Brave New Data: Pricing Expertise and Discovery
Listen
Radio episode about this paper
Transcript
Introduction to the show: ident: Economics. The day in economics: theory, mechanisms and the evidence beneath them.
Ruth: I'm Ruth.
Warren: And I'm Warren, here with Iris.
Ruth: We're looking at the paper: "Brave New Data: Pricing Expertise and Discovery".
Warren: Here's the gist of the paper. We study how a data buyer procures evidence from an expert whose initial knowledge and subsequent investigation are private The optimal contract pays only for disclosed evidence, independently of the initial report.
Ruth: Model Setup and Incentives The model studies a data procurement setting in which an expert provides an initial assessment and undertakes further investigation.
Warren: The agent has private information about the state, denoted as µ, where Pr(θ = 1 µ) = µ.
Ruth: Investigation can produce a certificate, which is verifiable evidence establishing that the targeted state is true.
Paper summary: Ruth: So, we're looking at this new paper called "Brave New Data: Pricing Expertise and Discovery". It’s about how someone who knows something privately, like an expert who assesses a state or condition, gives you an initial report and then does some more digging. The core idea is figuring out the best contract for paying that expert.
Warren: Right. The paper studies this data procurement setting where the buyer has private information about a state, let's call it mu, and the expert can either give a certificate proving that state or they can just stay silent and do more investigation to find evidence. The main thing they claim is that the optimal contract only pays for the evidence that actually gets disclosed, regardless of whether it was the initial report or what came from the expert's further search.
Iris: It seems like they are setting up a model where we have an expert who knows something private and can produce verifiable evidence later on. The research focuses on how this structure affects what happens next for both the buyer and the expert, which is kind of like looking at moral hazard in delegated information acquisition CHADE AND KOVRIJNYKH two thousand sixteen <ref:2611.07278#pg1>.
Ruth: Exactly. They look at two benchmark environments where they control what's being observed, and they find that when acquiring evidence is cheap, it makes sense for the contract to give equal rewards for both potential certificates. But as costs go up, rewarding only one certificate becomes the optimal move even if the environment is symmetric <ref:2611.07278#pg3>.
Warren: That's interesting because it means that when acquisition costs are high, we switch from paying for both possibilities to just one of them, even though the situation looks identical on paper. It suggests that the value of discovery changes depending on how much effort it takes to find out.
Iris: The authors characterize this optimal contract using some thresholds, k1 and k2, which depend directly on those acquisition costs. These thresholds dictate when the principal switches from offering rewards for both certificates to rewarding only one or when they change the intensity of the search for a single certificate <ref:2611.07278#pg3>.
Ruth: So, those thresholds essentially map out two different payment structures, one where you reward both possibilities and another where you only reward one. The choice between those two structures is about balancing how easy it is to get the discovery versus how valuable that discovery actually is for making decisions <ref:2611.07278#pg2>.
Warren: That sounds like the decision value of discovery, which they define as the expected loss avoidance before subtracting any payments or acquisition costs. It's about how much avoiding a bad decision is worth compared to what you pay the expert <ref:2611.07278#pg3>.
Iris: They show that truthful reporting by the expert turns out to be weakly optimal under these report-independent payments, which is something interesting given the setup of their equilibrium selection <ref:2611.07278#pg3>. That means if you tell the truth, you're not really punished in this model.
Ruth: So, what does this mean for us? It highlights that we need to look at the expert's contribution beyond just their first assessment. The contract structure itself is what shapes how much discovery happens along two different lines: which certificates get rewarded and what the resulting search targets are <ref:2611.07278#pg3>.
Warren: And the choice between those payment structures reflects how those two margins work together to determine both the ease of inducing discovery and its actual value for a decision-maker. It’s not just about one thing; it’s about that trade-off <ref:2611.07278#pg2>.
Iris: The model's limitations, as stated by the authors, is that they are studying a setting where the expert provides an initial assessment and then undertakes further investigation <ref:2611.07278#pg3>. They focus on how this structure works under report-independent payments <ref:2611.07278#pg3>, but it doesn't necessarily cover every single nuance of real-world data procurement where the initial report is heavily relied upon CHADE AND KOVRIJNYKH two thousand sixteen.
Ruth: That’s fair. The paper’s main contribution seems to be establishing this framework for pricing expertise and discovery, showing how cost regimes shift the optimal payment structure from rewarding both options to rewarding only one <ref:2611.07278#pg3>.
Warren: It really boils down to when acquisition costs become high enough that the symmetry of the environment doesn't matter anymore, and you just pay for one outcome because it’s more efficient, even if you lose some potential accuracy on the other side <ref:2611.07278#pg3>.
Iris: So, to sum up, "Brave New Data: Pricing Expertise and Discovery" shows that the contract design is crucial; it determines whether we incentivize both potential discoveries or just one, based on how expensive it is to find the evidence <ref:2611.07278#pg3>.
Ruth: And for the world, this means that when you're buying data from an expert, you need to think about what the contract is doing—it’s not just about the initial assessment, but how that further search is incentivized <ref:2611.07278#pg3>.
Warren: It gives us a way to analyze why some experts might be more willing to do deep investigation than others, depending on those cost thresholds <ref:2611.07278#pg3>.
Iris: We've covered the main points about the optimal contract structure and how costs drive the choice between rewarding both or just one certificate in this paper <ref:2611.07278#pg3>.
Conclusion: Ruth: So we've been looking at how this paper, "Brave New Data: Pricing Expertise and Discovery," tackles the whole process of buying information from an expert who does some deep digging after an initial report.
Warren: It’s about figuring out what kind of contract you need to have when you're paying that expert. Basically, it’s not just about paying for their first opinion; it’s about how much you pay for the actual new evidence they find.
Iris: The authors are setting up a model where the principal has some private information, and the expert can either give a certificate proving that state or keep searching privately. They use this setup to figure out the optimal payment rules.
Ruth: And what they found is that if you make the contract right, it can be set up so you only pay for evidence that actually gets disclosed, regardless of where it comes from. It's about making sure your money matches the actual discovery made.
Warren: The big shift they show is how the optimal payment structure changes depending on how much effort is needed to find that proof. If finding that evidence is cheap, you might pay for both possibilities equally.
Iris: But as the cost of acquiring that evidence goes up, the contract switches to rewarding only one piece of evidence instead of both. That’s a big structural change in how the principal makes decisions about what they fund.
Ruth: It means you have to choose between two ways to incentivize discovery, and that choice depends on whether you value the certainty of finding one thing or having a shot at finding either thing.
Warren: What this really means for anyone buying data is that your contract design is critical; it dictates whether you’re incentivizing broad search or narrow focus based on those cost thresholds they define.
Iris: So while this model shows how cost regimes determine the optimal payment, we still have to ask what happens when the expert's initial assessment itself is flawed, which isn't covered here.
Ruth: Exactly. We’ve seen how costs drive the incentive structure for discovery, but next up we need to look at what those thresholds actually mean for the actual decisions made by the buyer.
Dihan Zou
University of North Carolina at Chapel Hill
econ.TH
Submitted: 2026-11-05
Updated: 2026-11-05
Project page: https://archiquan.github.io/papers/Learning-Optimal-Data-Markets.pdf
License: http://creativecommons.org/licenses/by/4.0/
Importance score: 76/100
The gist: We study how a data buyer procures evidence from an expert whose initial knowledge and subsequent investigation are private The optimal contract pays only for disclosed evidence…
Key concepts
- Initial Assessment
- The expert provides a first opinion or assessment about the true state of affairs ($\mu$). This is private information known only to the expert before further investigation begins. The buyer observes this report and decides whether to proceed with further evidence gathering.
- Certificate
- This is verifiable evidence produced by the expert's investigation that confirms a specific targeted state ($ heta=1$) is true. It serves as concrete proof of discovery, which the data buyer can use to make decisions about acquiring more information.
- Extensive Margin
- This margin relates to which certificates receive rewards in the contract structure. It concerns the decision of whether to pay for both possible outcomes or only one outcome when multiple pieces of evidence are available. This choice is influenced by how easy it is to induce discovery.
- Intensive Margin
- This margin focuses on the effort expended by the expert once a specific search target has been chosen. It describes the level of investigation or search intensity an expert undertakes, conditional on what they are trying to prove.
Terminology
Summary
We study how a data buyer procures evidence from an expert whose initial knowledge and subsequent investigation are private The optimal contract pays only for disclosed evidence, independently of the initial report. p. 0 The buyer balances the ease of inducing discovery against its value for
decisions beyond the expert’s initial assessment. An optimal contract pays only
for disclosed evidence, independently of the initial report.Where the paper says it
Model Setup and Incentives
The model studies a data procurement setting in which an expert provides an initial assessment and undertakes further investigation. The agent has private information about the state, denoted as µ, where Pr(θ = 1 µ) = µ. p. 5 The principal and agent initially share the
same prior, and payments can depend on reports and the subsequently observed state. In our setting, the expert’s private initial information affects both the probability of
discovering evidence and its incremental value for the buyer’s decision. The agent then reports her assessment and privately chooses which state to
investigate and how much effort to exert. Investigation can produce a certificate: verifiable evidence establishing that the targeted state is true. Acquired evidence
can be withheld but cannot be fabricated. The principal observes the initial report
and either one certificate or silence (no disclosure), and then chooses his action optimally.Where the paper says it
Where the paper says it
Where the paper says it
Benchmark Analysis
The paper presents optimal contracts in two benchmark environments where the principal directly observes µ. In the first benchmark, the principal also controls evidence acquisition. p. 9 3 Benchmark Analysis: Free Expertise
This section presents the optimal contracts in two benchmark environments in which
the principal directly observes µ. In the first benchmark (“first best”), the principal
also controls evidence acquisition. In the first benchmark (“first best”), the principal
also controls evidence acquisition. In the second benchmark (“second best”), the agent
privately chooses her search target and effort. The principal therefore chooses which certificates to
reward and the size of their bonuses to balance the reduction in decision loss against
expected payments. When acquisition is sufficiently cheap, equal rewards for both certificates are optimal. p. 4 The principal therefore chooses which certificates to
reward and the size of their bonuses to balance the reduction in decision loss against
expected payments. When acquisition is sufficiently cheap, equal rewards for both certificates are optimal.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Optimal Contract Characterization
The optimal contract can be chosen to pay nothing when no certificate is disclosed and to offer either equal rewards for both certificates or a reward for only one. The choice between these payment structures balances the ease of inducing discovery against its value for the principal’s decision. p. 4 These payments do not depend on the initial report, although the principal uses
that report when interpreting silence and choosing his action. The choice between
these payment structures balances the ease of inducing discovery against its value for
the principal’s decision. When acquisition is
cheap, equal rewards for evidence of either state induce intensive search whose
failure is informative. As costs rise, rewarding evidence of only one state becomes
optimal despite the symmetry of the environment, reducing expected payments at
the expense of decision accuracy.Where the paper says it
Where the paper says it
Key Findings on Cost Regimes
The optimal contract shapes discovery along two margins: the extensive margin concerns which certificates receive rewards and the resulting search targets; the intensive margin concerns effort conditional on a target. When acquisition is sufficiently cheap, equal rewards for both certificates are optimal. As acquisition becomes more costly, the optimal payment structure switches to rewarding only one certificate. p. 4 The value of investigation comes from both its discoveries and
what the principal learns when it fails. As acquisition becomes more costly, the optimal payment structure switches to
rewarding only one certificate, even though the two states are symmetric in the primitives.Where the paper says it
Decision Value and Thresholds
The value of discovery measures expected loss avoidance before subtracting payments or acquisition costs. The first threshold, k1, marks an adjustment along the extensive margin: the principal switches from rewarding both certificates to rewarding only one. p. 4 The value of investigation comes from both its discoveries and
what the principal learns when it fails. As acquisition becomes more costly, the optimal payment structure switches to
rewarding only one certificate, even though the two states are symmetric in the primitives. Payments
depend on certificates generated through costly search, and the state remains unverifiable when no certificate is produced. 2
The second strand concerns contracting with adverse selection and moral hazard. 3 Optimal equal rewards . 4 The two parameter thresholds .Where the paper says it
Where the paper says it
Where the paper says it
Conclusion and Implications
These findings highlight the importance of evaluating expert-generated data by its contribution beyond the expert’s initial assessment. The paper establishes that truthful reporting is weakly optimal under report-independent payments, as allowed by the model’s equilibrium selection. p. 8 The following lemma establishes that these conditions provide a lossless formulation of the contracting problem. Without loss, the principal may restrict attention to truthful reporting and
full disclosure, for which the feasible set of the contracts is characterized by nonnegativity, (FD), and (IC). The same conflict shapes
discovery bonuses: the additional payment for a certificate relative to the payment
when no certificate is disclosed. Consider two distinct types with the same search
target. These payments do not depend on the initial report, although the principal uses
that report when interpreting silence and choosing his action. The choice between
these payment structures balances the ease of inducing discovery against its value for
the principal’s decision.Where the paper says it
Where the paper says it
Where the paper says it
References
BEN-PORATH, E., E. DEKEL, AND B. L. LIPMAN (2026): “Mechanism Design for Acquisition of/Stochastic Evidence,” Theoretical Economics, 21, 99–131. p. 0 Brave New Data: Pricing Expertise and
Discovery
Dihan Zou*
November 6, 2026
Abstract
We study the procurement of verifiable evidence from an expert who privately
assesses a binary state and chooses the direction and intensity of evidence acquisition. The buyer balances the ease of inducing discovery against its value for
decisions beyond the expert’s initial assessment.Where the paper says it
BERGEMANN, D. p. 6 Their multiplicative-separability condition does not hold in our acquisition technology, so our pooling result requires a separate argument. 3
The third strand concerns the design and sale of information; see Bergemann and
Bonatti (2019) for a survey. Their multiplicative-separability condition does not hold in our acquisition technology, so our pooling result requires a separate argument. 3
The third strand concerns the design and sale of information; see Bergemann and
Bonatti (2019) for a survey. Their multiplicative-separability condition does not hold in our acquisition technology, so our pooling result requires a separate argument. 3
The third strand concerns the design and sale of information; see Bergemann and
Bonatti (2019) for a survey. Their multiplicative-separability condition does not hold in our acquisition technology, so our pooling result requires a separate argument. 3
The third strand concerns the design and sale of information; see Bergemann and
Bonatti (2019) for a survey.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
BERGEMANN, D., A. BONATTI, AND T. p. 6 Their multiplicative-separability condition does not hold in our acquisition technology, so our pooling result requires a separate argument. 3
The third strand concerns the design and sale of information; see Bergemann and
Bonatti (2019) for a survey. JEL Classification: D82, D83, D86, O33. *Department of Economics, University of North Carolina at Chapel Hill.Where the paper says it
Where the paper says it
BERGEMANN, D., A. BONATTI, AND A. p. 6 Their multiplicative-separability condition does not hold in our acquisition technology, so our pooling result requires a separate argument. 3
The third strand concerns the design and sale of information; see Bergemann and
Bonatti (2019) for a survey. Bergemann et al. (2018) analyze a seller who designs and
prices experiments for a buyer with private initial information.Where the paper says it
Where the paper says it
CHADE, H. AND N. KOVRIJNYKH (2016): “Delegated Information Acquisition with Moral Hazard,” Journal of Economic Theory, 162, 55–92. p. 5 2Foundational contributions include Lambert (1986), Demski and Sappington (1987), and Osband
(1989). See also Chade and Kovrijnykh (2016), Rappoport and Somma (2017), and Sharma et al.Where the paper says it
CHE, Y.-K. AND K. MIERENDORFF (2019): “Optimal Dynamic Allocation of Attention,” American Economic Review, 109, 2993–3029. p. 9 6Che and Mierendorff (2019) likewise show that, conditional on experimentation, searching for
evidence of the less likely state is optimal in their one-period example. In their dynamic setting,
however, the option to continue learning can make searching for evidence of the more likely state
optimal.Where the paper says it
DASGUPTA, S. p. 6 We establish the reduction needed for our specific contracting protocol. Related work studies screening existing knowledge (Dasgupta, 2023; Dasgupta and Xia, 2025). We establish the reduction needed for our specific contracting protocol. Related work studies screening existing knowledge (Dasgupta, 2023; Dasgupta and Xia, 2025). We establish the reduction needed for our specific contracting protocol. Related work studies screening existing knowledge (Dasgupta, 2023; Dasgupta and Xia, 2025). Brave New Data: Pricing Expertise and
Discovery
Dihan Zou*
November 6, 2026
Abstract
We study the procurement of verifiable evidence from an expert who privately
assesses a binary state and chooses the direction and intensity of evidence acquisition. The buyer balances the ease of inducing discovery against its value for
decisions beyond the expert’s initial assessment.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
DASGUPTA, S. AND Z. XIA (2025): “Screening Knowledge with Verifiable Evidence,” Working paper, https://www. p. 6 We establish the reduction needed for our specific contracting protocol. Related work studies screening existing knowledge (Dasgupta, 2023; Dasgupta and Xia, 2025). We establish the reduction needed for our specific contracting protocol. Related work studies screening existing knowledge (Dasgupta, 2023; Dasgupta and Xia, 2025).Where the paper says it
Where the paper says it
DEMSKI, J. S. AND D. E. M. SAPPINGTON (1987): “Delegated Expertise,” Journal of Accounting Research, 25, 68–89. p. 5 2
The second strand concerns contracting with adverse selection and moral hazard. 2Foundational contributions include Lambert (1986), Demski and Sappington (1987), and Osband
(1989).Where the paper says it
EPOCH AI (2026): “About FrontierMath,” Accessed October 19, 2026. p. 2 1OpenAI (2024) describes its initial concerns about benchmark reliability and the subsequent review by professional software developers. Epoch AI (2026) documents OpenAI’s commission and the
contracts with researchers to produce problems for the Tier 4 expansion.Where the paper says it
OSBAND, K.: “Optimal Forecasting Incentives,” Journal of Political Economy, 97, 1091–1112. p. 5 A further increase in acquisition cost produces a discrete reduction in the optimal
reward. With weaker search incentives, failure becomes less informative, and some
agents who strongly favor the rewarded state generate no improvement in the principal’s decision: both a certificate and silence lead to the same action.Where the paper says it
PREMISE (n. p. 2 For example, OpenAI engaged professional software developers to assess and review software
engineering tasks for inclusion in SWE-bench Verified, an expert-validated benchmark for evaluating the coding capabilities of large language models. Other examples include Epoch AI’s contracts with mathematicians to create original problems
for FrontierMath and Premise’s payments to local contributors to collect geospatial
data. For example, OpenAI engaged professional software developers to assess and review software
engineering tasks for inclusion in SWE-bench Verified, an expert-validated benchmark for evaluating the coding capabilities of large language models. Other examples include Epoch AI’s contracts with mathematicians to create original problems
for FrontierMath and Premise’s payments to local contributors to collect geospatial
data.Where the paper says it
Where the paper says it
QUAN, L. AND Y. SONG (2026): “Learning-Optimal Design of Data Markets,” Working paper, https://archiquan. p. 6 Bergemann
et al. (2022) and Quan and Song (2026) examine data markets with informational externalities and
endogenous participation, respectively. Bergemann
et al. (2022) and Quan and Song (2026) examine data markets with informational externalities and
endogenous participation, respectively.Where the paper says it
Where the paper says it
RAPPOPORT, D. p. 5 2Foundational contributions include Lambert (1986), Demski and Sappington (1987), and Osband
(1989). See also Chade and Kovrijnykh (2016), Rappoport and Somma (2017), and Sharma et al. 2Foundational contributions include Lambert (1986), Demski and Sappington (1987), and Osband
(1989). See also Chade and Kovrijnykh (2016), Rappoport and Somma (2017), and Sharma et al. Brave New Data: Pricing Expertise and
Discovery
Dihan Zou*
November 6, 2026
Abstract
We study the procurement of verifiable evidence from an expert who privately
assesses a binary state and chooses the direction and intensity of evidence acquisition. The buyer balances the ease of inducing discovery against its value for
decisions beyond the expert’s initial assessment.Where the paper says it
Where the paper says it
Where the paper says it
SHARMA, S., E. TSAKAS, AND M. VOORNEVELD (2025): “Procuring Unverifiable Information,” Mathematics of Operations Research, 50, 1433–1453. p. 5 In our setting, the expert’s private initial information affects both the probability of
discovering evidence and its incremental value for the buyer’s decision. Payments
depend on certificates generated through costly search, and the state remains unverifiable when no certificate is produced.Where the paper says it
WHITMEYER, M. p. 5 His analysis allows
researchers to choose both how much and how to learn, and shows how screening
distorts experimentation when realized posterior beliefs are contractible. Whitmeyer
and Zhang (2023) characterize minimum-cost contracts that induce the acquisition
and truthful reporting of soft information. His analysis allows
researchers to choose both how much and how to learn, and shows how screening
distorts experimentation when realized posterior beliefs are contractible. Whitmeyer
and Zhang (2023) characterize minimum-cost contracts that induce the acquisition
and truthful reporting of soft information.Where the paper says it
Where the paper says it
——— (2026): “Testing Experts Who Can Learn,” Working paper, https://arxiv. p. 6 Whitmeyer and Zhang (2026) study whether outcome-contingent contracts can distinguish experts from nonexperts when initial beliefs are private and experts can
learn. They identify conditions under which learning makes screening possible.Where the paper says it
WITTBRODT, C. p. 6 Related work studies screening existing knowledge (Dasgupta, 2023; Dasgupta and Xia, 2025). Wittbrodt
and Yoder (2026) study costly experimentation by a privately informed agent through menus of observable experiments without transfers. Related work studies screening existing knowledge (Dasgupta, 2023; Dasgupta and Xia, 2025). Wittbrodt
and Yoder (2026) study costly experimentation by a privately informed agent through menus of observable experiments without transfers. Bergemann et al. (2018) analyze a seller who designs and
prices experiments for a buyer with private initial information.Where the paper says it
Where the paper says it
Where the paper says it
YODER, N.: “Designing Incentives for Heterogeneous Researchers,” Journal of Political Economy, 130, 2018–2054. p. 6 Bergemann et al. (2018) analyze a seller who designs and
prices experiments for a buyer with private initial information.Where the paper says it
Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.
Improvements for AI systems
-
Adaptive evidence acquisition strategy: The system can dynamically adjust its search intensity based on whether acquiring evidence of a state confirms or overturns its initial assessment, as described by
The principal chooses an action a in arg min a in arg min E h (a - θ) 2 µ, e ˆ
and the findings thatSearching the more likely state therefore improves the decision only if effort makes silence sufficiently informative to reverse that action.
-
Optimal contract design for varying costs: The system can select between equal rewards for both certificates when acquisition is cheap or rewarding only one certificate when it is more costly, based on whether
acquisition is cheap
versusmore costly,
which reduces expected payments at the expense of decision accuracy in high-cost regimes. -
Prior-informed incentive compatibility: The system's reporting and search choices will be constrained by joint incentive compatibility, meaning
those investigating a given state consider it at least as likely as do those investigating the other state,
which prevents the principal from inducing every agent to investigate her less likely state. -
Threshold-based decision switching: For sufficiently high acquisition costs, the system can switch its strategy based on cost thresholds k1 and k2, where
the optimal bonus switches to rewarding only one certificate
or wherethe two distinct bonuses in part (ii) are both optimal,
leading to different search intensities. -
Decision value maximization: The system can evaluate the net benefit of evidence by calculating the value of discovery V(µ, ˆθ, q) as
min(µ, 1 − µ) − E h (a(µ, e) - θ) 2 µ, ˆθ, qi,
measuring expected loss avoidance before subtracting payments or acquisition costs. -
Targeted effort allocation: The system can determine the optimal effort schedule q(µ) based on the contract type, such as
q(µ) = min (min[µW1/k, 1], 1),
which balances the cost of investigation against expected rewards under different incentive structures like equal rewards or one-certificate rewards. -
Non-asymmetric search target selection: The system's search target will be determined by the interaction between expertise and evidence, such as
The principal values evidence that changes his action: a certificate of the less likely state overturns the action favored by the prior, whereas one of the more likely state confirms it.
-
Cost-benefit comparison for contract selection: The system can compare different contract candidates like
the positive equal-reward candidate S(k) and the capped one-certificate candidate C(k)
to select the optimal payment structure based on which yields a lower expected cost, as demonstrated by the crossing point k1 whereS(k) k1,
C(k) is therefore the best one-certificate cost."
Abstract
We study the procurement of verifiable evidence from an expert who privately assesses a binary state and chooses the direction and intensity of evidence acquisition. The buyer balances the ease of inducing discovery against its value for decisions beyond the expert's initial assessment. An optimal contract pays only for disclosed evidence, independently of the initial report. When acquisition is cheap, equal rewards for evidence of either state induce intensive search whose failure is informative. As costs rise, rewarding evidence of only one state becomes optimal despite the symmetry of the environment, reducing expected payments at the expense of decision accuracy.