecast, such as a full Markov-chain Monte Carlo over the same scenarios, would establish whether the quoted $\\sigma(N_{\\mathrm{eff}})$ values hold; the Fisher approximation itself is the fragile link."],"forward_implications":["A tenfold helium-4 improvement combined with projected next-generation CMB priors reaches $\\sigma(N_{\\mathrm{eff}}) = 0.0246$, which could detect a single new thermal fermion ($\\Delta N_{\\mathrm{eff}} \\ge 0.047$) at just under 2$\\sigma$.","A tenfold deuterium improvement alone changes the $N_{\\mathrm{eff}}$ constraint by only a few percent once nuclear-rate uncertainties are marginalized, so deuterium is not a competitive probe of the light-relic density.","To unlock deuterium, the rates of $^2\\mathrm{H}(p,\\gamma)^3\\mathrm{He}$ and $^2\\mathrm{H}(d,n)^3\\mathrm{He}$ must improve by roughly an order of magnitude together with a CMB-quality measurement of $\\eta$; with fixed rates and such an $\\eta$ prior, deuterium can give about a factor-of-two better $N_{\\mathrm{eff}}$ than the CMB alone.","The neutron lifetime is not a limiting factor at current precision, but it becomes one if $Y_p$ improves by more than an order of magnitude.","The rankings apply to a fixed BBN relation between $Y_p$ and $N_{\\mathrm{eff}}$; physics that changes that relation is outside the forecast."],"supporting_citations":[{"why":"The BBN review that supplies the deuterium-bottleneck picture and the sensitivity of abundances to the baryon-to-photon ratio, and that is cited for the deuterium degeneracy with Neff.","marker":"[1]"},{"why":"Current CMB constraints on eta and Neff used as the baseline priors in the forecasts.","marker":"[4]"},{"why":"Projected next-generation CMB sensitivity to Neff (sigma(Neff)=0.03) and eta used for the optimistic forecast scenarios.","marker":"[9, 20]"},{"why":"The current measurement of the primordial helium-4 abundance Yp that defines the baseline from which order-of-magnitude improvements are considered.","marker":"[23]"},{"why":"The current measurement of the primordial deuterium abundance D/H that defines the baseline discussed in the forecasts.","marker":"[24]"},{"why":"The publicly available BBN code modified so that each nuclear reaction rate can be varied, providing the abundance derivatives.","marker":"[38, 39]"},{"why":"The compilation of nuclear reaction rates and their uncertainties that supplies most of the rate errors marginalized over in the Fisher matrix.","marker":"[42]"},{"why":"The neutron lifetime value and uncertainty that set the error on the n-p rate and bound the helium-4 forecast.","marker":"[45]"}],"fun_headline_variants":["Helium-4, not deuterium, is key to light relics","For light relics, helium-4 wins; deuterium can't compete","Deuterium plateau: only helium-4 moves N_eff","Tenfold helium-4 gain rivals CMB; deuterium stuck","Light relics: helium-4 data cuts limits, deuterium lags"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The forecast treats the abundance likelihood as Gaussian and assumes abundance responses are linear over the parameter ranges considered, and the paper does not check this against a full likelihood calculation; if either fails, the quantitative sigma(Neff) values in Table II could be biased.","fun_headline_variants_meta":{"raw":{"variants":["Helium-4, not deuterium, is key to light relics","For light relics, helium-4 wins; deuterium can't compete","Deuterium plateau: only helium-4 moves N_eff","Tenfold helium-4 gain rivals CMB; deuterium stuck","Light relics: helium-4 data cuts limits, deuterium lags"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00102,"raw_usage":{"total_tokens":4297,"prompt_tokens":930,"completion_tokens":3367,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":546,"completion_tokens_details":{"reasoning_tokens":3273}},"tokens_in":546,"tokens_out":3367,"duration_ms":21834,"temperature":1.0,"reasoning_tokens":3273,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T13:17:42.098463+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the full posterior (for example with a Markov-chain Monte Carlo) for the best-ranked scenario — a tenfold helium-4 improvement with projected next-generation CMB priors — using the same BBN network and uncertainties. If the posterior sigma(Neff) differs from the Fisher forecast of 0.0246 by more than about 10 percent, the Gaussian-linear assumption fails and the quantitative conclusions are unreliable. A separate test: if an improved deuterium measurement at 0.12 percent fractional uncertainty, with current nuclear-rate uncertainties and no CMB eta prior, sharpens sigma(Neff) far below the forecast value near 0.22, then the paper's claim that deuterium is non-competitive is contradicted.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"The current measurement of the primordial helium-4 abundance Yp that defines the baseline from which order-of-magnitude improvements are considered."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"The compilation of nuclear reaction rates and their uncertainties that supplies most of the rate errors marginalized over in the Fisher matrix."},{"cited_title":"Tanabashi, K","cited_arxiv_id":null,"evidence_quote":"The neutron lifetime value and uncertainty that set the error on the n-p rate and bound the helium-4 forecast."}],"review_version":1}