Pith. sign in

REVIEW 3 major objections 4 minor 55 references

Competitive mediator games and urban CAV routing markets

T0 review · 3 major / 4 minor · reviewed 2026-08-11 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read A slightly preferred routing app takes the whole market

desk verdict A solid but narrowly scoped monopoly theorem for competitive mediators; the abstract oversells it. read the letter →

arxiv 2608.09894 v1 pith:G2Y2KJRL submitted 2026-08-10 cs.GT cs.MAecon.THmath.OC

Grzegorz Jamróz This is my paper · ORCID
classification cs.GTcs.MAecon.THmath.OC MSC 91A1090B20
keywords competitivemediatorgamemediatedequilibriumnon-atomiccongestionautonomousroutinganddrivingmarket-sharemaximizationmonopolycoarsecorrelated
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper introduces competitive mediator games, a framework in which users choose between acting directly or committing to one of several strategic mediators, and mediators choose routing or recommendation strategies to maximize their market share. It proves that in anonymous congestion games with market-share-maximizing mediators, if one mediator is weakly preferred by every user and strictly preferred by some, then every competitive mediator equilibrium is a monopoly: the dominant mediator can randomize its routing so that no matter what the other mediator does, all users prefer to delegate to the dominant one. The motivating application is future markets of autonomous routing and driving (ARAD) services, where users delegate both route choice and driving to an app. A consequence the paper draws is that a slight quality advantage can produce a monopoly, so market design should account for this when choosing mediator incentives and regulations.

What carries the argument

The central object is the competitive mediator equilibrium (CME): a Nash equilibrium of the two-stage game in which mediators first commit to recommendation patterns, defined for every possible subset of users, and then users choose a mediator or an unmediated action, with the resulting action profile forming a user equilibrium. The proof engine is the randomized unbalanced-split routing R1: mediator 1 places users who strongly prefer it on the slower route and randomly assigns the rest between the two routes in a 50-50 split, with the fast-route share w>0.5. Because the delay function is strictly increasing, any routing R2 of mediator 2 puts some users on a route with expected travel time no better than what mediator 1 offers, so those users defect. The no-HDV assumption, meaning users cannot choose to drive and route themselves, is what lets this comparison be made purely through the two mediators' induced flows.

What would settle it

Fix the two-route, two-mediator continuum game with an explicit strictly increasing continuous delay function such as t(q)=q, pick a strictly dominant discount-factor distribution with a positive-measure set of users at ratio gamma1/gamma2=1+D and the rest at ratio 1, and implement the routing R1 from Appendix D. Then exhaustively search pure routings R2, consisting of splits and assignments, for one whose induced user equilibrium keeps a positive mass on mediator 2; if any such R2 exists, Theorem 7.17 is false. A concrete candidate to test is a delay function with a nearly flat stretch between the fast and slow flows, where the strict inequalities used in the proof become numerically fragile.

Watch

Extended reading notes

Core claim

Formally, the central result is Theorem 7.17. Consider a discounted share-maximizing independent routing game with a continuum of users, two mediators, two equivalent routes, and a common strictly increasing continuous delay function. If mediator 1 is strictly dominant, meaning every user's discount factor for mediator 1 is no larger than for mediator 2 and some users strictly prefer it, then mediator 1 has a randomized routing strategy R1 such that for every routing R2 of mediator 2, the induced user equilibrium sends the whole user mass to mediator 1. Hence every competitive mediator equilibrium is a monopoly, and no non-monopoly profile can be a competitive mediator equilibrium. The construction splits mediator 1's users unevenly between the two routes with probability one half each, so that any split chosen by mediator 2 leaves some users with higher expected disutility under mediator 2; strict monotonicity of the delay function makes the comparison robust. The result is stated for the case where users have no independent-driving option and must delegate to one of the two mediators.

Load-bearing premise

The load-bearing premise is that users cannot choose to drive and route themselves: in the theorem every user must delegate to one of the two mediators, so the dominant mediator's randomized routing sees the entire flow and can make the competitor unattractive, and if independent driving remains available the monopoly lock-in need not survive.

Editorial extensions

If this is right

  • In fee-free ARAD markets where mediator revenue is market share, a provider that is weakly preferred by all users and strictly preferred by some can secure 100 percent market share in every competitive mediator equilibrium.
  • Because the dominant mediator can choose its randomized routing before users move, no routing chosen by the competitor attracts any users; the non-dominant mediator's strategy is irrelevant to the equilibrium outcome.
  • A small improvement in perceived quality, captured by discount factors, can flip the market to monopoly, so competition for quality rather than price is the margin that matters.
  • If a competitive mediator equilibrium exists, it is a monopoly; non-monopoly profiles are not CME, which sharply constrains what market designers can expect from this fee-free mechanism.
  • The paper argues that such monopolies need not be consumer-harmful because the threat of losing dominance incentivises the incumbent to keep improving service, though regulators could still add welfare terms to mediator objectives.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • If the no-HDV assumption is dropped, the monopoly conclusion may fail: a user who can drive independently along a less congested route has an outside option that the dominant mediator's randomized routing does not control, exactly the direction the paper lists as future work.
  • The same competitive-mediator formalism could be applied to other one-sided platform markets with network effects, such as ride-hailing or navigation apps, where routes are service choices and discount factors are user-specific platform preferences; the monopoly prediction would then be a testable hypothesis.
  • A natural dynamic extension is to let mediators adjust routings over time; the paper notes that the non-dominant mediator has an incentive to keep the system out of equilibrium, so CME may not be reached by learning dynamics and convergence to monopoly is not guaranteed.
  • The proof's reliance on strict monotonicity of the delay function suggests that in road networks with flat or non-monotone travel-time functions the dominant mediator's lock-in may disappear; testing the theorem on empirical or simulated delay curves would show how robust the monopoly result is.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 4 minor

Summary. The paper introduces competitive mediator games, a framework in which users either act independently or delegate their action to one of several strategic mediators, and defines competitive mediated equilibrium (CME) in both finite-player and non-atomic distributional settings. It proves existence of user equilibrium under continuity assumptions and then specializes to discounted, share-maximizing routing games. The main mathematical result, Theorem 7.17, constructs a randomized routing for a strictly dominant mediator in a two-mediator, two-route, no-independent-driving (no-HDV) game with strictly increasing delay, and shows that this routing makes the dominant mediator the unique user-optimal choice for all users against any rival routing, yielding a monopoly CME. The paper also discusses CAV market structures, interprets the result as a tendency toward monopoly, and lists several open problems, including extension to settings with viable independent driving.

Significance. Read under its stated hypotheses, Theorem 7.17 is an interesting and nontrivial result: a small quality advantage, combined with a carefully randomized routing, can enforce a monopoly equilibrium even when a large fraction of users are close to indifferent. The proof is detailed, the algebra in Propositions 7.15 and 7.19 checks out, and the framework has no fitted parameters, so the core result appears sound. The distributional formulation and the explicit discussion of its limitations are also useful contributions. However, the abstract and Section 9 advertise a much broader monopoly claim that omits the theorem's load-bearing assumptions (no HDV, two equivalent routes, strict dominance, exactly two mediators). Since the paper itself identifies viable independent driving as future work, the advertised claim is not supported by the present proof.

major comments (3)
  1. [Abstract and Section 9 (Discussion)] The abstract claims that 'in the generic setting of anonymous congestion(routing) games with market-share maximizing mediators all competitive mediator equilibria are monopolies whenever one of the mediators is weakly preferred to other mediators by all users.' Theorem 7.17 does not prove this: it assumes no HDV option, exactly two equivalent routes, strict dominance, and two mediators (Remark 7.14). The no-HDV condition is not cosmetic: in Appendix D, the normalization step explicitly states that it 'will no longer be possible with the independent choice (HDV) mode available.' With independent driving, a deviating user can choose the less congested route, so the deviation argument used in case iii of the proof fails. The abstract and the 'main conclusion' paragraph in Section 9 must carry the theorem's qualifiers or supply an additional proof for the broader claim.
  2. [Theorem 7.17 and Appendix D] The no-HDV assumption appears only in the theorem's heading and in Remark 7.14, not in the formal statement in Section 7.3. This matters because Definition 6.5 and Definition 6.14 define user strategies over A∪F, which includes the option of independent route choice. The proof's case iii argument, that a deviating user to mediator 2 faces an expected travel time equal to the average travel time and independent of mediator 2's routing, relies on the user not being able to choose a route himself. If independent actions are allowed, the user equilibrium condition must also consider deviations to A, and the claimed dominance of mediator 1 is no longer established. The theorem should be restated with the restriction on the users' action sets made explicit, and the abstract should not imply that the result covers independent-driving settings.
  3. [Abstract versus Definition 7.13 and Theorem 7.17] The abstract's condition that one mediator is 'weakly preferred to other mediators by all users' is insufficient; Theorem 7.17 requires strict dominance, i.e., strict preference for a non-negligible set of users (Definition 7.13(ii)). Section 9 correctly adds 'strongly preferred by some of the users,' but the abstract omits this. The distinction is load-bearing: the proof constructs a set IA of users with ratio γ1_i/γ2_i > 1+D for some D>0 and uses the positive measure of this set in the inequalities (13)-(15) of Appendix D. If all users are exactly indifferent between the mediators, the theorem's argument collapses, and the paper gives no reason to believe the monopoly conclusion holds. The abstract should state the strict-dominance requirement.
minor comments (4)
  1. [Definition 7.8] In the definitions of the natural basins B0 and Bf, the tuple is written as (γ1,γ2,...,γN), but the discount factor distribution lives on [0,∞)^|F|, so the tuple should be (γ1,...,γF) with |F| mediators.
  2. [Example 4.1, Eq. (1)] The first integral in the expression for USO is written with limits from P^{-1}(q2) to 0, which appears to be a typo; the intended integral is presumably over γ from 0 to P^{-1}(q2), with the second integral from P^{-1}(q2) to ∞.
  3. [Abstract] The sentence beginning 'which have become a popular research area recently as they not only can be more socially efficient...' has an ambiguous antecedent for 'which'; it likely refers to (coarse) correlated equilibria, but the syntax should be clarified.
  4. [Introduction, page 2] There is a typo: 'slighltly' should be 'slightly.'

Circularity Check

0 steps flagged · score 1.0 of 10

No circularity: Theorem 7.17 follows from stated assumptions; the abstract's monopoly claim outruns the theorem's no-HDV/two-route qualifiers, which the paper itself flags.

full rationale

The paper contains no fitted parameters and no prediction that reduces to its inputs. Theorem 7.17's monopoly claim is a constructive result: starting from strict dominance, the proof partitions users into IA (ratio > 1+D) and IB, defines R1 as a 50-50 randomized split with q_fast < q_slow, and verifies by explicit inequalities (Eqs. 15, 11-12 and Appendix D) that for any deterministic or stochastic R2 a positive-measure set of users strictly prefers mediator 1, so a user equilibrium in which anyone uses mediator 2 is impossible; the CME conclusion then follows because mediator 1 already holds the maximum feasible market share. The 'normalization' to (1,1+D)/(1,1) is a monotone rescaling of relative disutilities, not an import of the conclusion, and the proof explicitly notes it fails when HDV independent routing is added — a scope limitation that the paper itself flags in Appendix D and Section 9 ('Extension of the monopoly results to settings with viable independent driving (HDV)'). The abstract's broad phrasing omits these qualifiers, which is a correctness/scope concern rather than circularity. The only author self-citation, [28] (Jamróz et al.), motivates the fee-free market-share model and the mature-market interpretation; it is not cited in the proof of Theorem 7.17, and no uniqueness theorem or ansatz is imported from it. Score 1 rather than 0 only acknowledges the presence of that non-load-bearing self-citation; it does not indicate circularity of the derivation.

Assumptions & free parameters 0 free parameters · 6 assumptions · 0 invented entities

The central theorem is a pure theory result built on standard fixed-point and equilibrium existence theorems plus domain assumptions about routing games. No parameters are fitted to data; the constants w, epsilon, and D in the constructed strategies are proof-level design choices constrained by inequalities, not empirical fits. The paper introduces no new physical or ontological entities.

assumptions (6)
  • standard math Nash's existence theorem for finite games
    Used in Proposition 6.7 to prove existence of user equilibrium in finite-player competitive mediator games.
  • standard math Mas-Colell's Theorem 1 on equilibrium existence in non-atomic games
    Invoked in Section 6.1 and Theorem 6.19 as the non-atomic analog in the distributional formulation.
  • standard math Kakutani-Glicksberg-Ky Fan fixed point theorem
    Used in Appendix A in the proof of Theorem 6.19 to obtain a fixed point of the best-response correspondence.
  • domain assumption Delay function t is strictly increasing and continuous
    Theorem 7.17 and Propositions 7.15, 7.19 rely on monotonicity of t to compare travel times on the two routes.
  • domain assumption No independent driving (HDV) option in Theorem 7.17
    The monopoly proof requires users to choose between mediators only; the abstract omits this restriction, and Section 9 lists its removal as future work.
  • domain assumption Mediators maximize market share and users are non-atomic
    Definition 7.1 and Section 6.1; the continuum approximation makes user deviations infinitesimal and lets the proof use average comparisons.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Competitive mediator games and urban CAV routing markets." pith.science (2026). https://pith.science/paper/G2Y2KJRL

@misc{pith2026260809894,
  author       = {Pith},
  title        = {Pith review of: Competitive mediator games and urban CAV routing markets},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/G2Y2KJRL}},
  note         = {Machine review of arXiv:2608.09894}
}
read the original abstract

Inspired by possible future markets of autonomous routing and driving (ARAD), we introduce competitive mediator games and their equilibria which generalize the (coarse) correlated equilibria, which have become a popular research area recently as they not only can be more socially efficient than Nash equilibria but also are limits of algorithmic no-regret multi-agent learning dynamics. We discuss the basic properties of competitive mediator games and prove that in the generic setting of anonymous congestion(routing) games with market-share maximizing mediators all competitive mediator equilibria are monopolies whenever one of the mediators is weakly preferred to other mediators by all users. We apply and interpret these results in the context of new markets of competing ARAD service providers. We also provide a comprehensive overview of these markets and discuss the future mechanism design thereof.

Figures

Figures reproduced from arXiv: 2608.09894 by the authors.

Figure 1
Figure 1. Distribution of discount factors (relative value of time in AV compared to HDV) in [PITH_FULL_IMAGE:figures/full_fig_p009_1.png] view at source ↗
Figure 2
Figure 2. Equilibria in multi-agent games with increasing level of mediation complexity. When [PITH_FULL_IMAGE:figures/full_fig_p012_2.png] view at source ↗
Figure 3
Figure 3. Natural basins of belonging in a discounted competitive mediator game. E.g. a user [PITH_FULL_IMAGE:figures/full_fig_p025_3.png] view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: ii) Mediator fd is called strictly dominant if fd is weakly dominant and for a non-negligible (positive measure) set of users Id ⊂ N we have γ fd i < γf i for every i ∈ Id and f ∈ F\{fd}. Remark 7.14. Below, we consider settings with two mediators only and no independe…
Figure 4
Figure 4. Figure 4: Left. Admissible support of discount factor distribution when mediator 1 is dominant [PITH_FULL_IMAGE:figures/full_fig_p027_4.png]
Figure 5
Figure 5. Figure 5: The shaded areas indicate the users which are forfeited by mediator [PITH_FULL_IMAGE:figures/full_fig_p036_5.png]
Figure 6
Figure 6. Figure 6: Two possible splits between two routes induced by routings of mediator 1 (gray), [PITH_FULL_IMAGE:figures/full_fig_p037_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

55 extracted references · 55 canonical work pages

  1. [1]

    (2023).Combining human expertise with artificial intelligence: Experimental evidence from radiology(No

    Agarwal, N., Moehring, A., Rajpurkar, P., and Salz, T. (2023).Combining human expertise with artificial intelligence: Experimental evidence from radiology(No. w31422). National Bureau of Economic Research

    work page 2023

  2. [2]

    (2024).Algorithmic pricing and competition: Empirical evidence from the German retail gasoline market.Journal of Political Economy, 132(3), 723-771

    Assad, S., Clark, R., Ershov, D., and Xu, L. (2024).Algorithmic pricing and competition: Empirical evidence from the German retail gasoline market.Journal of Political Economy, 132(3), 723-771

    work page 2024

  3. [3]

    (2007, June)

    Ashlagi, I., Monderer, D., and Tennenholtz, M. (2007, June). Mediators in position auctions. In Proceedings of the 8th ACM conference on Electronic commerce (pp. 279-287)

    work page 2007

  4. [4]

    Aumann, R. J. (1974).Subjectivity and correlation in randomized strategies.Journal of mathematical Economics, 1(1), 67-96

    work page 1974

  5. [5]

    Aumann, R. J. (1987).Correlated equilibrium as an expression of Bayesian rationality.Econometrica: Journal of the Econometric Society, 1-18

    work page 1987

  6. [6]

    (2016).Mechanism design with strategic mediators.ACM Transac- tions on Economics and Computation (TEAC), 4(2), 1-48

    Babaioff, M., Feldman, M., and Tennenholtz, M. (2016).Mechanism design with strategic mediators.ACM Transac- tions on Economics and Computation (TEAC), 4(2), 1-48

    work page 2016

  7. [7]

    B., and Winsten, C

    Beckmann, M., McGuire, C. B., and Winsten, C. B. (1956). Studies in the Economics of Transportation, Yale Uni- versity Press, New Haven, CT

    work page 1956

  8. [8]

    Theorie mathematique de la richesse sociale

    Bertrand, J. (1883).Review of "Theorie mathematique de la richesse sociale" and of "Recherches sur les principles mathematiques de la theorie des richesses."Journal de savants, 67, 499. 30

    work page

Show all 55 references
  1. [9]

    Calder-Wang, S., and Kim, G. H. (2024).Algorithmic pricing in multifamily rentals: Efficiency gains or price coor- dination?Available at SSRN 4403058

    2024

  2. [10]

    H., Looff, E., van Cranenburgh, S., Snelder, M., and van Arem, B

    de Almeida Correia, G. H., Looff, E., van Cranenburgh, S., Snelder, M., and van Arem, B. (2019). On the impact of vehicle automation on the value of travel time while performing work and leisure activities in a car: Theoretical insights and results from a stated preference sur...

    2019

  3. [11]

    W., and Papadimitriou, C

    Daskalakis, C., Goldberg, P. W., and Papadimitriou, C. H. (2009).The complexity of computing a Nash equilibrium. Communications of the ACM, 52(2), 89-97

    2009

  4. [12]

    (2008).A multiagent approach to autonomous intersection management.Journal of artificial intelligence research, 31, 591-656

    Dresner, K., and Stone, P. (2008).A multiagent approach to autonomous intersection management.Journal of artificial intelligence research, 31, 591-656

    2008

  5. [13]

    (2025).Algorithmic Pricing and Algorithmic Collusion.Business & Information Systems Engineering, 67(6), 971-979

    Durmann, J., Oberlechner, M., and Bichler, M. (2025).Algorithmic Pricing and Algorithmic Collusion.Business & Information Systems Engineering, 67(6), 971-979

    2025

  6. [14]

    Forges, F. (1986). An approach to communication equilibria. Econometrica: Journal of the Econometric Society, 1375-1385

    1986

  7. [15]

    Feldman, J., Mirrokni, V., Muthukrishnan, S., and Pai, M. M. (2010, June).Auctions with intermediaries.In Pro- ceedings of the 11th ACM conference on Electronic commerce (pp. 23-32)

    2010

  8. [16]

    P., and Vohra, R

    Foster, D. P., and Vohra, R. V. (1997).Calibrated learning and correlated equilibrium.Games and Economic Behavior, 21(589), 40-55

    1997

  9. [17]

    Fudenberg, D., and Levine, D. K. (1999).Conditional universal consistency.Games and Economic Behavior, 29(1-2), 104-130

    1999

  10. [18]

    (2018, July)

    Gan, J., Elkind, E., and Wooldridge, M. (2018, July). Stackelberg security games with multiple uncoordinated defenders. In Proceedings of the 17th international conference on autonomous agents and multiagent systems. ACM Press

    2018

  11. [19]

    Garcia, D., Tolvanen, J., and Wagner, A. K. (2026). Strategic responses to algorithmic recommendations: evidence from hotel pricing. Management Science, 72(1), 609-626

    2026

  12. [20]

    Geffner, I., Karpas, E., and Tennenholtz, M. (2025). When Competition Helps: Achieving Optimal Traffic Flow with Multiple Autonomous Planners. arXiv preprint arXiv:2508.07145

    2025 arXiv

  13. [21]

    (1989).Nash and correlated equilibria: Some complexity considerations.Games and Eco- nomic Behavior, 1(1), 80-93

    Gilboa, I., and Zemel, E. (1989).Nash and correlated equilibria: Some complexity considerations.Games and Eco- nomic Behavior, 1(1), 80-93

    1989

  14. [22]

    (2000).A simple adaptive procedure leading to correlated equilibrium.Econometrica, 68(5), 1127-1150

    Hart, S., and Mas-Colell, A. (2000).A simple adaptive procedure leading to correlated equilibrium.Econometrica, 68(5), 1127-1150

    2000

  15. [23]

    R., Parsley, H., Schwieg, T., and Williams, K

    Hortaçsu, A., Natan, O. R., Parsley, H., Schwieg, T., and Williams, K. R. (2024).Organizational structure and pricing: Evidence from a large us airline.The Quarterly Journal of Economics, 139(2), 1149-1199

    2024

  16. [24]

    Huang, Y. (2025). Pricing frictions and platform remedies: the case of Airbnb. Available at SSRN 3767103

    2025

  17. [25]

    P., and Caines, P

    Huang, M., Malhamé, R. P., and Caines, P. E. (2006). Large population stochastic dynamic games: closed-loop McKean-Vlasov systems and the Nash certainty equivalence principle

    2006

  18. [26]

    Ivanov, D., Zisman, I., and Chernyshev, K. (2023). Mediated multi-agent reinforcement learning. arXiv preprint arXiv:2306.08419

    2023 arXiv

  19. [27]

    A., Masoomi, H., Fiondella, L., and Mosleh, A

    Jafary, B., Rabiei, E., Diaconeasa, M. A., Masoomi, H., Fiondella, L., and Mosleh, A. (2018, September).A survey on autonomous vehicles interactions with human and other vehicles.In 14th PSAM International Conference on Probabilistic Safety Assessment and Management. Los Angel...

    2018

  20. [28]

    and Kucharski, R

    Jamróz, G., Gorczyca, Ł. and Kucharski, R. (2026)Randomized routing strategies of fleets of CAVs may prove market efficient.Accepted to 14th Symposium of the European Association for Research in Transportation (hEART 2026). https://arxiv.org/abs/2607.14859

    2026 arXiv

  21. [29]

    and Gentzkow, M

    Kamenica, E. and Gentzkow, M. (2011).Bayesian persuasion.American Economic Review, 101(6), 2590-2615

    2011

  22. [30]

    M., and Lions, P

    Lasry, J. M., and Lions, P. L. (2007).Mean field games.Japanese journal of mathematics, 2(1), 229-260

    2007

  23. [31]

    (1978).On generalized Stackelberg strategies.Journal of optimization theory and applications, 26(4), 637-643

    Leitmann, G. (1978).On generalized Stackelberg strategies.Journal of optimization theory and applications, 26(4), 637-643

    1978

  24. [32]

    Mas-Colell, A. (1984). On a theorem of Schmeidler. Journal of Mathematical Economics, 13(3), 201-206

    1984

  25. [33]

    Maskin, E., and Tirole, J. (1988). A theory of dynamic oligopoly, II: Price competition, kinked demand curves, and Edgeworth cycles. Econometrica: Journal of the Econometric Society, 571-599

    1988

  26. [34]

    Monderer, D., and Tennenholtz, M. (2009). Strong mediated equilibrium. Artificial Intelligence, 173(1), 180-195

    2009

  27. [35]

    Morandi, V. (2024). Bridging the user equilibrium and the system optimum in static traffic assignment: a review. 4or, 22(1), 89-119

    2024

  28. [36]

    Moulin, H., and Vial, J. P. (1978).Strategically zero-sum games: the class of games whose completely mixed equilibria cannot be improved upon.International Journal of Game Theory, 7(3-4), 201-221

    1978

  29. [37]

    Nash, J. F. (1951).Non-cooperative gamesThe Annals of Mathematics, 54(2), 286-295

    1951

  30. [38]

    (2024).Mediated collusion.Journal of Political Economy, 132(4), 1247-1289

    Ortner, J., Sugaya, T., and Wolitzky, A. (2024).Mediated collusion.Journal of Political Economy, 132(4), 1247-1289

    2024

  31. [39]

    Procaccia, Ariel D., and Moshe Tennenholtz.Approximate mechanism design without money.ACM Transactions on Economics and Computation (TEAC) 1.4 (2013): 1-26

    2013

  32. [40]

    Rietveld, J., and Schilling, M. A. (2021). Platform competition: a systematic and interdisciplinary review of the literature. Journal of Management, 47(6), 1528-1563

    2021

  33. [41]

    C., and Tirole, J

    Rochet, J. C., and Tirole, J. (2003).Platform competition in two-sided markets.Journal of the european economic association, 1(4), 990-1029

    2003

  34. [42]

    (2007, January).Routing Mediators.In IJCAI (pp

    Rozenfeld, O., and Tennenholtz, M. (2007, January).Routing Mediators.In IJCAI (pp. 1488-1493)

    2007

  35. [43]

    Shladover, S. E. (2018).Connected and automated vehicle systems: Introduction and overview.Journal of Intelligent Transportation Systems, 22(3), 190-200

    2018

  36. [44]

    Schmeidler, D. (1973). Equilibrium points of nonatomic games. Journal of statistical Physics, 7(4), 295-300

    1973

  37. [45]

    Smith, M. J. (1979).The existence, uniqueness and stability of traffic equilibria.Transportation Research Part B: Methodological, 13(4), 295-304

    1979

  38. [46]

    Talebpour, A., and Mahmassani, H. S. (2016). Influence of connected and autonomous vehicles on traffic flow stability and throughput. Transportation research part C: emerging technologies, 71, 143-163

    2016

  39. [47]

    (2015, July).Implementing the Wisdom of Waze.In IJCAI (Vol

    Vasserman, S., Feldman, M., and Hassidim, A. (2015, July).Implementing the Wisdom of Waze.In IJCAI (Vol. 15, pp. 660-666)

    2015

  40. [48]

    Wardrop, J. G. (1952). Road paper. Some theoretical aspects of road traffic research. Proceedings of the institution of civil engineers, 1(3), 325-362

    1952

  41. [49]

    Zhang, Y., Liu, F., Wang, Z., Chen, Y., Feng, S., Wu, Q., and Hou, Y. (2022). On Nash–Stackelberg–Nash games under decision-dependent uncertainties: Model and equilibrium. Automatica, 142, 110401. 32

    2022

  42. [50]

    Act of July 2, 1890(Sherman Anti-Trust Act), July 2, 1890; Enrolled Acts and Resolutions of Congress, 1789-1992; General Records of the United States Government; Record Group 11; National Archives

    1992

  43. [51]

    YouGov,https://yougov.com/en-gb/articles/53188-do-britons-trust-driverless-taxis-and-autonomous-veh icles, Accessed 5th May 2026

    2026

  44. [52]

    2:23-cv-01495-JHC, U.S

    Federal Trade Commission et al., Second Amended Complaint against Amazon.com, Inc., No. 2:23-cv-01495-JHC, U.S. District Court for the Western District of Washington (Sept. 2023; second amended filing 2024). A Proof of Theorem 6.19 Before we proceed to the proof of Theorem 6.1...

    2023

  45. [53]

    To see this, consider two examples

  46. [54]

    Then, disutilitiesu 1,u 2,(9)–(10)become u1 γ = (1−w)τ 1 +wτ 4, u2 γ =sτ 1 + (1−s)τ 4

    Consider the extreme casep= 1. Then, disutilitiesu 1,u 2,(9)–(10)become u1 γ = (1−w)τ 1 +wτ 4, u2 γ =sτ 1 + (1−s)τ 4. Clearly,u 2<u 1 whenevers<1−wasτ 1>τ 4. 38

  47. [55]

    Let us compute, for which range ofpthis is a Nash user equilibrium

    Consider mediator2with zero market share, i.e.σ2 B = 0. Let us compute, for which range ofpthis is a Nash user equilibrium. Forσ2 B = 0we haveτ 1 =τ 3 andτ 2 =τ 4 andτ 1>τ 4. Disutilities(9)–(10)become u1 γ = (1−w)[pτ 1 + (1−p)τ 3] +w[pτ 4 + (1−p)τ 2] = (1−w)τ 1 +wτ 4, u2 γ =s...

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.