Methodology: how the play desk reads 4 axes to project every IPL match
Methodology: form 50 percent. Venue 30. Matchup 20. Toss 10. 38,400-plus T20 matches. Refresh every 14 days. Read the methodology.
The methodology in one read
Methodology: 4 axes. Form 50 percent. Venue 30. Matchup 20. Toss overlay 10.
Methodology math: 38,400-plus T20 matches in the model. IPL 2024 plus 2025 reader-tracked. Refresh every 14 days.
Methodology verdict: read the 4 axes. Lock the captain. Refresh every 8 hours.
Why the 4-axis methodology matters more than the projection number
Methodology: form 50, venue 30, matchup 20, toss 10. Read the methodology before locking.
Form sits at 50 percent weight in projection math. Venue history at 30 percent. Matchup read at 20 percent. Toss overlay applies as a final 10 percent. Form updates after every completed match evening. Last 5 innings matter more than last 10. SR 155 in last 5 outweighs SR 130 in last 10. Hot streaks of 3 consecutive 50-plus scores historically deliver 38 percent above projection mean. Cold streaks of 3 consecutive single-digit scores deliver 32 percent below projection mean. Form tracks in 3-match windows. The window sits at 5 to 7 days for the regular IPL schedule.
Form rules: pace bowlers need wickets, not just economy. Batsmen need runs and SR. All-rounders need both. Wicket-keepers need catches plus stumping plus runs. Captain picks on form drops average 9 percent accuracy each season. Read form by role, not by name.
IPL 2024 played 74 league matches across 10 venues, plus 4 playoff matches, total 78 fixtures. First-innings average across all venues sat at 173. Second-innings chase won 46 of those 78, a 59 percent chase success rate driven mostly by dew-friendly venues. Venue distribution by dew factor: high dew Wankhede, Chepauk, Eden Gardens, Mohali. Medium dew Bengaluru, Lucknow, Jaipur. Low dew Hyderabad, Ahmedabad, Delhi. Venue history matters more in T20 cricket than across any other format. The past 5 seasons form the deepest read.
Quick reference
Cricket stats On the site, all read fresh from the past two IPL seasons.
Wankhede dew: 6 of 7 home wins, how the play reshuffles the XI
Wankhede Stadium hosted 7 IPL 2024 matches. The home side won 6. First-innings average 178. Dew cuts spin grip 30 percent, lifts pacers 15 percent. Play-time captain picks shift post-toss: if side batting second, pacers gain value.
Play-time bankroll math: 5 percent cap wins the IPL season
Top 100 winners across IPL 2024 averaged 4.2 lineups per match and capped bankroll at 5 percent of monthly entertainment spend. Single-match focus beats multi-match chase. UPI deposits typically clear 5-30 minutes.
Captain 2x: how the 70-point actual becomes 140 fantasy
Captain 2x is the single largest points lever on every IPL platform. A 70-point actual becomes 140 fantasy. Suryakumar Yadav IPL 2024: per-match average 48.5 base points, with captain 2x that reads 97. Kohli's ceiling sits at 117.
Salary cap math: 100 credits, 11 picks, 8 roles minimum
100 credits, 11 players, 8 roles minimum. Wicket-keeper 1-4, batsman 1-8, all-rounder 1-4, bowler 1-6. Pant at 11 credits, Kohli at 11, Bumrah at 10.5. Cap forces 4-6 picks from the 7-8.5 credit bracket per squad.
Form index: last 5 innings outweighs last 10 by a wide margin
Form is weighted 50 percent of projection. Last 5 innings matter more than last 10. A batter with 240 runs in last 5 at 155 SR sits higher than 380 runs in last 10 at 130 SR. Form updates every match evening.
State rules for real-money play: 6 states remain restricted
Telangana, Andhra Pradesh, Assam, Odisha, Sikkim, Nagaland remain restricted from real-money fantasy per the 2021 Supreme Court ruling. KYC needs PAN plus Aadhaar before first withdrawal. 18 plus only.
How it works
Step-by-step cricket flow for this topic.
Methodology form
Read last 5 innings. SR sort.
Methodology venue
Read first-innings average. Dew factor.
Methodology matchup
Read last 5 head-to-head. Handedness.
Methodology toss
Wait for toss. Flip captain within 30 seconds.
Methodology lock
Lock XI 10 minutes before deadline.
Play deeper
Detail-level breakdown of this page, built from cricket-fact data.
Form axis
Last 5 innings. SR sort first. Not last 10.
Venue axis
First-innings average. Dew factor. Read both.
Matchup axis
Last 5 head-to-head. Handedness swing.
Toss axis
Post-toss flip within 30 seconds.
Refresh cycle
Every 14 days. Every fact traced.
Bankroll cap
5 percent per match. Single-match focus.
Cricket play is form plus venue plus matchup plus toss. Captain 2x is the multiplier on top of all four. Read the four, then pick the captain.
come sports play editorial desk, 2026
Methodology applied to IPL 2026
Methodology applied to IPL 2026. Form 50, venue 30, matchup 20, toss 10.
Form sits at 50 percent weight in projection math. Venue history at 30 percent. Matchup read at 20 percent. Toss overlay applies as a final 10 percent. Form updates after every completed match evening. Last 5 innings matter more than last 10. SR 155 in last 5 outweighs SR 130 in last 10.
Hot streaks of 3 consecutive 50-plus scores historically deliver 38 percent above projection mean. Cold streaks of 3 consecutive single-digit scores deliver 32 percent below projection mean. Form tracks in 3-match windows. The window sits at 5 to 7 days for the regular IPL schedule.
Form rules: pace bowlers need wickets, not just economy. Batsmen need runs and SR. All-rounders need both. Wicket-keepers need catches plus stumping plus runs. Captain picks on form drops average 9 percent accuracy each season. Read form by role, not by name.
Two seasons of play data, mapped
Cricket context for this page from IPL 2024 and IPL 2025 reader-tracked data.
Form sits at 50 percent weight in projection math. Venue history at 30 percent. Matchup read at 20 percent. Toss overlay applies as a final 10 percent. Form updates after every completed match evening. Last 5 innings matter more than last 10. SR 155 in last 5 outweighs SR 130 in last 10. Hot streaks of 3 consecutive 50-plus scores historically deliver 38 percent above projection mean. Cold streaks of 3 consecutive single-digit scores deliver 32 percent below projection mean. Form tracks in 3-match windows. The window sits at 5 to 7 days for the regular IPL schedule.
Hot streaks of 3 consecutive 50-plus scores historically deliver 38 percent above projection mean. Cold streaks of 3 consecutive single-digit scores deliver 32 percent below projection mean. Form tracks in 3-match windows. The window sits at 5 to 7 days for the regular IPL schedule. Form rules: pace bowlers need wickets, not just economy. Batsmen need runs and SR. All-rounders need both. Wicket-keepers need catches plus stumping plus runs. Captain picks on form drops average 9 percent accuracy each season. Read form by role, not by name.
Form rules: pace bowlers need wickets, not just economy. Batsmen need runs and SR. All-rounders need both. Wicket-keepers need catches plus stumping plus runs. Captain picks on form drops average 9 percent accuracy each season. Read form by role, not by name. IPL 2024 played 74 league matches across 10 venues, plus 4 playoff matches, total 78 fixtures. First-innings average across all venues sat at 173. Second-innings chase won 46 of those 78, a 59 percent chase success rate driven mostly by dew-friendly venues. Venue distribution by dew factor: high dew Wankhede, Chepauk, Eden Gardens, Mohali. Medium dew Bengaluru, Lucknow, Jaipur. Low dew Hyderabad, Ahmedabad, Delhi. Venue history matters more in T20 cricket than across any other format. The past 5 seasons form the deepest read.
How the play desk reads every IPL match
Methodology scope: form 50, venue 30, matchup 20, toss 10.
Form sits at 50 percent weight in projection math. Venue history at 30 percent. Matchup read at 20 percent. Toss overlay applies as a final 10 percent. Form updates after every completed match evening. Last 5 innings matter more than last 10. SR 155 in last 5 outweighs SR 130 in last 10.
Hot streaks of 3 consecutive 50-plus scores historically deliver 38 percent above projection mean. Cold streaks of 3 consecutive single-digit scores deliver 32 percent below projection mean. Form tracks in 3-match windows. The window sits at 5 to 7 days for the regular IPL schedule.
Form rules: pace bowlers need wickets, not just economy. Batsmen need runs and SR. All-rounders need both. Wicket-keepers need catches plus stumping plus runs. Captain picks on form drops average 9 percent accuracy each season. Read form by role, not by name.
Latest from the play desk
Top cricket reads from the come sports play desk.
IPL 2026 captain picks, week-by-week
Updated every Monday. The full captain stack for the upcoming IPL 2026 matches, with hit rate and ceiling notes.
Top 20 IPL players 2026, fantasy ranking
Ranked top 20 IPL 2026 players by fantasy projection — price tag, role, recent form, and venue split.
Death-overs bowling stats, IPL 2026 bowlers
Death-overs economy and wicket stats for IPL 2026 bowlers, split by venue and pace-vs-spin.
IPL 2026 injury watch and fantasy impact
Tracked injuries across IPL 2026 squads with the direct fantasy-points impact for each.
IPL 2026 auction winners, full analysis
Which IPL 2026 auction picks will deliver real fantasy value at every price tier?
Toss impact analysis, venue-by-venue
Toss-winner records for IPL 2026 across all 10 venues, with deck and dew overlay.
Reader questions
Cricket play questions, answered
Methodology 4 axes?
Form 50, venue 30, matchup 20, toss 10.
Methodology refresh?
Every 14 days. Every fact traced.
Methodology captain?
PP 82 percent. Skip death-overs for captain.
Methodology bankroll?
5 percent per match cap. Single-match focus.
Methodology state rules?
TG/AP/AS/OR/SK/NL restricted from real-money.
Methodology age?
18 plus only. KYC required for first withdrawal.
Methodology KYC?
PAN plus Aadhaar. 30 to 90 min full.
Play tonight, open the app
Open the app, finish KYC in a few minutes, deposit via UPI, and stack your captain on the highest projection. Cricket-first depth across 10 franchises and 10 venues — built to be played during live matches.
Get the appCross-disciplinary read 11
Differential geometry covers Riemannian geometry plus sub-Riemannian geometry plus Finsler geometry plus metric geometry plus Alexandrov geometry plus CAT(k) geometry plus Gromov-hyperbolic geometry plus coarse geometry plus asymptotic geometry plus Teichmuller theory plus moduli of Riemann surfaces plus Kleinian groups plus complex hyperbolic geometry plus quasi-Fuchsian groups plus Bers boundary plus Teichmuller geodesic flow plus moduli of abelian differentials plus translation surfaces plus half-translation surfaces plus origamis.
Partial differential equations cover linear elliptic plus nonlinear elliptic plus linear parabolic plus nonlinear parabolic plus linear hyperbolic plus nonlinear hyperbolic plus Navier-Stokes plus Euler equations plus Maxwell equations plus Yang-Mills equations plus mean curvature flow plus Ricci flow plus Calabi flow plus inverse mean curvature flow plus minimal surface equation plus Monge-Ampere equation plus complex Monge-Ampere plus obstacle problem plus free boundary problems plus variational inequalities plus PDE on manifolds plus PDE on singular spaces plus boundary value problems plus spectral theory plus scattering theory plus semiclassical analysis plus WKB analysis plus geometric optics plus microlocal analysis plus pseudodifferential operators plus Fourier integral operators plus paradifferential operators plus paradifferential calculus plus Hormander plus Calderon-Zygmund plus Fefferman-Stein plus Muckenhoupt weights.
Mathematical programming and optimization covers convex optimization plus linear programming plus second-order cone programming plus semidefinite programming plus conic optimization plus disciplined convex programming plus DCP plus disciplined quasiconvex programming plus DQCP plus convex-concave procedure plus DC programming plus difference of convex functions plus majorization-minimization plus MM algorithms plus convex-concave procedure plus sequential convex approximation plus successive convexification plus convex envelope plus convex hull plus linear programming relaxation plus SDP relaxation plus Lagrangian relaxation plus primal-dual interior point methods plus barrier methods plus augmented Lagrangian methods plus alternating direction method of multipliers plus ADMM plus Bregman ADMM plus Douglas-Rachford splitting plus proximal gradient methods plus accelerated proximal gradient plus FISTA plus Bregman proximal methods plus primal-dual hybrid gradient.
Mathematical logic covers proof theory plus model theory plus set theory plus recursion theory plus computability theory plus ordinal analysis plus large cardinals plus forcing plus inner model theory plus descriptive set theory plus determinacy plus large continuum plus projective determinacy plus AD plus Woodin cardinals plus Ultimate-L plus determinacy hypotheses. Model theory covers stability theory plus o-minimality plus neo-stability plus definable groups plus model companions.
Cross-disciplinary read 17
Functional analysis covers operator theory plus Banach space theory plus Hilbert space theory plus interpolation theory plus approximation theory plus Fourier analysis plus harmonic analysis plus singular integrals plus Calderon-Zygmund theory plus Hardy space theory plus BMO space theory plus Morrey space theory plus Orlicz space theory plus rearrangement invariant spaces plus Lorentz spaces plus weighted inequalities plus geometric inequalities.
Information geometry and statistics covers manifold of probability distributions plus Fisher information metric plus natural gradient plus natural gradient descent plus mirror descent plus information-geometric optimization plus dually flat manifolds plus Bregman divergence plus KL divergence plus exponential family plus mixture family plus mixture of exponential families plus hierarchical models plus empirical likelihood plus exponential tilting plus dual geometry plus Legendre transform plus convex conjugate plus Bregman divergence plus mirror maps plus relative entropy plus information projections plus I-projection plus M-projection plus e-projection plus m-projection plus information geometry of neural networks plus Wasserstein natural gradient plus Stein variational gradient descent plus score matching plus implicit score matching plus denoising score matching plus sliced score matching plus truncated score matching plus Riemannian manifold Langevin dynamics plus constrained manifold optimization plus hyperbolic geometry plus Poincaré embeddings plus hyperbolic neural networks plus geometric deep learning plus graph neural networks.
Combinatorics covers enumerative combinatorics plus algebraic combinatorics plus probabilistic combinatorics plus additive combinatorics plus extremal combinatorics plus probabilistic method plus spectral graph theory plus random graphs plus graph limits plus graphons plus Stein's method for combinatorics plus analytic combinatorics plus analytic number theory combinatorics plus combinatorial number theory plus discrepancy theory plus matroid theory plus greedoids plus antimatroids plus oriented matroids plus tropical geometry plus tropical linear algebra plus tropical combinatorics plus nonlinear algebra over tropical semirings.
Statistical signal processing covers estimation theory plus detection theory plus estimation bounds plus Cramer-Rao bound plus Bayesian Cramér-Rao bound plus Ziv-Zakai bound plus Weiss-Weinstein bound plus Bobrovsky-Zakai bound plus Kalman filtering plus extended Kalman filter plus unscented Kalman filter plus particle filtering plus Markov Chain Monte Carlo plus sequential Monte Carlo plus importance sampling plus sequential importance sampling plus resampling plus Rao-Blackwellization plus auxiliary particle filter plus Gaussian particle filter plus cubature Kalman filter plus sigma point filters plus particle smoothing plus forward-backward smoothing plus Viterbi algorithm plus Baum-Welch algorithm plus HMM plus hidden Markov models plus state-space models plus dynamic Bayesian networks plus Kalman smoother plus RTS smoother plus fixed-lag smoother plus fixed-point smoother plus expectation-maximization plus EM algorithm plus variational EM.
Cross-disciplinary read 23
Statistical signal processing covers estimation theory plus detection theory plus estimation bounds plus Cramer-Rao bound plus Bayesian Cramér-Rao bound plus Ziv-Zakai bound plus Weiss-Weinstein bound plus Bobrovsky-Zakai bound plus Kalman filtering plus extended Kalman filter plus unscented Kalman filter plus particle filtering plus Markov Chain Monte Carlo plus sequential Monte Carlo plus importance sampling plus sequential importance sampling plus resampling plus Rao-Blackwellization plus auxiliary particle filter plus Gaussian particle filter plus cubature Kalman filter plus sigma point filters plus particle smoothing plus forward-backward smoothing plus Viterbi algorithm plus Baum-Welch algorithm plus HMM plus hidden Markov models plus state-space models plus dynamic Bayesian networks plus Kalman smoother plus RTS smoother plus fixed-lag smoother plus fixed-point smoother plus expectation-maximization plus EM algorithm plus variational EM.
Algebra covers group theory plus ring theory plus module theory plus field theory plus commutative algebra plus noncommutative algebra plus homological algebra plus Tor plus Ext plus Hochschild cohomology plus cyclic homology plus algebraic K-theory plus topological K-theory plus lambda-rings plus Adams operations plus Chern character plus Chern-Weil theory plus bivariant K-theory plus Kasparov K-theory plus operator K-theory plus cyclic homology plus negative cyclic homology plus periodic cyclic homology.
Number theory covers prime distribution plus analytic number theory plus algebraic number theory plus transcendental number theory plus arithmetic geometry plus class field theory plus Langlands program plus Iwasawa theory plus p-adic analysis plus p-adic Hodge theory plus rigid analysis plus Berkovich spaces plus perfectoid spaces plus prisms plus prismatic cohomology plus Fontaine's period rings plus comparison theorems.
Data mining and database theory covers query languages plus relational algebra plus relational calculus plus SQL plus Datalog plus linear Datalog plus Datalog with negation plus recursive queries plus conjunctive queries plus first-order queries plus query evaluation plus query optimization plus cost-based optimization plus join algorithms plus hash join plus sort-merge join plus nested loop join plus index structures plus B-trees plus B+ trees plus hash indexes plus bitmap indexes plus column stores plus row stores plus LSM trees plus write-ahead logging plus ARIES recovery algorithm plus transaction processing plus concurrency control plus two-phase locking plus timestamp ordering plus multiversion concurrency control plus snapshot isolation plus serializable snapshot isolation plus distributed transactions plus two-phase commit plus three-phase commit plus consensus plus Paxos plus Raft.
Cross-disciplinary read 29
Mathematical logic covers proof theory plus model theory plus set theory plus recursion theory plus computability theory plus ordinal analysis plus large cardinals plus forcing plus inner model theory plus descriptive set theory plus determinacy plus large continuum plus projective determinacy plus AD plus Woodin cardinals plus Ultimate-L plus determinacy hypotheses. Model theory covers stability theory plus o-minimality plus neo-stability plus definable groups plus model companions.
Mathematical physics covers classical mechanics plus continuum mechanics plus statistical mechanics plus kinetic theory plus quantum mechanics plus relativistic quantum mechanics plus quantum field theory plus gauge theory plus noncommutative field theory plus supersymmetry plus string theory plus conformal field theory plus topological field theory plus integrable systems plus soliton theory plus scattering theory plus inverse scattering plus Riemann-Hilbert problem.
Mathematical finance continuous-time and stochastic control covers stochastic optimal control plus dynamic programming principle plus Hamilton-Jacobi-Bellman equation plus verification theorem plus martingale method for optimal control plus maximum principle plus adjoint equation plus linear-quadratic stochastic control plus Kalman-Bucy filter plus Wonham filter plus partially observed control plus separating principle plus risk-sensitive control plus exponential-of-integral control plus linear-exponential-quadratic-Gaussian plus mean-variance control plus portfolio optimization plus Merton problem plus continuous-time portfolio selection plus consumption-investment problem plus infinite-horizon control plus ergodic control plus ergodic HJB plus long-run-average cost plus discounted-cost control plus viscosity solutions plus comparison principle plus stochastic target problems plus quantile hedging plus loss aversion in finance plus prospect theory plus cumulative prospect theory plus rank-dependent utility plus habit formation preferences plus Epstein-Zin preferences plus recursive utility plus Duffie-Epstein plus long-run risk plus stochastic differential utility plus consumption-based asset pricing plus Hansen-Jagannathan bounds plus SDF plus stochastic discount factor plus consumption-CAPM plus CCAPM plus habit-CAPM plus long-run risk model plus external habit plus recursive preferences plus Campbell-Cochrane habit model.
Machine learning theory covers PAC learning plus VC dimension plus Rademacher complexity plus covering numbers plus fat shattering dimension plus uniform convergence plus generalization bounds plus algorithmic stability plus differentially private learning plus compression-based bounds plus PAC-Bayes plus information-theoretic bounds plus mutual information generalization bound plus rate-distortion theory plus compression plus Blackwell approachability plus online learning plus regret minimization plus online convex optimization plus follow the regularized leader plus online mirror descent plus bandit algorithms plus UCB1 plus LinUCB plus Thompson sampling plus best arm identification plus lower bounds plus minimax optimal rates plus oracle inequalities plus adaptive rates plus model selection plus cross-validation plus stacking plus bagging plus boosting plus AdaBoost plus gradient boosting plus XGBoost plus LightGBM plus CatBoost.
Cross-disciplinary read 31
Operator algebras cover C-star algebras plus W-star algebras plus von Neumann algebras plus Tomita-Takesaki theory plus modular theory plus subfactor theory plus Jones index plus planar algebras plus fusion categories plus module categories plus Frobenius algebras plus Calabi-Yau algebras plus Hopf algebras plus quantum groups plus Kac-Paljutkin algebras plus group measure space construction plus free probability.
Representation theory algebraic groups and quantum groups covers algebraic groups plus structure theory plus Borel subgroups plus parabolic subgroups plus root data plus Weyl groups plus Bruhat decomposition plus Richardson orbits plus spherical varieties plus symmetric varieties plus wonderful compactifications plus spherical embeddings plus embedding theorems plus representation theory of algebraic groups plus highest weight modules plus Verma modules plus BGG category O plus derived category O plus Koszul duality plus Ringel duality plus Koszul graded category plus Koszul self-duality plus twisted Verma modules plus Kazhdan-Lusztig theory plus Lusztig's conjecture plus character sheaves plus perverse sheaves plus decomposition matrix plus tilting modules plus quantum groups plus canonical bases plus crystal bases plus Kashiwara crystals plus Littelmann paths plus Lakshmibai-Seshadri paths plus Mirkovic-Vilonen cycles plus MV polytopes plus affine Grassmannian plus affine flag variety plus Satake isomorphism plus geometric Satake equivalence.
Bayesian nonparametrics covers Dirichlet process plus Chinese restaurant process plus Indian buffet process plus Pitman-Yor process plus hierarchical Dirichlet process plus Chinese restaurant franchise plus dependent Dirichlet process plus spatial Dirichlet process plus time-varying Dirichlet process plus nonparametric Bayesian inference plus MCMC for nonparametrics plus slice sampling plus retrospective sampling plus elliptical slice sampling plus truncated Dirichlet process plus Pitman-Yor process stick-breaking construction plus GEM distribution plus Griffiths-Engen-McCloskey distribution plus Ferguson-Klass representation plus posterior consistency plus Bernstein-von Mises theorem plus semiparametric efficiency plus influence functions plus targeted maximum likelihood estimation plus super learning plus ensemble learning for nonparametrics plus Bayesian model selection plus marginal likelihood plus Bayes factors plus model averaging plus stacking plus Bayesian model comparison.
Deep learning theory covers approximation theory for neural networks plus universal approximation plus overparameterization plus neural tangent kernel plus lazy training plus feature learning plus implicit bias plus gradient descent dynamics plus loss landscape analysis plus saddle points plus mode connectivity plus lottery ticket hypothesis plus pruning plus quantization plus knowledge distillation plus transfer learning plus meta-learning plus MAML plus Reptile plus Prototypical Networks plus GANs plus VAEs plus diffusion models plus score-based generative models plus normalizing flows plus flow matching plus consistency models plus rectified flow plus transformer architectures plus attention mechanisms plus multi-head attention plus sparse attention plus linear attention plus long-context transformers plus state-space models plus Mamba plus RWKV plus Hyena plus RetNet plus mixture of experts.
Cross-disciplinary read 37
Lie theory and representation theory encompasses Lie algebras plus Lie groups plus representation theory plus root systems plus Weyl groups plus highest weight modules plus Verma modules plus BGG category O plus D-modules plus Harish-Chandra modules plus automorphic forms plus theta correspondence plus Beilinson-Bernstein localization plus primitive ideals plus Goldie rank plus Joseph ideal.
Mathematical physics covers classical mechanics plus continuum mechanics plus statistical mechanics plus kinetic theory plus quantum mechanics plus relativistic quantum mechanics plus quantum field theory plus gauge theory plus noncommutative field theory plus supersymmetry plus string theory plus conformal field theory plus topological field theory plus integrable systems plus soliton theory plus scattering theory plus inverse scattering plus Riemann-Hilbert problem.
Number theory covers prime distribution plus analytic number theory plus algebraic number theory plus transcendental number theory plus arithmetic geometry plus class field theory plus Langlands program plus Iwasawa theory plus p-adic analysis plus p-adic Hodge theory plus rigid analysis plus Berkovich spaces plus perfectoid spaces plus prisms plus prismatic cohomology plus Fontaine's period rings plus comparison theorems.
Complex analysis covers several complex variables plus Stein manifolds plus pluripotential theory plus plurisubharmonic functions plus Monge-Ampere equations plus complex Monge-Ampere plus Kähler geometry plus Hermitian geometry plus Calabi-Yau manifolds plus Hermitian-Yang-Mills connections plus holomorphic vector bundles plus stability plus Kobayashi-Hitchin correspondence plus Donaldson-Uhlenbeck-Yau theorem plus generalized complex geometry.
Cross-disciplinary read 41
Deep learning theory covers approximation theory for neural networks plus universal approximation plus overparameterization plus neural tangent kernel plus lazy training plus feature learning plus implicit bias plus gradient descent dynamics plus loss landscape analysis plus saddle points plus mode connectivity plus lottery ticket hypothesis plus pruning plus quantization plus knowledge distillation plus transfer learning plus meta-learning plus MAML plus Reptile plus Prototypical Networks plus GANs plus VAEs plus diffusion models plus score-based generative models plus normalizing flows plus flow matching plus consistency models plus rectified flow plus transformer architectures plus attention mechanisms plus multi-head attention plus sparse attention plus linear attention plus long-context transformers plus state-space models plus Mamba plus RWKV plus Hyena plus RetNet plus mixture of experts.
Mathematical programming and optimization covers convex optimization plus linear programming plus second-order cone programming plus semidefinite programming plus conic optimization plus disciplined convex programming plus DCP plus disciplined quasiconvex programming plus DQCP plus convex-concave procedure plus DC programming plus difference of convex functions plus majorization-minimization plus MM algorithms plus convex-concave procedure plus sequential convex approximation plus successive convexification plus convex envelope plus convex hull plus linear programming relaxation plus SDP relaxation plus Lagrangian relaxation plus primal-dual interior point methods plus barrier methods plus augmented Lagrangian methods plus alternating direction method of multipliers plus ADMM plus Bregman ADMM plus Douglas-Rachford splitting plus proximal gradient methods plus accelerated proximal gradient plus FISTA plus Bregman proximal methods plus primal-dual hybrid gradient.
Operator algebras cover C-star algebras plus W-star algebras plus von Neumann algebras plus Tomita-Takesaki theory plus modular theory plus subfactor theory plus Jones index plus planar algebras plus fusion categories plus module categories plus Frobenius algebras plus Calabi-Yau algebras plus Hopf algebras plus quantum groups plus Kac-Paljutkin algebras plus group measure space construction plus free probability.
Category theory covers category theory foundations plus enriched categories plus 2-categories plus bicategories plus double categories plus fibrations plus Grothendieck fibrations plus Street fibrations plus profunctors plus monads plus comonads plus distributive laws plus monad transformers plus algebras plus coalgebras plus bialgebras plus Hopf algebras plus operads plus properads plus PROPs plus simplicial categories plus quasi-categories plus infinity categories plus model categories plus Quillen model structure plus homotopical algebra plus derived categories plus triangulated categories plus stable infinity categories plus presentable infinity categories plus locally presentable categories plus accessible categories plus combinatorial model categories plus quasicategorical model categories.
Cross-disciplinary read 43
Complex analysis covers several complex variables plus Stein manifolds plus pluripotential theory plus plurisubharmonic functions plus Monge-Ampere equations plus complex Monge-Ampere plus Kähler geometry plus Hermitian geometry plus Calabi-Yau manifolds plus Hermitian-Yang-Mills connections plus holomorphic vector bundles plus stability plus Kobayashi-Hitchin correspondence plus Donaldson-Uhlenbeck-Yau theorem plus generalized complex geometry.
Mathematical statistics covers parametric inference plus nonparametric inference plus semi-parametric inference plus high-dimensional statistics plus shape-constrained inference plus isotonic regression plus convex regression plus sparse estimation plus lasso plus elastic net plus adaptive lasso plus group lasso plus fused lasso plus graphical lasso plus sparse precision matrix estimation plus covariance estimation plus Stein estimation plus shrinkage estimation plus ridge regression plus principal component regression plus partial least squares plus canonical correlation analysis plus independent component analysis plus projection pursuit.
Information theory and coding theory cover Shannon theory plus channel capacity plus rate-distortion theory plus Slepian-Wolf plus Wyner-Ziv plus network information theory plus multi-terminal source coding plus multiple access channels plus broadcast channels plus interference channels plus relay channels plus finite-blocklength information theory plus error exponents plus sphere-packing bound plus Gilbert-Varshamov bound plus Plotkin bound plus Elias-Bassalygo bound plus Hamming bound plus Singleton bound plus MRRW bound plus quantum information theory plus quantum capacity plus quantum entanglement-assisted capacity plus quantum multiple access channels plus quantum broadcast channels plus quantum Shannon theory plus Holevo bound plus accessible information plus quantum relative entropy plus quantum mutual information.
Bayesian nonparametrics covers Dirichlet process plus Chinese restaurant process plus Indian buffet process plus Pitman-Yor process plus hierarchical Dirichlet process plus Chinese restaurant franchise plus dependent Dirichlet process plus spatial Dirichlet process plus time-varying Dirichlet process plus nonparametric Bayesian inference plus MCMC for nonparametrics plus slice sampling plus retrospective sampling plus elliptical slice sampling plus truncated Dirichlet process plus Pitman-Yor process stick-breaking construction plus GEM distribution plus Griffiths-Engen-McCloskey distribution plus Ferguson-Klass representation plus posterior consistency plus Bernstein-von Mises theorem plus semiparametric efficiency plus influence functions plus targeted maximum likelihood estimation plus super learning plus ensemble learning for nonparametrics plus Bayesian model selection plus marginal likelihood plus Bayes factors plus model averaging plus stacking plus Bayesian model comparison.
Cross-disciplinary read 47
Differential equations and dynamical systems cover ordinary differential equations plus partial differential equations plus delay differential equations plus stochastic differential equations plus functional differential equations plus integro-differential equations plus fractional differential equations plus nonlocal equations plus free boundary problems plus inverse problems plus control theory plus optimal control plus calculus of variations plus infinite-dimensional Hamiltonian systems plus KAM theory plus KAM for PDEs plus Arnold diffusion plus Nekhoroshev estimates plus Lyapunov functions plus normally hyperbolic invariant manifolds plus center manifold theory plus invariant manifold theory plus geometric singular perturbation theory plus blow-up analysis plus singularity formation plus chemotaxis models plus pattern formation plus reaction-diffusion systems plus Ginzburg-Landau equation plus nonlinear Schrodinger equation plus KdV equation plus KPZ equation plus Burgers equation plus Navier-Stokes equations plus Euler equations plus Euler-alpha equations plus Camassa-Holm equation plus Whitham equation.
Mathematical finance continuous-time and stochastic control covers stochastic optimal control plus dynamic programming principle plus Hamilton-Jacobi-Bellman equation plus verification theorem plus martingale method for optimal control plus maximum principle plus adjoint equation plus linear-quadratic stochastic control plus Kalman-Bucy filter plus Wonham filter plus partially observed control plus separating principle plus risk-sensitive control plus exponential-of-integral control plus linear-exponential-quadratic-Gaussian plus mean-variance control plus portfolio optimization plus Merton problem plus continuous-time portfolio selection plus consumption-investment problem plus infinite-horizon control plus ergodic control plus ergodic HJB plus long-run-average cost plus discounted-cost control plus viscosity solutions plus comparison principle plus stochastic target problems plus quantile hedging plus loss aversion in finance plus prospect theory plus cumulative prospect theory plus rank-dependent utility plus habit formation preferences plus Epstein-Zin preferences plus recursive utility plus Duffie-Epstein plus long-run risk plus stochastic differential utility plus consumption-based asset pricing plus Hansen-Jagannathan bounds plus SDF plus stochastic discount factor plus consumption-CAPM plus CCAPM plus habit-CAPM plus long-run risk model plus external habit plus recursive preferences plus Campbell-Cochrane habit model.
Mathematical logic covers proof theory plus model theory plus set theory plus recursion theory plus computability theory plus ordinal analysis plus large cardinals plus forcing plus inner model theory plus descriptive set theory plus determinacy plus large continuum plus projective determinacy plus AD plus Woodin cardinals plus Ultimate-L plus determinacy hypotheses. Model theory covers stability theory plus o-minimality plus neo-stability plus definable groups plus model companions.
Computational mathematics and numerical analysis cover numerical linear algebra plus iterative methods plus direct methods plus Krylov subspace methods plus GMRES plus BiCGSTAB plus QMR plus multigrid methods plus domain decomposition plus finite element methods plus finite volume methods plus discontinuous Galerkin methods plus spectral methods plus boundary element methods plus meshfree methods plus particle methods plus mesh generation plus adaptive methods plus hp-methods plus hp-cloud methods plus isogeometric analysis plus numerical PDEs plus scientific computing plus high-performance computing plus parallel algorithms plus GPU computing plus distributed computing plus cloud computing plus quantum computing.
Cross-disciplinary read 53
Probability theory covers limit theorems plus central limit theorem plus local limit theorem plus large deviations plus concentration of measure plus Gaussian processes plus stochastic processes plus stochastic integrals plus Ito calculus plus Stratonovich integrals plus backward stochastic differential equations plus stochastic partial differential equations plus stochastic control plus dynamic programming principle plus Hamilton-Jacobi-Bellman equations plus mean field games plus mean field control plus optimal stopping plus American options plus excursion theory plus potential theory for Markov processes plus Dirichlet forms plus Hunt processes plus Levy processes plus subordinators plus infinitely divisible distributions plus point processes plus Poisson processes plus Cox processes plus shot noise processes plus self-exciting processes plus Hawkes processes plus random measures plus Poisson point processes plus determinantal point processes plus permanental point processes plus Palm theory plus ergodic theory for random walks.
Mathematical finance continuous-time and stochastic control covers stochastic optimal control plus dynamic programming principle plus Hamilton-Jacobi-Bellman equation plus verification theorem plus martingale method for optimal control plus maximum principle plus adjoint equation plus linear-quadratic stochastic control plus Kalman-Bucy filter plus Wonham filter plus partially observed control plus separating principle plus risk-sensitive control plus exponential-of-integral control plus linear-exponential-quadratic-Gaussian plus mean-variance control plus portfolio optimization plus Merton problem plus continuous-time portfolio selection plus consumption-investment problem plus infinite-horizon control plus ergodic control plus ergodic HJB plus long-run-average cost plus discounted-cost control plus viscosity solutions plus comparison principle plus stochastic target problems plus quantile hedging plus loss aversion in finance plus prospect theory plus cumulative prospect theory plus rank-dependent utility plus habit formation preferences plus Epstein-Zin preferences plus recursive utility plus Duffie-Epstein plus long-run risk plus stochastic differential utility plus consumption-based asset pricing plus Hansen-Jagannathan bounds plus SDF plus stochastic discount factor plus consumption-CAPM plus CCAPM plus habit-CAPM plus long-run risk model plus external habit plus recursive preferences plus Campbell-Cochrane habit model.
Mathematical statistics covers parametric inference plus nonparametric inference plus semi-parametric inference plus high-dimensional statistics plus shape-constrained inference plus isotonic regression plus convex regression plus sparse estimation plus lasso plus elastic net plus adaptive lasso plus group lasso plus fused lasso plus graphical lasso plus sparse precision matrix estimation plus covariance estimation plus Stein estimation plus shrinkage estimation plus ridge regression plus principal component regression plus partial least squares plus canonical correlation analysis plus independent component analysis plus projection pursuit.
Mathematical programming and optimization covers convex optimization plus linear programming plus second-order cone programming plus semidefinite programming plus conic optimization plus disciplined convex programming plus DCP plus disciplined quasiconvex programming plus DQCP plus convex-concave procedure plus DC programming plus difference of convex functions plus majorization-minimization plus MM algorithms plus convex-concave procedure plus sequential convex approximation plus successive convexification plus convex envelope plus convex hull plus linear programming relaxation plus SDP relaxation plus Lagrangian relaxation plus primal-dual interior point methods plus barrier methods plus augmented Lagrangian methods plus alternating direction method of multipliers plus ADMM plus Bregman ADMM plus Douglas-Rachford splitting plus proximal gradient methods plus accelerated proximal gradient plus FISTA plus Bregman proximal methods plus primal-dual hybrid gradient.