{"credit":"Taylor (gtmcknight): unmodified replica, Gaussian/filter mathematics, qm tables and solved reference. qt1, TwoComp and COMBO from linked IdentityMD artifacts.","status":"complete replica experiment; no contract or submission","blocked":null,"source_commit":{"sha":"4812a6382311624e1506aca131ca35c378f930d9","verification":"GitHub git-commit API returned exact SHA; pinned raw source downloads","url":"https://api.github.com/repos/gtmcknight/amm-sim/git/commits/4812a6382311624e1506aca131ca35c378f930d9"},"hashes":{"gates.py":{"url":"https://raw.githubusercontent.com/gtmcknight/amm-sim/4812a6382311624e1506aca131ca35c378f930d9/research/solved/gates.py","sha256":"603280c1f595ca163ffaf2ba1ebdc7945fdb08dabae67fc53c922d39e1cce604","bytes":792},"replica2.py":{"url":"https://raw.githubusercontent.com/gtmcknight/amm-sim/4812a6382311624e1506aca131ca35c378f930d9/research/solved/replica2.py","sha256":"6e37b4e61e1ffa5ffb1e43912ddb8659105989aa986eb48fcd1d7608f8dca685","bytes":7503},"evalpol.py":{"url":"https://raw.githubusercontent.com/gtmcknight/amm-sim/4812a6382311624e1506aca131ca35c378f930d9/research/solved/evalpol.py","sha256":"6e67667bfed7798bdd3384652a6efc6ba0b1fce653740d0f9af2bcbcb8e6a27c","bytes":4610},"real.py":{"url":"https://raw.githubusercontent.com/gtmcknight/amm-sim/4812a6382311624e1506aca131ca35c378f930d9/research/solved/real.py","sha256":"1305b5333731c046c3e6736cc2578eff4635c88132a39c22f5780d78aef2c81e","bytes":9586},"latt.py":{"url":"https://raw.githubusercontent.com/gtmcknight/amm-sim/4812a6382311624e1506aca131ca35c378f930d9/research/solved/latt.py","sha256":"925c0fe996b19fd51fb79a85d5e4dba4bc2f5453d0a156cc1e480f65a6a00cc8","bytes":8893},"pol_mid.json":{"url":"https://raw.githubusercontent.com/gtmcknight/amm-sim/4812a6382311624e1506aca131ca35c378f930d9/research/solved/pol_mid.json","sha256":"efbb5ee102db4d312f64565d9bf7232b6becbe02b624164f744a402bbbf2983b","bytes":780061},"qm_k1e-5.json":{"url":"https://raw.githubusercontent.com/gtmcknight/amm-sim/4812a6382311624e1506aca131ca35c378f930d9/research/solved/qm_k1e-5.json","sha256":"5ff53e42ca87390e5f65d1caa3f7a480e67a8212fb0670274dfd562b05ba1afd","bytes":7892018},"dp.py":{"url":"https://raw.githubusercontent.com/gtmcknight/amm-sim/4812a6382311624e1506aca131ca35c378f930d9/research/solved/dp.py","sha256":"ca4e7c378d23c46c7def4716611135c471a831d15234b984fe30d58d9eec5734","bytes":15595},"qmdp.py":{"url":"https://raw.githubusercontent.com/gtmcknight/amm-sim/4812a6382311624e1506aca131ca35c378f930d9/research/solved/qmdp.py","sha256":"63b8c35048718c37eb0a5dae9e70849fa3af257c335e4a2e3a625c833f1dc1eb","bytes":8609},"calib_out_2048.json":{"url":"https://raw.githubusercontent.com/gtmcknight/amm-sim/4812a6382311624e1506aca131ca35c378f930d9/research/solved/calib_out_2048.json","sha256":"fdf24c7aa2b604d770336a5783276da8ed4d50bca9dbcc02e98b4da58f608d26","bytes":2948144},"readme":{"url":"https://raw.githubusercontent.com/gtmcknight/amm-sim/4812a6382311624e1506aca131ca35c378f930d9/research/solved/README.md","sha256":"02d2b51f17adbb5411a2a67f7a37255143ba381a73ff10abba557d7d23c0cef2","bytes":10483},"qt1":{"url":"https://api.imdnotdead.fun/artifacts/8a6e303b662e1a5ea9bb47f31acedd3988ef65b3ddb66d074517aa59ad41ca40","sha256":"8a6e303b662e1a5ea9bb47f31acedd3988ef65b3ddb66d074517aa59ad41ca40","bytes":62223},"TC":{"url":"https://api.imdnotdead.fun/artifacts/d3c849db2ac31e567b57e52b80b86e8c24c842e5b4125b9d75c78e3cb07115b9","sha256":"d3c849db2ac31e567b57e52b80b86e8c24c842e5b4125b9d75c78e3cb07115b9","bytes":23156},"COMBO":{"url":"https://api.imdnotdead.fun/artifacts/d360bea6c8c982717b7b43a8069a87013169c89ee3e4da25622db4d880b936d4","sha256":"d360bea6c8c982717b7b43a8069a87013169c89ee3e4da25622db4d880b936d4","bytes":25626},"QMS1":{"url":"https://api.imdnotdead.fun/artifacts/17d249ab0c33fc44931d9b105cf3f9e3846a560bd518b67310c2a56defacdecc","sha256":"17d249ab0c33fc44931d9b105cf3f9e3846a560bd518b67310c2a56defacdecc","bytes":29486},"commit.json":{"url":"https://api.github.com/repos/gtmcknight/amm-sim/git/commits/4812a6382311624e1506aca131ca35c378f930d9","sha256":"b3e292db8053fbb0ac56c8e7f122e482a05046bde97873a477fdc376a52b101f","bytes":1301},"extracted":{"passed":true,"expected":{"qmref.py":"1dfc9b7a3c2237ff7a93dd9f0af54b9bdab16f2ce0a7eb00784cb4432598db92","Strategy.sol":"c68887cff61180e12ba2b1ceb333edf76a830f91f63b5d775a8806fb9ec08aca","twocomp.py":"9d3c1833321493fb51bbde4e1b4dabc2f8978bd6d774ee2af8dc8aed13f0f0af","combo.py":"c5b711a0b4fb4783e8879a484bf87a427171f4a11aa5eb5ec3513c24ead1a796","blob.bin":"3ac7cb8811ecfb6a17f499aa4e9b70dd55f7fb6432b3aec3b904e11a3dece38e","qt1":"8a6e303b662e1a5ea9bb47f31acedd3988ef65b3ddb66d074517aa59ad41ca40","TC":"d3c849db2ac31e567b57e52b80b86e8c24c842e5b4125b9d75c78e3cb07115b9","COMBO":"d360bea6c8c982717b7b43a8069a87013169c89ee3e4da25622db4d880b936d4","QMS1":"17d249ab0c33fc44931d9b105cf3f9e3846a560bd518b67310c2a56defacdecc"}}},"gates_output":"seed   perfect-info      gaussian+plain    gaussian+qm       lattice+qm\n10000  448.261189  443.455598  443.522608  443.663520\n10001  576.779199  571.463409  571.623551  571.977293\n10002  531.886497  525.168028  525.865859  526.411631\n","gates_pass":{"P":true,"all_12_match_to_6dp":true,"before_policy_changes":true,"qt1_edges":[[10000,443.1017411146846],[10001,571.4544833819259],[10002,525.908441161224]],"Q":"complete paired report for nominee only; teacher/compact comparison on full choose; raw calibration pilot is separately scoped","F":"not applicable; no contract"},"findings":{"nominee":"compact","nominee_report_vs_COMBO":{"mean_diff":0.028164469516271523,"se":0.00829932267062314},"nominee_report_vs_qt1":{"mean_diff":0.625420727193216,"se":0.013546117627163672},"selection_rule":"compact when within 0.05 mean edge of teacher on all choose seeds, else teacher","negative_results_are_reported":true},"seed_coverage":{"training_trajectories":[73000,73511,512],"label_fit":[73000,73447,448],"training_validation_no_selection":[73448,73511,64],"calibration_choose_pilot":[64000,64127,128],"representation_choose":[64000,65023,1024],"report":[10000,12047,2048],"steps_per_run":10000,"configs":"replica2 defaults, unmodified; fresh policy object per seed and tested representation","report_not_used_for_training_calibration_or_selection":true,"oracle_diagnostics":["Unmodified P perfect-info gate only; no oracle inputs in learned policies"]},"frozen_spec":{"credit":"Taylor (gtmcknight): replica/filter/qm tables","utc":"2026-11-10T08:20:26.251685+00:00","nominee":"compact","rule":"compact when within 0.05 mean edge of teacher on all choose seeds, else teacher","choose_gain_vs_COMBO":{"teacher":0.04268041979110149,"compact":0.02592771064529331},"sha256":{"driver.py":"a6bd531e0cd4bd54bd36b97f872237071644f7f7a831d6298a18cd9bd9867b6b","calibrate.py":"5ec803eba783e58d82d54158857091368aab700ae5fb95c131e843afb01e2cf9","select.py":"9d4a7b4ce7bb1b0fe91878550ff2a6da5ab1be3f87f4da3c2cd4afb236bec1f4","validate.py":"bcd95041097621a9489628a50f46fca470386668e04fe76dd50d982e3f414786","model.json":"dc00f5fc28911c1ad89b8d2dc4a48d36c7f4b5ce6bfab050195e2205fc31864e","choose_spec.json":"1ab961f62554c62db8416148f9ba30544625c4ffa6859c210e5025dae1f72b9a","report_spec.json":"09e23d44288f48d5d3047e3c07faad2dfaf455591ca989ae7f1c50c2af16b8e1","replica2.py":"6e37b4e61e1ffa5ffb1e43912ddb8659105989aa986eb48fcd1d7608f8dca685","combo.py":"c5b711a0b4fb4783e8879a484bf87a427171f4a11aa5eb5ec3513c24ead1a796","twocomp.py":"9d3c1833321493fb51bbde4e1b4dabc2f8978bd6d774ee2af8dc8aed13f0f0af","qmref.py":"1dfc9b7a3c2237ff7a93dd9f0af54b9bdab16f2ce0a7eb00784cb4432598db92","blob.bin":"3ac7cb8811ecfb6a17f499aa4e9b70dd55f7fb6432b3aec3b904e11a3dece38e"}},"features":{"order":["logit_pi","log_dt","signed_impact_bps","innovation_sd","prior_sd_bps","band_width_bps","rival_displacement_bps","synced","previous_count","elapsed_fraction","noise_ewma_bps2","retail_logimpact_mean","retail_logimpact_sd","callback_rate","same_direction","log_ax_over_x","log_ay_over_y","mixture_weight"],"formulas":{"logit_pi":"ln(max(1e-8,pi_raw)/max(1e-8,1-pi_raw))","log_dt":"ln(1+timestamp-previous_timestamp)","signed_impact_bps":"(1 if isBuy else -1)*1e4*abs(ln(post_spot/previous_spot)); sign follows callback flag","innovation_sd":"(arb_log_reading-prior_mixture_mean)/max(prior_sd,1e-9)","prior_sd_bps":"1e4*sqrt(prior_mixture_variance)","band_width_bps":"1e4*(ln(previous_spot/(1-ask_fee))-ln(previous_spot*(1-bid_fee)))","rival_displacement_bps":"1e4*(ln(inferred_rival_spot)-prior_mixture_mean)","synced":"previous inferred rival-sync Boolean","previous_count":"min(previous own inferred retail count,8)","elapsed_fraction":"callback timestamp/10000","noise_ewma_bps2":"initial 89.3025; before next callback update N=(1-a)*N+a*min(900,(1e4*(lp-last_first_lp))^2/max(1,t-last_first_t)); a=.005*pi_raw","retail_logimpact_mean":"initial ln(11.12); update r+=a*(ln(max(1e4*d,1e-5))-r); a=.005*(1-pi_raw)","retail_logimpact_sd":"sqrt(max(.01,V)); V initialized 1.176^2 and updated (1-a)*V+a*(logimpact-old_r)^2","callback_rate":"number of preceding own callbacks/(timestamp+1)","same_direction":"previous callback isBuy == current isBuy (False initially)","log_ax_over_x":"ln(max(amountX/post_reserveX,1e-12))","log_ay_over_y":"ln(max(amountY/post_reserveY,1e-12))","mixture_weight":"weight of first carried TwoComp component before current propagation"},"clip":"every feature clipped to [-10000,10000] after computation","timing":"features read before posterior, phase/branch quoting; estimator state updated using raw pi, never the true label","hidden_inputs":"on_trade forwards only the six live callback fields; init ignores sigma/lam/size_mean","caveat":"noise/retail-law fields are observable proxies, not unbiased estimates of hidden parameters"},"policy_rule":"bin each feature by bisect_right; teacher residual is sum of 48 depth-2 GBT leaves, compact residual is bias+sum(table[k][bin[k]]). pi=sigmoid((logit(pi_raw)+alpha*residual)/temperature), clipped sigmoid logit to +/-35. Replacement occurs before COMBO is_arb threshold, mixture posterior and all branch quoting. Other COMBO logic stays unchanged.","calibration":{"teacher":{"selection":"minimum logloss on raw closed-loop choose pilot","grid":[{"alpha":0.5,"temp":0.7,"logloss":0.36699762186510115},{"alpha":0.5,"temp":0.85,"logloss":0.35251177114226334},{"alpha":0.5,"temp":1.0,"logloss":0.3481867137255523},{"alpha":0.5,"temp":1.15,"logloss":0.3494315769801973},{"alpha":0.5,"temp":1.3,"logloss":0.35383620510521974},{"alpha":1.0,"temp":0.7,"logloss":0.36373015758693067},{"alpha":1.0,"temp":0.85,"logloss":0.35011846904879357},{"alpha":1.0,"temp":1.0,"logloss":0.3465148201601598},{"alpha":1.0,"temp":1.15,"logloss":0.348377160904226},{"alpha":1.0,"temp":1.3,"logloss":0.35331744048358255}],"selected":{"alpha":1.0,"temp":1.0,"metrics":{"n":55986,"error":0.15846818847569036,"brier":0.11023373513118823,"logloss":0.3465148201601598,"ece10":0.005338272707254356}}},"compact":{"selection":"minimum logloss on raw closed-loop choose pilot","grid":[{"alpha":0.5,"temp":0.7,"logloss":0.36584164692194554},{"alpha":0.5,"temp":0.85,"logloss":0.3519295292401374},{"alpha":0.5,"temp":1.0,"logloss":0.3479891505908891},{"alpha":0.5,"temp":1.15,"logloss":0.34949977685025},{"alpha":0.5,"temp":1.3,"logloss":0.35409178087379484},{"alpha":1.0,"temp":0.7,"logloss":0.36387049511781816},{"alpha":1.0,"temp":0.85,"logloss":0.35067171670245956},{"alpha":1.0,"temp":1.0,"logloss":0.3473219782624882},{"alpha":1.0,"temp":1.15,"logloss":0.3493448747669273},{"alpha":1.0,"temp":1.3,"logloss":0.3543887081974404}],"selected":{"alpha":1.0,"temp":1.0,"metrics":{"n":55985,"error":0.1594712869518621,"brier":0.1106561738013748,"logloss":0.3473219782624882,"ece10":0.003580453077775537}}}},"probability_error_and_shift":{"training_validation_COMBO_trajectories":{"n_train":195835,"n_validation":28061,"distill_logit_rmse":0.26834903753439326,"quantization_logit_max":0.001338856426228996,"quantization_logit_rmse":0.0003380768863940658,"teacher":{"n":28061,"error":0.16057873917536794,"brier":0.11110900172548314,"logloss":0.3502826556153649,"ece10":0.00474896112722945},"compact":{"n":28061,"error":0.15965218630839956,"brier":0.11149974540155821,"logloss":0.3524518338375053,"ece10":0.005187468503041529},"raw":{"n":28061,"error":0.16246748155803428,"brier":0.11231122354921086,"logloss":0.3560265812425931,"ece10":0.007241564082630589}},"choose_closed_loop":{"teacher":{"error":0.15753199723075284,"brier":0.1096987079654995,"logloss":0.34475406302496814,"n":450664,"ece10":0.002455256572725984,"feature_sample_n":14554,"feature_means":[-0.8514526821706732,0.8449956420276824,-0.2704374059195447,0.015281853654950862,15.162456885289163,73.37118807030595,0.01814958133982935,0.6031331592689295,0.9811048509000962,0.4813435412944895,279.8402017008988,2.4275343837757357,1.1522033169548567,1.0495063057788152,0.5103064449635839,-7.643770107228851,-7.643888153285497,0.6730520645813548],"label_prevalence":0.5171774082726398},"compact":{"error":0.15791938042983789,"brier":0.11013909662671485,"logloss":0.3473134144616384,"n":450635,"ece10":0.0023952989476019606,"feature_sample_n":14548,"feature_means":[-0.80659437630158,0.839721323841049,0.2431396817931288,-0.024421684464850626,15.213268396149656,74.3497566721951,0.8711681953360857,0.5956832554302997,0.9775226835303822,0.48116102557052515,279.8124763977933,2.42765305863861,1.1519798068554317,1.0493367458487364,0.506117679406104,-7.647949128098829,-7.648137765356819,0.673022020056132],"label_prevalence":0.5281825680505912}},"report_closed_loop":{"compact":{"error":0.15986549900602165,"brier":0.11126394542083316,"logloss":0.35010065869788987,"n":899919,"ece10":0.002044318096169823,"feature_sample_n":29085,"feature_means":[-0.8034899642431976,0.8435572630459305,-0.11878026065923562,0.017926018611662313,15.18995266863052,74.11635635961987,-0.2831079204380347,0.6004813477737665,0.9794739556472408,0.48153528623001546,279.9875449255009,2.4250924769372126,1.1511931222106921,1.0472357767879847,0.5073405535499398,-7.641962495182363,-7.641923364103628,0.6706066508181041],"label_prevalence":0.5224342444559051}},"interpretation":"training metrics are off-policy held-out COMBO trajectories. Choose/report are own-policy closed-loop metrics, sampled every 16th first-own callback. Report metrics are diagnostic only; no report retuning."},"policy":{"pilot":{"qt1":{"COMBO":{"n":128,"mean_diff":-0.603432888439102,"se":0.06128054218907142,"worst5":[[64075,-2.6427585511460165],[64003,-2.5579592306520453],[64044,-2.482025272988949],[64065,-1.7358785901802776],[64114,-1.6697144020498627]],"sum_d":-77.23940972020506,"sum_d2":107.65503576643215}},"COMBO":{"qt1":{"n":128,"mean_diff":0.603432888439102,"se":0.06128054218907142,"worst5":[[64101,-2.7769830832156686],[64072,-0.6198058143504568],[64116,-0.5976360933805154],[64053,-0.5087301499639239],[64094,-0.50678623624799]],"sum_d":77.23940972020506,"sum_d2":107.65503576643215}},"teacher":{"COMBO":{"n":128,"mean_diff":0.038200954687916955,"se":0.03076347287285427,"worst5":[[64099,-0.9719686764527751],[64093,-0.6479794624690385],[64011,-0.6271467559922712],[64105,-0.5795404632456211],[64028,-0.5406019650985172]],"sum_d":4.88972220005337,"sum_d2":15.571328430761096},"qt1":{"n":128,"mean_diff":0.641633843127019,"se":0.05986809806151223,"worst5":[[64101,-1.7766653553080687],[64011,-0.7974275291410322],[64093,-0.6726911427992945],[64116,-0.5816737533514242],[64084,-0.49488298247695184]],"sum_d":82.12913192025843,"sum_d2":110.96140962109558}},"compact":{"COMBO":{"n":128,"mean_diff":0.026099166914557603,"se":0.03441947788808375,"worst5":[[64079,-0.8123481008137219],[64099,-0.7892858446981563],[64105,-0.7524545601946784],[64078,-0.6960601014418444],[64050,-0.6745362107112101]],"sum_d":3.340693365063373,"sum_d2":19.345679960428324},"qt1":{"n":128,"mean_diff":0.6295320553536596,"se":0.060541127819134434,"worst5":[[64101,-1.7534485769058392],[64049,-0.7015676460811733],[64078,-0.6538279356213934],[64011,-0.5944830431889159],[64072,-0.554029674581102]],"sum_d":80.58010308526843,"sum_d2":110.30970684603206}}},"choose":{"qt1":{"COMBO":{"n":1024,"mean_diff":-0.6096250136973232,"se":0.019402886773559457,"worst5":[[64697,-3.4913405586955832],[64075,-2.6427585511460165],[64003,-2.5579592306520453],[64505,-2.490816220900001],[64044,-2.482025272988949]],"sum_d":-624.256014026059,"sum_d2":774.9360935131383}},"COMBO":{"qt1":{"n":1024,"mean_diff":0.6096250136973232,"se":0.019402886773559457,"worst5":[[64101,-2.7769830832156686],[64845,-1.1638684799397652],[64374,-1.0234338236045915],[64795,-0.9831714544176293],[64555,-0.9461893679029458]],"sum_d":624.256014026059,"sum_d2":774.9360935131383}},"teacher":{"COMBO":{"n":1024,"mean_diff":0.04268041979110149,"se":0.01072070857311321,"worst5":[[64338,-1.1925217115001487],[64197,-1.0311982567895939],[64817,-0.9943982918538268],[64099,-0.9719686764527751],[64869,-0.8861458850583404]],"sum_d":43.70474986608792,"sum_d2":122.26425156227998},"qt1":{"n":1024,"mean_diff":0.6523054334884247,"se":0.01953132551180329,"worst5":[[64101,-1.7766653553080687],[64522,-1.2884251851703539],[64761,-1.2883506272463592],[64795,-0.975148132474601],[64590,-0.9551159368145363]],"sum_d":667.9607638921469,"sum_d2":835.32690059289}},"compact":{"COMBO":{"n":1024,"mean_diff":0.02592771064529331,"se":0.0116201483191013,"worst5":[[64715,-1.3644557939308015],[64522,-1.3037705411322804],[64957,-1.1992801593555669],[64338,-1.1408221130287188],[64505,-1.0703021979294363]],"sum_d":26.54997570078035,"sum_d2":142.13707122406484},"qt1":{"n":1024,"mean_diff":0.6355527243426166,"se":0.019501707271001492,"worst5":[[64522,-1.7585886717549783],[64101,-1.7534485769058392],[64509,-1.1342293101089354],[64698,-1.0576768810601607],[64761,-0.9866964293947831]],"sum_d":650.8059897268394,"sum_d2":812.0229205936974}}},"report":{"qt1":{"COMBO":{"n":2048,"mean_diff":-0.5972562576769443,"se":0.013389082248862141,"worst5":[[10035,-3.763936966649908],[11776,-2.7239732542892625],[11211,-2.6915414776889293],[11242,-2.683378630478046],[10546,-2.4149308007862373]],"sum_d":-1223.180815722382,"sum_d2":1482.0877473193957}},"COMBO":{"qt1":{"n":2048,"mean_diff":0.5972562576769443,"se":0.013389082248862141,"worst5":[[10146,-3.5524969294904167],[11062,-2.334849143696829],[11808,-1.8227381435364123],[10519,-1.7826914653087442],[11376,-1.750295687514381]],"sum_d":1223.180815722382,"sum_d2":1482.0877473193957}},"compact":{"COMBO":{"n":2048,"mean_diff":0.028164469516271523,"se":0.00829932267062314,"worst5":[[10416,-2.2994635115367714],[11911,-1.706591645006256],[11696,-1.565161967862423],[10401,-1.0915223797054523],[11918,-1.0715996822000307]],"sum_d":57.68083356932408,"sum_d2":290.3819315088466},"qt1":{"n":2048,"mean_diff":0.625420727193216,"se":0.013546117627163672,"worst5":[[10519,-2.637827418420443],[11376,-1.8201499503721834],[10416,-1.5865458851168341],[10401,-1.5046991422774454],[11062,-1.390441687632233]],"sum_d":1280.8616492917063,"sum_d2":1570.3450926508863}}}},"validation":{"cached_surfaces_exact_equality":true,"compiled_teacher_exact_binned_vectors":10000,"teacher":{"fixed_callback_count":5000,"hidden_init_and_callback_invariance":"exact fees","fast_vs_original_moments":"exact fees"},"compact":{"fixed_callback_count":5000,"hidden_init_and_callback_invariance":"exact fees","fast_vs_original_moments":"exact fees"},"hook_identity":{"73000":0.0,"73001":0.0,"73002":0.0},"offline_bundle":{"passed":true,"sha256":"231f1cfc9e02085a34baec19177545794e5318a09d94c0ec2e553e70d120bd47","files":13,"check":"unpacked file bytes and complete validation results equal working copy"}},"model_counts":{"teacher_trees":48,"teacher_max_depth":2,"compact_table_entries":235,"cutpoints":217,"compact_weight_bytes_q12_including_bias":472,"cutpoint_bytes_float64":1736,"model_json_bytes":12830,"model_json_zlib_bytes":5574,"compact_inference_operations":"18 binary searches (<=4 comparisons each with <=15 cuts), 18 table reads/adds plus bias, alpha multiply, temperature divide, sigmoid exp; feature calculations and existing COMBO separate","teacher_inference_operations":"<=96 split comparisons and 48 leaf additions per first-own callback","extra_live_history_scalars":7,"extra_history_names":["noise","rm","rv","pin","pint","calls","lastdir"],"diagnostic_only_state":"sample rows, callback ordinal, latest features; omitted by a deployment adaptation","gas_estimate":null,"gas_note":"No EVM/fixed-point port built; no measured or inferred gas claim. Counts exclude existing COMBO table/filter work."},"quantization":{"weights":"signed Q12 residual logit coefficients, round-to-nearest ties-to-even; decode as integer/4096; cutpoints float64, input scales as feature formulas","global_logit_error_bound":0.0023193359375,"global_probability_error_bound_unit_calibration":0.000579833984375,"max_logit_error_training":0.001338856426228996,"rmse_logit_training":0.0003380768863940658,"max_probability_error_bound_training_at_unit_calibration":0.000334714106557249,"note":"Model and cutpoints are exact in embedded model.json. Bound is quantization-only on observed training rows; distillation error is separate."},"data_regeneration":{"sampling":"COMBO with model=None; collect one first-own callback every 16 classifier opportunities per seed, starting at opportunity 0; true arb Boolean stored only as supervised label","manifest_sha256":"057124ba2b1b7a82f90f57e5261bf7f0544242703620e22f36e99698580f1c5f","manifest_format":"UTF-8 compact JSON [[seed,sha256(data/<seed>.json.z)],...] sorted by seed, no trailing newline","manifest_entries":512,"compressed_data_bytes":31161913,"model_training":"448 whole-seed fit +64 whole-seed validation. 16 quantile bins/feature, logistic baseline logit(pi_raw),48 depth-2 histogram Newton GBT trees, L2=20, shrink=.12, leaves capped +/-3 before shrink, leaf min64, node min256. Distillation: 12 additive backfit passes on teacher residual logits, bin ridge20, then Q12 quantization. No tuning on validation/report."},"rerun":{"python_version":"3.14.4","zlib_version":"1.3.1","dependencies":"Python standard library only; no installations","bundle_encoding":"zlib-compressed UTF-8 JSON mapping filename -> UTF-8 text (blob.bin only is base64), itself base64 encoded","bundle_sha256":"231f1cfc9e02085a34baec19177545794e5318a09d94c0ec2e553e70d120bd47","bundle_zlib_base64":"eNrsvQl32si2NvxXdPuuswyxjAbEZDe9rgBPcTzgoRPH8fJiEBibKQiDnV7937/n2VVCAuN0+gzvd993JaePkUo17Nrz3lUl/fFLe9KbBZPM+OWX7V/+26hOgnZvum1cNl76o4mR6k4Hrcdhr3s/TZvGJBj3e62G1en1p8HE+jowpo1mPwh3jMN2MJz2pi/HNaN6elw5tb5OHVTvBJNg2ArCzJfhl1/4v4tpu99rbo2G/Rej0e9vjZphMJlJJ0bwPA4mvQE62jHGL9P70TBrbB0aC/iM49ParnHTmHTD24zq7suwNxiPJlNjFJrhS2gOGtN78yEcDc37RniPkcwpOjSbvTBoTc1w2pj2wmmvFZqt0bD1NAFw00znafo0CULzG6s3G2GQ974MO5PRwBijNxQaeowz3H4Znp+eXpZ5mbq7Ax6Cu7t0Bs1H/VmQSmfGDfb5ZQhgMmye6Q0xwWnKxuCTFNum0wugB0/9aW88GQFBYW/YNce91mM/WDzW2HaNRmica5C+DoDUCKD68XnQ0Q9ao0FzFD2o8kY/mM5HeDZeasMeL6u6+X8bldH0HthvtKYxyVi/H5AYwNloGBrtoDVqo1gI3Wr0UWHwJOQ3QsA5xM/TpNNQxJauM3etRus+KOuR9C0HPEVtUBuz600NhSGMwR7BfI1pAOZoBp3RBLVA+hfFUsZTGBijjjHuZYzTSa/bGwKIcPQ0AbCg7EuInqaYA4afBs/TsmDb2hDEgHs2SKZG+47PUqDBqN8ubxiJf73wrjFpGmUMYPxmZHIbX4aNkKAZbJJpjZ6G0xRapctl58tw2BgE4RjTLf+xcXfHu7u7je2NmJ/vZOCNP78Mg+eglSINwC0p6YuURVP2Zi7BgKHLRhj0O1GVu3EvNe6ZRn8ElIcpMM+X4cYmoTA3flVoQZv70ejxtw1zgyNtpM0FbJjmAR4JO5QXpTcbUrBx+2V4uetXD3bP76r8Kf8BYNvgDg0rcUWKTVIDUL6f3v4yNB6DlzLlK9N+GoxD9eBmYzoJgnDj1gwD8H9jOpqE5RTAMYEQsrvR67ChMRyR2sbSoOzU4KDgh0lqKoMYbNCDSICsYMXU1EyBtGanP2pM0+ntSQCBHVI82EDqP5ots29OytMd/XAjtbEpPfbTmxvsrnmzsUkRfMT97a9lddPiw6APvtK1JyhIb7DLJSBvAP1tOZg1+qmNfmPQbDeM5nb4NEhhFMwy8zDqDVN6Agb41pB5LiMHPZvp9Ib5xwYabrD1n4Rdw/t6uC/DvV3/8up896J8s9EfdXtT8AIGw+Vde4qLsNcdgkaQUzD9XXMcoqw3HI5mIrB3YRv340lvNMGlftxsDNt38157eq8LoFob/bt2LxRmo7TrB+ELEK96CGa90VN4J/xPHus3xiHG7UwwLAZC0XAE/XoXzAcNtnbZbzBt9Pp3BFuBNwgaw3XlAiVYu99stB7vwDsBB6c0tXuTIBqAc248342gDO6eo/sXdf+C+0HvmTr8bh7QUJGxyVFA0GDUa6dmwlSz8qDxnNrK5jLmANTi7yyd3okI4FgpZ5PmIwM6prZm5Ft2IphPjRdcJ1VQmGJvTrBVNMdpa3HtbI2jhu1eNwiTLbVFyoT3DTeXT4kJGWul1HyZBpTuzH3wrBuim1YfCsg4G0FPvqQWgiyz+R+xZa1BADPZxj1HHIxIwFBUjZKj/zZ2RauH46DVa/R734Q1qESn90Gs6zdCWgno70Eo3NsEqdtB2xgNAwsPttghrodT6tlpkGHXEKl+MFRjQSEqsZ2bz+aszLIb+3ZnUJ6/e5byCHcmEWVnzPm71Gwz9bw1SL9754qGWOnPXdffV/PFfFJ3ju598+u7lx8YAdVST5upl+SAusUCrZkl9GmU3t31hhC8uxR1sikCXT4BKuSyTNtCLt4wG/3xfaPsZMxpMBjztzXq98G+5T1o7UARI3yCewMXIe5TewTWRrM/amaavSG05Y5ofzWS/I0L5F7dqvHkryqQcflH3UbD618ZnuWdRjgtv6HeSQK5MqAl5Kpc3tBVtJbk1Bd9TUbzsHxzqwYcTdq0x2Vb3YJxpxBgwZWGCDIeRo9hv1rDadle9CUapFwsZbK2m1N1JoPyQtYcJ+O4GjWTGfDrFPIgpCoY94aJcXA3jaGAVE3LGmrSE3pxOmm0A0VPWPzm04vZeDYbL+YUvIbrSVPRzAQZbTBS6Krfh7IdydQpnVcKUKc3CSETvWcj0mAoCvptuLJwQSZGH44rDIGWsoxB96IHT6rX6A5H9EONfqMJfNMbzixwoZGjsDYU2o2ecK/ZZzGF19BHgpSYPMej2SUOlEx1zEljbkKtt8uJepHVTXKP8EGSuv9w8uWyvb2gfqYxHgfDdioFVMI3wvTSi86VkC013yw76+hixPyRqKH4R80xIa7AhCZlwkESYsJLaikKtcHiNxuwkbc7/TEv+2NctqUQF/KUD0NewDLvNJ551Xjm5YtcvuBSClkmRS+0KnDQBuZ4VtYiqfXFjTKzMPE7YVsxbfgVPj9V0XhmwjA4RYWOTvlG25Ne2oy42xmn2giuUg7xj87UnDdulchtOel37XdO4Jmp/nhrPEhbYdsMdQlq3/c2brc4R/hzaVW6EBsBM3Tv02wnz8SF0uVi4dNiDBN8ZxbT5tQCy9sIWRaiaWqRNJdnl7Ed/QT20ozpaKWmm046OVpE0HJigjEKpLPGs/UsyIKCXnnyYr0snmgOFZOA/26VTO5NRt+CodEJGnQEjFa/N4Ym6EIEB71paEIK2wF5lUYM5u++18alQUUc28WMJtGWzD5DcsyMX43oVsiRfPSbkXwyE9M5o+PXuY2EdtxLiqRoVyWFzfKNCkoz6uduQs8l1XoCsDPlQ6pr9vetN07FHdxs8Am87U76dklsV1S2GghyEvbaT1DNCwOQairZJNjby4/1AM1eAwNs0sNFhHfzcKuc2kbTfFgHj0SBAKiphV7UlXa+Uh0QaTM2Wu+iAcHIkeFKL7SA1gwpEWd0k1bUvRh1EJo25luIdZ7GbVYy6JiSmuIk9aB3J6B+0BYtDG00MoRxLeVxbvUbcwM99wYxpSO0gU9ea8nesFWmZJSo+1MUvaiuCBLdCOX1mdPFE8QmO41yxrZz7xTdk7Yt5Ww10u/igs3GO4yxmDiNWH+cMGGikfvt8pIgKLGHJOTiodAxtMkOCDFtlPvtLS2psZkebJYb7+TxwoAmgJnMNvVT9XdV1ypnljHMHdV6rzVN8Qa03laBXTJO44OUSL0p6IhcXz5g8GiCc81+0Jmawu5lXZ6otRiEtVTc9ohwDe2UlEk7kwysAJs3JoM7lYRKLQLUdY7VjspFqGKdqciohMvGqwg1ma/Y1rwiNwgOeuGUQyWrqPBQdQZaNdowClJ7ybdnPyJywhrSPoWGGrLMfNKbBrqy7krF2K9Ggn0dTUcw0uVcFGxoi333MGpCMqHWOci4rEOHyBu8nDwF6Z1J+TwzeRpKPVNkj4kqjRnIVmPDSm38o52RMP/bxj+kvx1JZiVhZLpM9DAEOkwlcgJj8Q3W5wIycISgM4CNROwrgND5Vy3TZhQ6YUg4FDcbQbsbREGdnosidioBdDozeISFSQmF7kaParaoNB21R+UbFdkwRwRFMuwGqUKWJq6QzcGqkL4kfeptJBB2TXtq3SW+431vem+8TipmzlR272w06u8+B2Dj0YSSfDcfTR4DoMYzB2OE1ipjtZIPzHSDafQstQHwH4FBpu6CZ+EhGTMB8t2gMex1gDmBug+MjGDyUhvADVtBnLUHCET0TKCayAiGT4OAcTcQlxk0xqkEK5lEndm6fxo+hr1vQdlL6/QM+1LMkCQ8ekxvbjBBld5hhU7/KbxPKZMgmuIfWZf+I7wleIsbeqANU2AxpXZEtCj2hvhwZsKroQyuow5cLegpD7fFIQWm6JAKUEoQhU2Zu9SM+n0SLwlskkfZe6wMG7AMAs1/Gx/vR/1giyBIGnS6I968qkMHBBcPmOVo8qKc/TADK2O07kcjMU4q0YvODZhsiaynk3I872VW9Tz4kTuzxpoKeKR5mUDTQYiR9BgzPaUsSitpYs4Q7ZRDgBG0UxP6VI/K2k/YajqBekJnkaev64XBNMV2NyQk++RN+t1Xy8mnVeuv8ZiOiVKFTGhiwLXe81m4PbHPIx4QYUovgcR5KVyWbyY3zvKznbA1mgR8IO7h8jNJxSnE/LfxAZ64BGH3+Bl1J42BsV+5NBFcjCFXrolY7hlU9YpimMId4ySYT0cYOmjM4H+I9H1wy66t4qR7cPVjGXFqRqNdUtsaByCcIBvC3QQAcWYqdhoF7tudbvnmZWssxS/mOMKEmq4pzdO3O/flm8j9HosDkFYdSX2pdBulVbtgrlSvHZoyLS2/+2W6dt2bnsJPj81QJ71zIA/uXz2QVph4hxN8pxJpOo+WMfet1MGma6e19wdJl7Eg6gb6IHuwi1/dXD7yB9iTcoPF24OLRc8rvdBOcBSWdRM5IdI9XSDQzty+E8ZrpRFo7Ny/LiKj2ckS1Tg5K90ffHHyJeYM1t/phjfN280ycYNu5ZroQH9y7WgY+uX7ftnO7Az7Kouhen5YlrTWQl+iAfoM4Unv3OPqXq6GuBryStehEer/mveSWNuSkm1agd7wKYh6g3Ipd/vvun0rhf6A+83U/la3L25p6mBLlW3tv4tIE4/Apr8R7cDOtkI/i8zUo/kQE1AqQLLgF4tPvEo3VC7rOjv01Mo3vSXUSh8LpP5axmyVx/fder9pVOjRbjCKKfwrLqMwFcJhVaQcwahM2lFMy6o+JDoV0wFaDIGjm1byH+ky3qjIQ2mM2eY6L1cJqNlcxD2sbBJmNSjt2vQfxYRdY0MYtemyRYPCqVHd9PuGDtDgLkx0ihs+CpSI0Wi3EY/OEGUWtxaRrEoxqoVW4ymUoDZoMO/09akBk2Yw4un0plN6DLQfjQn8BuiYraQKhB5dngFwcrvDOK8cr4pmmKVPqQ7g8TFFdMMqt+80FqV7wAEdvixuSnNzBEpqZHhg3Vr3CUPgKnFQEo6OlmVcetYCw1S0EmkZudGEGeL6gxLqqCzhzaz0JfTRsgdOpWCzS8jaZlnNDxy3xRnid1PFuHoEqaPEfBjMgUcrNaQIaTQOF2hEf6ZqouPw70ASjVRGlzcE53aLg8qVtJXJ444VhFnqTw2I/LdgEb+LqyPZI+Np2JuG5Bm1BGTUHXcnzjNynTRarm3rZV3lWXKRfdQxHMuzS3nyyteImhNm/FOzd3yQlsexYZrexktaUv9256swjmrEy0S7L3rdq/zHRku2EsDvX7OXYDta3aaZ5WYC01jsJkCsoLk/3NiOvBVTZT22xSfQa2rb8mOqdMW2AGXqZMS2mpq5IUBCNADGmewpMKTEmPaCcGs62mISwSQGFxgDxRrw1iSMJSziyt2Jk7mxfRP7YoVbcwN+T6+t19sWFSJfzLn9M45RBCnia8IvV3GU+PWri6nr4yayejCZlG9kkioxo7gnkqZlrlOzT6e1J9GMPQnyo0rQkvGYigalJKWMKHlby7mJknhqqnjWQHFbqbA7tR45GYTAdJwSXNUkqcZWkyZJYGjEOpTDm1rNQDFvfFW8rvCouoaLwZ6fUwxJGs3QxOzfqPpXUMzeJVJz0s2fWj8NZEMDDM0iZ2YulnRulb8mKxwv8nfSmCvXUdp2zBdzYt6x+azx/ZzeW55tlL+LU3Ak65INigyQEj5y+2KZ5tXiTJI1ooxdnK/T8hCl6MaxS699UcnTLZJzaUw6qvGS5tyjO7ECwjg3BOO2PAimEyA7NQ7NF/EU5dkGmmzED7ksgKc7Whqi8Oj7AsGO0jREtKu8WRckLg9PSgzLkkzAaCreiMxRwjtz9GIOXJCFS6JkJJahscL1g+XYv5bHv6ZSDzB0uJHCX0s6H5xhCtiJfQG4lRw1wtZN5MiZq2w5Dl952a/qvLyuc5sISiG4G9tDcwNMPZpsqA0J498yufR/lZujUR+EU1GB+bKQPEFS2kKj5qTHJPErmR1vvSxk9nVLWXvvj0Jouq1X0L5bSlaOo5T9JiIU9Lm8cr61ePzmSEErcGw1reE76AABTWoPTV1frKvMBhcbgvk/I8bgdg1JiEHRCLUlz8Sl8DJKdppBF040t4dlxsGko/Y3cIVtR9mvRA5hnQpP7CWSbAe0KPsHuBxlQ5wsQPTHn1SajS6vIrXDjThmq9OVqA/gZMD7A52+lCAAzyHdX6fOxva4rBKFa9eKVUYxbiH7gdhGFrS/1yYMthfZQWV2AM+NzHDjVl1Ltj66YZae18lEorjO0OLTtUhcyTISRTcRVm5vCPBtOZHaW93yMzYVdOltaSgojJr9saFlHlpfCz/C/Wwi1leJxFtzJT2gi9Oyw2RML0GVSHq3I/m21Ebnqd/fMNW6vZJxVelmezvrikVXU4G3MITBv103+y1hrlhQE5l0cuUT1cu8N2yP5qZg0ISqEM7khkLFq2TTBAvKbpFFjWXui9O3utPNjSgDuNNG9FZmukZvwxqrHS46iS1P/0gME6ZvFBvfxuk1aZIYMCN5Lu74QxfEBzdklm9SocwgvZpnXUxQtGkYpdg59P+qRKpM859JmEZ6xiQe/la2dL2v9+MpVM1CAlpEN32tBW2dwRw3ehPm+pikF1XEPa8CKo1hahL3tCqwW8kStrpNx8IlorUzK9+0daTUFjKjTNnj2RrDRbNx1+51Oq/t0IxSGiyVh9N2MMMDK/b1hqgFDginOdRUucl2aHINSG/Te95+hgpI32zn4K3Dkty1lUGR7nnrqvvnd88C9rM4dOmFEcGzQWMCcmpkJ3LQS2ITc31qvSSm3xQhIIg7M8sqaUFNg7B9gee0UjjlP1hp+w9ifXs9BZWnjyvl0ipjYIoZuZU8Cx79l+wC/XNhhlhVRo8jlQXkd2ruL9riveWkcZU28tGWi82erLeXJSfd42YqtU0WluoODj6CqjtZpBZvlhumIUAzUEtpKu3jRqsE24uFH4ZC/YQTrCIXnZRfeUjZ3Nhe6NzFIO6t7FhZ3GdhFJYKPBQsbnK3q/1q1AjvaP5I9I3av5i/tBBANako/gNb6zOLTeLc9S4b3mXre2KHe7QBXsUi/+qO9p3vbWjf0d2pnfpUnbUvwx8Qko1xrz+KV6veFJDFRuzFMkXSSC20T7pcTiT+8h4D9bznuLL5RgzqH8qpolOmfSV4ZTt0LkLZ/PxDQSHXXco335bHlrtvcrfwViQ8uol8DUn0/H3XsjccLy3mSNyp91iJggU0Ymb+RvS5dhvJUhRay3wnDl3Z1/xmQPrjm0nW7yNRU483mUEvmhGU5ovU6U567Rg34rIK7TI508ncLrKNdF9VecHMFPkM/zn8zapKEujXMotNK8Cwyi5srm5akf0qivZRnQRQEqUJ2Cq0V/5puZZJRKkrTikJeJve4Uyiqf6hne9t+TGV973Nv+aq6yt7yCWNz8UYdrJq/RYtbm8WsZvoM0qEDuH/UJ7/Nv+a0eCS31+EARoIVajCgT91rB8uegmDvt6zvb0BeHqDp4GhxzSYeWvMjRZugvZWfzQa61VQQ7QApIzQb2zLHHRPjKc4IkaKTnNIkzsJm76bO2CNN3NpUWdaQ3NH0F8mIsI3e1O2T5xfUfwK9v+o1l864sT9UAxC/vVTSz+isxUF/rNKO2e7nihtcYLKfzyKj5b0Ph+XnVGtzlf7tuh4ylW8EL5Wr//JU0lnoFTQ6jeoWcfxQShRIPfxSgzPnQRhdDDKmN8HKmpB3xk7ZzCilR32ahTQbDiCLASJ7eISCAnoNzEIv5Wjogi+2y3pT46nRGXakL22IGsEY8mOqK0Gay2gqQHcFo2gb25jmVNNf0DmVMU35YR8B/baWBzoU0c/Fk7SRiT26lonoPWThIU0X0/WfA2kFMn5OT1SdA4MqkwdiVM3cqJOXS7SJLcUpSD4FvzoKsYrAUZvT9PWxnYkm5nFxXA0Ty1u+OcbIuEMKqczvXAEZoMjl2IGXtEBcZK6wHyemLKIOGaJ8WxyCuOmVfaj2m30+5GqlUUKU/FUQgwUOrkOfDcL7zRjKHYETeS8Chils71yhEXxR+c7Z1iUyyJ7Ykn8P2OmUvj9Pj+pOm/wkw45VB2leBMc83/K4Ybaao8G/1HvejHyknutzgZdTniMsJaJDwa9OrWS3v47B0+igwza43x9ZuIdsxz6QMtSg8h54fOb7fxtOjqK9+rgguoCdrOswE/v1DIqWagW17gncaQPeJbHmUs8Tu6BzLT6QWMirZaSR2ofZbusa6+fZeZyYZeiITJ+uayaZnzZyrN4UFk8qCw/OFs8OPsy/Ns+/T+rw2k23q4ZJfWUcDOaUSxyHhmyY3VeQWd9a5kowSphiTwqJw+Vocars1EAHaz71KcdUQi4iw793slC813wFX4wZGdjmzknADEZdstKSDLn8pNys+kdOYdUy7xxEkkFYr3kcg3+CaRqRQedZthpcqvN0i6EpQiHjKxprra/c9Gg/Eacw2dvhjrR9JXZTkKup4/mPJM5kx1/TA4L5D8YWLY63bJQlHloccUXgQACFu2AOxlx+RvliIA6fy/130rcp3ea5Tf44Edax/hrZKhCUlz0KhZdM5M3nRJQ2VwUY75FhleubLhSidPWypaIcaw0brZz6OpW70QsNxJKotVUSXizhCjNdBwXsZvayDTHeEv1yGqmUyhkXDOXzThmqST1NMyzcnlupnrmzJyrFKRQMApZZJPf3eIEqjrlus1gipDpbckLeLkspc6OKF3K87SLtr3hrDHpcfkCFkrtu+gEajsD2I6WNVIgd1qeluuJE0qZE6s+egK69YE8cTgn3HVIR/SJKocnHtr6YDwth9roIefIIh7lqfC7njZm4MUowyGbVcmI0V4GkE3+usKGZK3E2s2Cz9LpaLkG3LRS5TsLTnGriIfALzEdVqG8YY+35cZWc+EqxDsR/sr9ZI8Lz0DfimcQ62I4Bn98+UU5dV9+2f7yy9/cnPLlF/PLL9H2FHRw8+WX6HC2PFLHs+Xy1QFtKV06oi0lyUPaUrB8TFuK1h/UVsPIQS7dU/KwthStHteWwpUD22qEdUe21z/RcC8d21agLB3cXuBjcXQ7LokOb0vJ8vHtL7/csnOobqKXp7EybtF1bK/oeq5bLGaz5paXKTgFz3Y8N+86uXzezZlbbiafK3pevpQrZgte0fHMLSeDx/lCqeDhiZ0vOGhqZ0oF282V7FIxW3SypVyOZZ5bKKCuVyq6Ja+EPky41PlC1is4JcdDf7mc46Asl7Wz2bxXtIt2Np8tllBUKrl2Po+aOfQPCKH+vDz6LWW9UsHNuU6WiaaSl/OyWXjNTlZ0pJvJ2U4uZxcLrmt7hWwxb0J1Aa6cXciWHIxd9FBSJJA5N1tw3CIgLpheBvPLASC75Hk5t1S6NW/sTL6UdQBq0c7lShgJ+jKD2nmH+Mrni45dyjPZVcxjdtm84zilYsnOo+mWW8pkPWLMcYr5IuAD1koYI5sHkvKum88BvSgjxotoCww5eUBdNEkYGAFCV7RRli245lYhA9R4OZvY9orFEhCeyzggTA4Kpgh4vKwjtAKuMFEXKAKSMRXiG48KtuOWck7OKeSLhTywVPDyOS/v5u1StmTn7KyZy7hZ4DVfxPyyhWIWqisDouUKOQeYAQaAJEJWyKIvB01KILJJ+NFtDgDkc0A45unaGccB6J5ne6UsYTOBDEwcBPLQjVN0HI8oymYKpbztgI1ADdfLqQngF2XZrOeAUh4AA7eBRHmMWoQtygJkKcsW3WIJ0II7SjZmwjLbBb0wJIYq5dyiQw7Ml8BoXqmEaeWyDuvZmSwIky25oCOYHFzIMtsuuJ4HwheLxKrj5E1WtEE9NLPdPGDJmsIS4H0gGBQs5hxTRnUhR5xh0c0rhvDydt4lihwPGEQR8QCGz4GhIXHALmbqeWAgTAK4A2LIqIV8Lk/+g6TYrucAS6VMHlTzIBLgrXy+5Jrgorx06xLiPLgDZIEUlcBcIJ2TzzMDCyA80ApC52E2HvgKlj6bKxXykGQwpospmk42k82CN2wwPfgMyAE9MS66wdQgpjkP7cBoroveQF5ICvQDiwpAFMQH4psDxlArj1nnIGFFzDMHmYTjkIG4ZbOFEkwgpCePdsUMJp/PEaq87TkYrwQlBIUD/GECQH+W/IMapRJ437VdEMIzXegg6IUCEYA+Sw5FLOdmwF4ULZvDFkCwnAdQgRToBnJLyWNJ3i6AyiQues2ypCTiBakueJiCmctlqAMhGlB4BYxo5vLoJ+tACQIRFM8ciyB1EFLgAjxso2vIB4TRc4EpKMqSmStlHF7nAKFtY35mHswBPZQjPzg5kjjvgV1wSe1D/YESULiYI92gx3LozCyUwAeQU2hnMGmuUDBLwBtEF8zuOUAuxC5byGBGkGuoP6dYLIhAeZBhhzoCYFL2qb2z4GzoDbAT2BSc7YHZ3SwmbKMthI6ocKnSQS2oMQxYADsDYAoUwAf5oA6yOXJugUoHzzConXcAPliA+gWCVAL6gS+wFstymSwVBDgSoo1uhR9BEHTj5mwQPE9lAr8SOHCh8mkj8NcEYDnPA+0gxtRC4KI18HseSEHFJcCBXp7oapuiyCsTwqVKKAnQRi61IDqEowvuLhHmLNQyfwqiIaF/eQctwTuoZ8o5kMMfGBL8FKBQ8QMpZmGxwL8lsD/GAaOAh6GEPRdkLeRLJpghA70MUYeadws5KEBQOUdzB6LSoIClC3kg16PyK1APYPJF9EPqOYWcV7ILVKVFCFoBeg5I8bLgB5SUaGfz1EaAGuYHIpMBwYBZj7QHkdwSmQOqEiwIC4hnwCJNNqxYEa0IPeiDnjEMaAdL7AKtWToEYELIJfAGSwSF5JDe6BUMDkAhyyaUXwZiAB4ricYAhUAiiH4ekg9qFuySIAWKBZiG/ORgpGBhTGofEDtHT8AhW3ssApDQrzD6DhQO9IGbER1WRDUaFNp4FIG7ADakCuBBvqg7KdSIJKCkMAUpgbqATgQBoPsh6yxyWUXp/TyH8zgX3FEFlgpSBebGBQJKdEQKopWhTyDe8DSK1LYogKPjsQcoEhf8yyIwNeYHdi8SMOmaqhHqCnwK/8Cjic25UPWFAjwPoIEai84D5wlDApqTHbhcRi6FooZGhAdEDwOmh9qxBH4HA0iJSysCLwhWDYxLV8ghROQEkI8SJUVgT7ShJAIxUoRhoHYKsJBQrqoIsgstAX1OUjhSkoelxSRpOpSTQwsqeIPFQDWWwEpAddOzyNG2S1EBBIK7ARqIS0IZB/YKdBftEtiVttCh/oBiBypsWB/IfYYap1AgvqCGQCMR1pLjFCguHBRGRuQL3iQEBAxZoiFhUd6jKwOfB2yvpRB1yO2oB79EuWiYNuhAywc8sAA9AL10CmF7xZW0yatUHsAoTLw4dnBbbJoiYCivkMyOYIhxm6O+EyQXiHRgHJa0lBNcgeo0c/CBbIUECh3q0KRAZcmM+YCqk54iJgO1QjUOgcOkS5ACKFy6EEuabAtKA8odpLFB+yLB3CpS9+XBQ+BxlyhlUY7+L3Q5jB6Ud4FFcJDpEAMhWVKERfSwIf82WJcQi1NJ9qAXXyK35VhEgOALgTfgQWWlCJwHxBfpwsPUeyyC11AgUJAMQCcOKuB3sxQR0BRKECXAEzUrxBeEABdt5TOo65Dz4cTDduZZBOKBQaAoREw8FgHZ0r/N2KEotWj9IIKwTzYN+wI7eVomWAEgOqexk0Mdqp0CamvseDSxcNjgMBY1doBskKxEzZPV2IHIQ9BQEbKgsQOGB5tA6ktOhBzIFimcpUtP3FDxQJ26yltSuIFwQmUxsPAEE2DGIgQaqAA/OFmNHMwXaKB3lo2QA6cDej7vkJsWyAGDk7R5z9XIoXmmlw8W9BRy6NLlGQjgt6CYiPODdEL06I1S+GEE6eiAf6EpodxsFuESRATbwEekvIhVxExgAOj65nPaMMLUOvTd4YWxDuw/3VkHPgm0eYHOMuMk0MijaoH6YQwHxZ6llwwZxohih9EpkWWTUHRvKMJUwPREioy3CqwFFQ53DBBinjkxxGAitEMt+ml0oCn6dH0ohwUqnJKIDYNcyStKlAtcF6F4YcrEUfG8ogfXiJEraAHrKY4/9AzMHvQUvIoc5IMKCSbaUU89mmyXUSWjQx0sQJQQGkDJ4C/ELHvLJlkzj7/wFWUsm9QBF4AHGFWqznL0C0GOAuMKGJ+lkUDnAj1ZYJVqypNYBZSDbwUHHI419BeAu5UJueLpyLScPN36Ir16h3ZbQFQWh32za/qscNXhSUlQpqARMwBbhkDLIfLzehYJjNG4gMLwKxgauTkFKGpDyYNT8Zy6dwVhEg/nGSU7VPlqLI/BGqJ0UAxIKLiJsTTGHLpDiDShhhHPZd2oN7oB4D9wDAyPMLcpbqdHHZgXRAKhdKrYAAgvOQygICO48NSkCNxNUSEDaIe3JBkOUFBJAUgIlU83F2LGuTH2igCjmwGQaBng+MAPKmbVnOBWwX/KMucB7eJ6S1yQV3MCE2dpA+GtIibXoSa4HOi06QjDE4VnGM8K9gVIAifmGZppdgMe6d3Bt0E4i36WeECaweVAtzYdUeixLLUtZJv6M89IEfTCVAtLLMFQJ29LMANnqKBoywgGGgP8DcOM4FhPyZXxCoolaLvBiZR0/CsoFJEb8spcQ/dCtJfGggMFwcOYjNgAkhoLQ+AWcAEGBPxFPVhJT05RE7OBYc1TueQ0+mhg6CS6NHAMVm9JXFdPgKwMISgyN5LVxAXdwD9eQUJ5wKbiBC8xELGQQzwI/wLxpaNHgocMtzMnJhq9FRVX2CJgUHNwA7JQ2UVKelazH1Nleep/+F1g9MRQWYUN0pvJBMZ3sGpeMeIJ9FhiosUDQuCoJMeCUWWcBM5DIKhHytLth2sFA8w2iYFySnhthkdQpGgBUKBVtA5ymJCC6wVn3GX2JjkQkxZ4DgUFO1rUdAJtXP6FAgIAXixRjp4RnFCIDkCEJaLLoZmWORX6jRgPUErQrhtQbGHA0JCYKCodwUiEuT6HrgZsl56REEmyhAjuwZJQdXA8bTqvgjfGIeCjrPjs8EAiJhLGg6PhUYzg+0FCNQaYCSsWmADKM45M0MhW8MHQ5LI2HTvUANK1oQALoSdwBLQxGDCfxBwCJLAirWOe1NLKyKXswZfBA9pJZ82UPHFFKHaMVzytrkUPlSTL6ogqF/i0mSgyxwA9ygje0ULB5KuCADiEy7BmJGZZRL1AW+b0OHTwmd1DqAf3PhJ2oZ5kkoBm5m/tYmQrS5RXpglLcM9Ka2jk0McFcenAFRjYqoFI1yzFCP4VvMnlGQHbAATEAtYZDksxE6ygdIG5G+hxb81QcEDh3iKE84T4WtcAMjyBLmRyyikmFRGcHMZfsA4unXRHGzTaJcSfdElzcIXUSMVYO7hUK3D0oAgg27CIqhnYMUs5gtRS1vIJdrA5MuIXm0kKqDDNEMQ4Iiq6DhAvBDrxpAqRNadTyGg3L3luT2ZKsoHvS7SrBZHYSN8xL4bYlX4/fC8EeUr0i9BlsEgYGJofrtaSdlDEh/V0SvTCCuQyjXSqF1hfJsiYRraXWNyTYEwcSzqpGndknZwkA5nvs711UwLPFUuSdBXtKlMCvwFyZszpXbiK9zRoEt0ze0pGymvbZNMnpLMHxQQH1VlGnuJal7lKZpDgdkLr5IXxS0xp0slhlCvuXYL3XEdSJVD+8GVdCRBIPrhgsPXkJWjYpC5aDAU1lWdYTTxCS6ihpB06goUEVUpqUo4dSVuBZktSmTBDOu3MrAhT+dQdGG/dSOBK5iZLCH0gfkqYwSF0PKhz85ICi5kc3MI8G3id1F2YMpokl+s7JWfBELE0gUuKNBYFuvXAk/gv9Cty0DekEt29gvJUFFnxFFVzEqMhCmcmWznHBfEWocHxL5ctrmEI1AbX0CHwJPMh0pyn9YSPlGNmP5/kcWgGWjAaVjiJTHsJ5tCYnka+ROutZEm8FEVepa4cLnLQtsLlygrNwRvoinwC1xyOpuBOow4zZPaHekKQrhUEJJ56nYsVAMW1NTu4CezRKhXpRAL7uYWHCL6H9uTiFEa0S1rBaqOZY1xK76YEvRi5AS5FiT6bK9Mueq+UEaQV8llgMkYl23VDSjS8JGCH9qHoJbURZM0lN8B+oV1Oo9WhLYVph6Vj5OCWXvOeDYYr0OahxxLTcTYzVozOACcUG9fyYrsOtQPTUxKhES9DgwYfyZO4hvQqrB1GNGPWFSdZsbhNP1iWVhj5ZN2F2LogISNNh3n2IsJuL9iy1VCAi4E200GSXxfsgbsEFYpQFFqEfAVmawqutoPQEFCGBbIkKuTdZbmV2AtczeCd4Vyk+CDNTExDXBDzZF+zOVGGJy7ULE2YuCSUQjCgwzUQ8mZScLngBysJFBYRzUVyy0w3dCfgoxb1cmscvTwhkFUuuC5Q+RrtiDW5FFMSMItugiEgDFTc0KEMp6HMlbCDswrMqoKJoBTX0alIk1rgUhP9jZIMz7UqUI9Bqlssugket7lcm2dmCv4ZDH8xih+hFiAOJBLTdsWYSgsNAcsKPVUgmYpMroiGICY9hutMbYh2jRgPbhLQRLS5xJEeB86o6zJphUdZt7BmGCAIYg5KIMSEQi0I45W4pu2Rw2B8SsXbOPgBL0CauUAqxIscQy77YI5clYY/4BXXIA4+oC1uQo4ZaL0SznyTICYnaxyOwBcpV5bQ1YWnR12qzC15HrzoSE6Q6ZlXzjjYBb4FHC7Mh7GGDovgkCCGYohBP1EyjZEzDncIzAUEMEDPRakKaBHYay5VA6ugtppTPhEIMvyFPyBsDg7QoarLpSxAnGMuF3KaVENEUE5iCI8LnSU38v8AKjQ8pIShgld4rV49BhjUdpRGaD/tX1OaQAv4jKCGF/Ge4maIJ3mhkKPc5IpRmAVVLktMBRq6wppQUCIfuBdcA3Gdhf0k13NFA055gbFgUpqA6gLNekkWC3RGwbUL3EgBJoY7Axxm17AfA+MSZRy8RD9E2I+pREY60DeRBxbJE+jATAi0FwR3EdGB6SjlcEEZ/xTcwhr+YyYawRAxAivryvDQ+FzOJIbwnxM7RtkMeJRr5vANYVhzDhSsYnVIM0xCCbqmSMsQJb7wX8y5mDjZoMB0ZClyqOD8k/1tpv0hOkUVAQimXK54QLGWmH/kAqSOpLnUCWbPw/4w4X8ruT0ee/ryy7aat8uhslyU+/KL7COSpJ/NjQt0Nou0xcw5Ulgg0mrFgJF+kUGpDoLoGpSiOwGDq3TaMpQYqKshVOzFvJx0pJINeW3juPAli5L6Sb7EiCa6K0FcwXm2ru0wM5W8Z7QNlymv+4UipXF31FPqLi/Rd5HWT4Os1wskRKHTm9Vt7JwXdQYMMaftcTCmzRmqSi7FXlSG9BS4XySrceBA7IsL6JaQoAwL0355mbv4FnFvoqSzMXIlklkgHixWcBcNmVUtRIimAxTNWbasFLIxFSS7H/dZshe9KI2TjycOuLN2PFuu+xYYUUqB6ivmBFurhAUDKJ/NU3wjmKGyTUwdbiNdwEVlLiJF/Qmuc0tEyDHCitmFWdNCErFRZc6iECMDVhQyXtLkcGPK0Gl2E4RPCIAgewmlnusl2JxZwQQSYcOYPnAWbJ94luxW0XuBQKV07QSvrjxMgL5uxm4SeeByeNWJjvN5LVHLQCiupWteLMWkiWurmKO4RMUVmnLnRLI2U6kx1VZ6VtsuIvHjRpz8EhU9+r4xergIkKi9JBvxhIUEsSahLtSE4rIKlxNzC5ZVfseCNZg9y2lt8Rr/zICWEhTg1rsEPPkEgvPcVLaYJr36YkJJLassbhZLcjYCjqQgOHT+vZh7EyqDsX08lNZQ2oQu8PyKwlElGPq8neBF+vHZhSwwce/GmFqSoCSfym2pECuJ1bFXybR8u9w0m5DYlTESmmnrFThvjR8tAUQabmulJ5XVyXP1ZIHAFTF7NfDq3L1isvP4drkrbTtWnq+K/9KkYgP4WnW+hjJG6xpFouJPzSTLBetptjTFZd232vWSUYcBjcXNXlZCrwiTVOBgY1ulONajeQWGVcQtAxGjLiJp/m9o3e+z6AogXlI210xrhT1itKvcTMLDeYWe3JK5fZNZlCKTfQO6Y1jOpRkgvix42quRTV4JUi/fJCfjcS9vIUZDIalJ7YTey3LnWVK7IWzTEqUsd95LKvmkQl92txBUStJpgexiwbOXdNuK1tpKwqw5O59LWlgukSxMgWNLXL9Qcdy2kYtvnULRdRPOFiL9hFWPbY+YnoLkvPPrZfUV5Vbsaey4Kfwl7TjX9LzYknFfTNJfzS2pK8UniSkk2Qv0yiVtEibH9HMsqjnEFgkPe9lZX/Yul1ljydauqpKYRZjvXnDdqpRxm1uMQa+Q9MOTzkxEVm6GSjCd7IOOOWfF/Vuad447wBL8uaxJV7k34cttrUiBclOTRmo1FuEu44RvbS88MjIz4j5n8QxjOpqNJdiK3vIpJzr+mfd8yjmExJs+GZ8tverzyy+r7/pUVRIv+/zF/EVeZRAfy1OHTb5OHdT940/0Ied1oht9DIy3X+SEikC/KEUNOYyFUkdqD8b6kq8ow+WUJ+x4VEIdIlvpZ1H6Y/38qd5Zs3Ku8N82AWifeOiM/TfBXm1NVK8e+P4+qP/6YNFRcoyjvj179jQJtiJeU8958LoZDOUwYXPSaL9YjcFgq3Xf6PeDYTcw/iffZFqmUzRSuO8Ng8wkNI1WZzDgxRfwY7MHJuwGTxN5AhaeBvpSDuHgMm0aPP405deO5Puu90H0FVQ5/hj0+nxHdPwxXX6P9fxk3wink6Ax4OfHOh1+LCdqfP4UTg0FjomejEYzHPUxrn5FgTEYzQKjqb54pr5zEEx3vgwFd5PAGPN0WI8f5JG3UvHAudTk26VETvjt36Gv6r0Y3WAaGo34g2mNDr+Upj7zKif62AFfbzyaD9Fm1Df4XTG+El1hANPv9x4D1e5i3hgDbeP+U8j3Z48mjVY/MGa9YG6kyNVGBzAZ40kPyBnLEV6ClrWN5jiUzjdCIxyPpkboakD4nMeYqUuG6sRU9LlEmZF8cwgwsC91vjk0Us1e+64TAHuN8JEX6dWvE8tRcSM6K/5leGSU+e2ojG2809+Qsr8Mj/1PeyiHcvsyPFFX0Jb8/NGQLwXo976pTxyxA/0tlH5jMO6kOunFp57kwDX77XWMjvGrXMp54JT0LqW/GXItxZ103NviOJ8mlWlA0U2DcagOsZpyPlU+fmh8bTXGchmN3JoA3uVjvo59h1b8v7EpjKDfYsd348nZMrRoTTJPwx7feZBySpA7w4XA6XqY23INO5M3jfg539XTWK2hz6UK6pjvjd+cB+Cjd/EngTA5jhl31umqSuPXE0rMYbL+KSYafWegPel1pqgEs5YDkVX/+lfVGDzd9YmD+GttEUxpY8tQzZyMq/6qJsFAISX+TCnu9YDjiKXU7bNjGi9OVGbGXCYPgaEX942HnaYJseFnkIUL1MlegXuBqwjOlQYROzYhH9F1Q1cKCEvgKq7W7YIgdFQBmRy3hnzS60lOSV8bjan6yGIswloQ1eFZ1cm8MZw+oBOolcZ0OklFrLvBB+Hdw4ZpqFd5LiSEQ0KjJEaG8tTKxbDkG408r6rrz6T6zEkCzvcGdvihaLloJB91Bavodss42dO1v06mEc14HX8tIvF9EC1p6QSLjo13CUorftpcMNJ4kuk2nsIwpQgIqUjHTf/b2NraMhbGxFxSIPLNyhSVH3TmJEh+GQ90m/KlIkBImJAVzgnsYoFv4lK+wsaFhhknQOa/4TMRw3czHqFFqksGHvM7uw0+eHaBmeHzchN0hYe/GfZKV/wnrXgau/FssvU7g9tp068rDieL/hvPO7h9wS1BwAPcBtM7FrB4C5PZMaZyK8UWyPa6w/EE7FYmAJgAGk1fXtfhC1VZbT3sS6IGMEwOv0Mp2CpLu7iJvFsQ+PztFT6fZzE+35G/LKBTYzOFh1sYIr1mCt9D6gJXm2p6XfcVwl5GiuhbfPAmclBrS2NoLXLwHBCIsfz34EkUBvQaeNFZ5kVnDS92F8LYabzCTvct5Kww8DL/Omv496/Q/ZqPnbf5OEEf54d42Vnl5fV9/gg7/xhLL5mXBLmciFw7Sr/Gd1Cfm2WMuKN1vtwAjlWqJP+tmqHFCx/USyEM4nEKRT81NTCmIa+BMMZmpKzgBsg7U5SNUK9y/qc7S/8AqN8zgLGUO2ukPMGpzR/m1BXVsEYzOOn1DPFX3LpgwEhBvKEfnLf0w4/qiL+tJ/5Z5nsZJZhPoIEFfxOk7zKg4gyyzMvoX+W/H+7r38F+yi1Qnk5cSh/U3jG+0qOdZJRTm0q0mt/3EFN9BU3E/Vw198Snw+bv3mgvX16YJD8Hx0ZwlFd6olupYIAzTIdFXqYyDVLiKNPHWUUCUMtAJPnxVvUpNvb0mzzbjrrlzXLr5tPL8pQlXMq9GgL1tok7IdiEt6HxSb0y/Po1UbqNN22OmAEnElmxAV26ci8OpdZdPHCVnccDFw8YaTRo8BprfJQ2Pd8GsDmRNq8rPHG8FJ+mXmjupV+iJC2mg+qhLR9sM/4LXKB4VDBmGW/1pqPLJxrcKLiUJqrwN9Veyp+cNWz7JDNlnS1UWKubnugIOcGW96Z6elHeySarRv7L80I/QSE8jxbe5uT5Tb3zPPoLVQMfBPz9xH5QF/pi57Wnsn4Kzg9NgRr2SXHCG1Nw/uUpkEmenMQMAlGCQXtHhUPRtVKU60jyHcW70Kkyi3/FoD9DAT45S0pQP0lq1IfJD5r0H+nu36FUjehD2gkdEQZ9BFefXvc+fY5431pnerrNN90A0VgL5fEiLEMD9kzl0Vwoj5eF8niOlEeTyqP5lvJofkd5zGLl8bxQHtPn9MKzeKU6ML+1imOWUByzpOJAA1X0G9uq76ivUxozzhA1tvB4rbzN/lJlLEKemftPhTxLrspfqgwZhH42HY8fUxkz54emQJUxU/T/Z7yyH5oCWUPGiGbwXZXxMvrnVIaM8C+6YTPnlev0T2mNH+/t36E0dAYJOOiAiDqNxLvGUiJX3knIZFQ5cFQSpxy40VJAmZQw6e2W6fGieFaecAKNWbnBX+K5LMg2X4PcaZY1EFaU5mWmMfA4g7KGaPXR624kHVWOs4PlRYawnEwT8n/qbZR7fN2eZr1XryrF2Olt9VpRRg6duJpKQkqVdw15P5v+cjirmvo3rr78ytLvNolBO5XVg/rTaKo/0W6oLL6UMLVvjFPO5iAtPmaz1+btFm/1QuN9YFyeX+0mE5bJxQ2uZwhQi8WBN3AwiHAgqd6/xIGsEqiMd/ZNDPTCOziuKtJo6EAXHC3sI8wdukl/PKQaIbcn3PjYsQ2ZFxnTLDiS2ieoyfRjJzZjutZWVAstE5lFPQMlNq9JUR0NxsEKMY6Go/lfLeVkDB/E+iqENECaYNJ6UrlkeYseKpKYzRfjSXN0E4+HpJLRDPrsn3Tuj2hZQWbd0ehpyrtkN7yXbnTrRpNLZcIWaP2XdO7TRcHvfa/MBG4S/4ItPperJ/2LmmSJuN3/FuaYkDlC6kch9skefY+mKnu3KIvrf21IRvx5lYuANVOyZJPGEuM8oXil6n0vneS5r02dX1vhONUjRpo0l9onetxa36PGoOLjr9SToZkQgK8NtWCu3yq+WBVWb+SEszIfbXGJFEHpcKpfI4nB1SsOg/ZWAwzTkFd1d8JgKkymPuD7nZdWGxK+6zdT7jP932sMF6+olAWUp76KgeP3VQ5TXcSqT02+bNeKO5OV6bA3MP6HJ6sa+WyRhiUMGpPWvSVvp26DiF9h7OM34afkxcppSM3lfETphAOnp2nwG3bphVzOG7PulppaGM3NSB0fforqTYLka7blA3gQQ8xErx8PgImLKbprTNqGvCKeKnTYf8kYJ1xUNuJ3uAMIqt4qOp32OnybbmhU+a7dZn/UejQTr0tN80sZPb6Fd4run0Lgf1utgwsZo/d2y8tMd9RPKvEqUyYaGhEB3GhBO6VXmtKyzr2nCE08tBqTCd/RypcqBzJjjbUNMJIUDoQ1+EbWxV2qduV/uKteXV6UXfZ4iikrnSPsGLStcDriMjsNu17KV6vRRI687d2gyQYO1ErPwuyolAX3FTFT3sOfTKmUE7nmmhcwPGzdk/FT457J9VcIvm1AlURkbEyN8T0/L2UD/Db/ih6Eg9MbjwHRdIQ21IqARn1QRpbUdMf84iP6HUC5zKbrenWkV4liZvJ+9tXOB1PduSxSve2FRf+4b6GtNi6AcAoHDUVBAazDb3Gnoq0G6i2qRqff6KqXB9Pl7L1r3BE5mwr8NG7R2DSasgdh3gsD5c7yg3WQ56lsf6Dd56v0Q2VzIjbQLzmVt/SDsJURqKFwg/bykbx4S8Z4EmyRihoJQTjtDfhee+NkxNf9d8mJ0ydwA3nu9VYD/Tp6LSIRW2vuM42Lw314kR9qd8fy98KUDQHQh/cg0HTyNGwlbbESA904/bbrRjm545uIl2zF6qvn41or1m4YzKVcXqdME2N/GcZj9Uejx6fxnWYaPeIc/7kEm/LUNjUJk8NrB05eHq+6MPAfeCmlvkLY6j9xp0daLTQHgaZj2BsNdyKOGPO1xfyONsW+J29YTth1/rugoeMcLnaMRwX5cvL0EcLGF0lfMHR2VY6ybfxWNi5uHpkKvd1GFeZR44bTqbZn3I4kBkwWelNot8Vmt3SlUlH7qCxpv6YNLips2C353AiQZJTLccKOS+X3YZqcvRGqD5Ko6arPjbQ2kl35plGJ5niZ8W/Q9a0Z3VbkNrFuPdeQy/OTj4BteQ5zmlI+u4+M8cel1es+I+W5u6i1tM4XzWMlepU2i0H3Kmq4rcVd3+US3pi1+m407F4FATSLJCmqSojO5IJNGKyMJH1I5bvxKHRTfeZZpBdSWE3H/ZCYTk9DxldSj+dpjTfBC5s+JB+7eDx09ROGQEDmltHbkfWhMR88xB1TLpr91GVzNZOuPRc8uend8ntb2s3pkNwsFJ5RD1Cmn3aFGXQbxVVvtIseStvE6hJVJgDyhTt12+mUjVWptOOT6TRu1FSNKmsbVd5opGcIL6iZ1BJ/w68tR3E9Nzdl6OLqi4eyvdbZ3THaFMlp5C5OoRynOxyjIZ6u9LdDLY8bLkUssS3aQtydFULJgsxWpJgXjhwU8D2XStryeGWVRL65MZwuqxkRgdHSZiLWC5c93E6TQqBiiNV6VrJeI72S5JA3lEd8P56MxjBw0yDVBkYZjMBzXoFlvDTGKyjUx35JHf2luLjq9wHp6902O8bn6EJb87Jxc7tcl44VTMTAVF9+lzmsSXqF7eTuHPWZ0pV/fUnOzJmouu+lUn1mxVQs21Z/Xzf5Frk65Bma1BTheANd/Pc5GuLb62d672f0ITyptXClXi+KfVZ5xKy9LseHvoirFLXwZ4Wd9Cqq1Hi3rxsvpqQiJvURgxSqr0uIr0ly6qgv/myHECUNrZf4aF1vW9GKisi+XYOquwRuFaioulytrUfiV8OXmEvzO4cEilY3YPTJ5J+B3oybj6i9aN6m9abXRJrTbxIT3EY9uVmRGM61zxRaityDnoXp9bWsdgvn25kV8kX6BB38ZqxZjlzrwK9wtOgIahQSml63+PIZzFlkK+GHr6GxVkqM1aLATX3RcNvgxu7InxePNfxLl3WNChP0NJLbGPvctyA7LhdFgymKRv32Xbiwt6F7v2aursaXZBOYNcjw5QgqFUD0q1vVUzq9tgMlEtLBZLmDyQ900OLHGlQXrxSdbr6q7kI9yqqyWwOdOGW6+9RiqF812K8UaVzlt6jK2rWGuN/tGANR2zVyL1ZtwUqKeUyNfFP3YCZ6faUPFJyo/Zf8Ks6iZtjbH1Ira9qvZ2+ltbQHNltVY4tuVkU5SDKmTmSu6Rj8GbFRsMxGwYKNoorptR0oD1z7Ekv+g4Qr6/BHmv3bOY/fmZKOf40BXstp5LKowluEUWhh7Z2VaXLN5Q3vZjkYEiqwF0d4d+ECsUCAcb/vst8ve0ux/0bjHEFF7knkpiUGjU1HeokclAeE4HTztiOtqbP8U350LzrdsB2nfFQaecVDidaLBAIVoMK9CBP5cr0mk1px2EiQV5XG92bCkVFMrh+mVxD0WunD9va6T6OnKG8kjjQscW8rSllGml8H4dCJw2TAvGr/flRVSPyzsr+hQc9+BTGL4F8hCDZjDXJklPTyY5uqZHkb+GIceNvNyXfHgSFaO45o36XHThL7s7j8Ta2oUkyo1mhIhklMM5s1JjtGM3rYXHnYnCz3wnB7Dc+scFq8ehMfHkg1cDtX9bc8kfAGJHoeeSb2Kreu64PJseU+muhj661O1uV6loVda7Boha6h42OVolokERhFmGsfNNYk8aMVpl/MX+K8sXwi8nIpTd9vTKdctwufmvBfpk9Tbs5rN8bkfvF0mA+n3luTnF+k4pZy5lEeHwpgKZO/kr03Ug2uYk2mW9kNyQYO240+vwtaHTUmIe3HdNIbbLmMiydBo8/vlYZM8st5IMn683tK4onVdquntd0anm7xg7wQUrUSwPT9hSN5OiBXvuBpZButQqtZLGYymXTGWJ9wX86Uy+dSo5SiZNZ21M/fy5QfDnvqq2fiw0cTYpJwWyXjnvp9Y9xvMHAb9e8GXHnjLJgvTtkm/pfWJ6T4dto8z7/zdc65Ykm81PV5UabdJ0+tKXMGTIKiBzmg7uWMd+9oROKcaFL585sgLrmb6z9OIQ/oTw4v3xh++RxXdLaKkcRzvCwXy48c8+F21GeabH4Uim85c/meeZ7KcVXzs+Xm0kjWslRHk06KG2+cjOfwkzK5vJstZPny2HTieNeaKHR7cTpqZvzKpNpCXngmNtUfsXTAP6gs4WEUrrwKmndEm7HJIjRWOix130sWfUMRZyNKmheN+IjWN/gb0CKOu4DjG0EVg8sM6kjpFGapZxKj6ehoDGsxFn0p/UkiGwOo7U+MAqA6xSP6trS7giE0U4GM4CZag6ieBbUpOWbDxrJDRXWAO4YIkyRix5RQSWxHCGUmBqFrnJiatFMdWIqEXyLepzBj5mk45kfUKF2pjV83MHSHVlPJ6Ygh8I78UE/qJtCbLe6viOund6JpzdQg+mN77OVm27uls9TcoALY0L2VDW9xFAmUHUK88Jc5dWABojIU0WynNiqVg4NKZYMk1h/RRAP2p1eUKSvqE96o/I92ZcP4Bxqno7SodLGRxpA7xkf3w3Lle6ksyVkxL3Fl1TeAChOpHeYp7hKnpvA4idQhyDhs4f+PC9DRWfz8cbI8+gFHTw0noKrHLwobj621FVpxhXBthUddIZELVagiJEC02Az68I9hMkHfXHw09yYG3QSY+H8rynwsToqxfmKy06jtKjpTU5L3nTG9cdSPy4WB9BJbKA5QOWvh3Nf7jk4+lskNH+0yGeKgrHTCPcTronyjbpJfh7+4NUne6JFOKkk6CcV4uFcp0wxzFmX+SUhQsxH2wtQjEASv8jm9tOFDPl791Oy1jEpkzqS6MWv0nwJZP3w2UtXR81Y7MCojOSfcehKPVBYRYW0NYEkv9vA7ifHiDX1PGXVHdsETGeLUacHgmge/hUMFNL0Z3FIRPkMR4VqlwCMXR6Y6iPly0ZETRaDcqMsM8PRmy7lNEFEHIdQ9g+/0EkOwGBzMKPBheEnUJ2rjyQ07vlX+WsxG7aVd+F5SdCrD9Z3h/+2VzfxrQZR626/9WwVumytUnICyCbowWrqKpvT6AJvsUwRsnPomT7XobqiLqZ4ruH69YX+11cpwYmmbqvXKuMRoZbgkC5Wb7eFtzKrR545TZ8tsunczvb3pIwbs395MRvPbm9aof6spy/Vy4dx0dKa1O+lRH7QD8DuQKV/yjNmSX35d1nlTmLzUaxVhGtX08qd1z242KFsbt0tqEbDoVS4836ts3HLawlO9xKIgiYnnkNaNZNb1aJLgDA8tb2bAX96L5ftxwg5v0Pbk48atWu+TyolOWn/ZSUs6EVjX9hCu9qDhvQC0a1tUBHClW46Ar55SSYnPSmuI02DLSitRGYh9UJUfEsYGkCUTSpUw0QJ07KsW/WXp0AAmG+6trFV8p9WKQNXY8obfra6EN2C0kDOu3sgCy5C7iyYMxGCuNg1gc6GaNfyPaTJ+UaurVnJmaVU7cUpnOFlNnu0uBq/dTMDfHLvSunmQy9c9voG/vx5nb2ETbxZBnUxZFgQmHHaXADys6+u7w66jfILRIXbR0HvL9hBPkjsiLiWcWhJ/36pEH2DmHpOvg/BOvWYmMx3dIUyTDxbL15sHRkr2i6o9LXm+jaY3aPRhDr+/WzDpX/LfGf3d2PPUWSyWnsWVTj7yXs11RzIFNym64xRze0OWVXF5sPEKPycfk2SRrn20/uNPPU5FbpZ5eNrowsqzj2+9ceqGOw9MY8MJ5a9cu3Lttja4eyBWpKts7gun3QiaUntkdLWMvWnMcWUa+VdEpszoiun02smsk7K99NoEsdrfABj8NQ8r0UPCN7/Z3oY9V3w4mrN3n1BI6dohfRkyZiWJnLfjVxkgOl5KjN61+C6dBLJ/dOMN7WBUrrLHQzVYRnW5vXR3s6jLqSn+To3G2jmUByDcpLmRziBYb6fSq7t3GA2/0eGOsKreLHK20u5C8ecF2VNvBVElZM+dxVaJBBPrnSNlzbg7i40b5ci4RT1JrBEZtNW9tUvxvW1H8jOTwH1nOUVs7ywtu+8k89rcNLAT5ayieUrSKt5yG6fCx8st1iX8hcCys0RRd5ak6sdoBMwp9lk/LienJZb/CAc/uQt4pcZvrDGk2bxdCfqlMLkbR1IFnDfPbW4lM3Vqc5PE930eKVnd3sCkjcoJbDKBYFivzgYBlo83qCau9Wxb72Hotdcdr1L7FpLPNNzSf4rnnT/e9EdqY9THm/ve7aJkdR/Z2v1j6e1kOlznpIxz/J6psyKq7f9z274SjPhz79fPvV//6/d+/atLGf/yMsbKmsLaExiTbvjqPMeKfWAOeZG0440pf/8X7Wb7gR1sy3bx37sFbUftRyqvLNu/3uSVyAhSmtqclWMajptetyg9o8pl9n9HNhKpIf5yi9ai1X9so9vOX+1Ek71vb2xA25G9XN/ZcLbzf9OWqJ1/yzaogRstoQJfS0uog2jRP7kXZEBXoCfZL0Gvm5ZFoeWWqRlfwpUaTBM1vrPX5fUoamPLOv/mNUBCzh/Yi7Lzz+w/2fl/Ys/Jv7jFZOfvbSv553aSrN/lkdha8h/d7vGfXE5fSoqojS76iNhScqQS5TuW1trhQU0D7vPcWXvUjacx1Lr10nmLxfEoum7/qbMoy3vJlgVTxe+s+j88tdVrDYLp/agdgxJtJ2sljsuIOlIvDJTdu8/RjltYcLFeK9vOZom6qSdqnGel8UXfJBo/rWkcrTEnqDhLJ7fJxxu4FbYS27gT8GokLBCynOyRrZ4/anYpXaxf1t3FRKQ803eWKawJIQVxSS8gidw1ezGXNwFy4isZnqmq8P2t4uxm+0f3dIM66kDVm1u7MeaP7d2ejqaNviY9s0v2Umpp+mrKzOBLE0TPb274fr3pWpD3as81S39oy/VU6v0l8tfuwHxjB+eiKeNDmdIbO9NJPr7IYbHD/fYV36MTXV8RJr2Ourf/q46N/E1H++e5j5/nPn6e+/h57uPfd+7j5/78n/vz/y/bn//3o6q/eGPAj8Rcf9HF//8R2S/mL7JBN/lGkp/7kX/uR/65H/nnfuSf+5F/7kf+uR/5537kn/uRf+5H/rkf+ed+5J/7kX/uR/65H/nnfuSf+5F/7kf+uR/5537kn/uRf+5H/rkf+ed+5J/7kX/uR/65H/nnfuSf+5F/7kf+v2s/8i/mL9Ga7y/bv1xdOZfH9eO9en2/6/tXh75fP6xV/Kv93uHz9Vlp08p9s8Iza9cqFiz3zOpZmx2rcGbNLOuDVTqzHMtqWiXL8ixraG3OrIJlfbOsjrVp4Z/v+xXfR5dzfxd9+5Wuf+D7Xb9a9498P/Rrvn/qV3b9Pd//7Fc8Fr74tTlrPvqHh/7I9/f3/dGufb57cNDb7U2rp0fv/Vb/6uLj5/fn/sDqXVePOv7488vX2dHs8Wu7Nz3+cFKcbl7uzs66ob8/r3r+bhd/Psz9++5p5fjqwq3W7MNS8bH98in3/lNrUHj53H1fnA9Pqr3uUWV33Lt4PDgaPX49vAiPP1S7s9Nv2bnfb+1Wu79/q7T9vWO/cVKZXe9/rbfq1fPD40r1+PrzxeHx+/Zxv9W7fv8+Hw5/v3jIHh164w8XE/toHE6zl22rcuzvnvsfC5Wrq/2ZTHf3YN/3dqfP748P3189TKvnw/edVr/90py+73vDl17LPTo6HO2+DN4f1R+/Di7Gx0duGE6rpfdnzbB6UL/sVvaP977Vr75Vxvb+od95qNhX72t+dze897z3Z17/rNrZPfIPv75Up6fH18XdYP5xrzLcPTi7B6lrj9fnz/f7+/vH193K/cU+iHV933sPWvhf78+9vePw6vN5vbV3dnzl+9XHvT3Qyvc7tWH3dK9eeF/rt85qXc/337dOz7stUGp0cNwNziovxYPBvPFcmfv+WT3Ig/ag49knv2DXamCyyuM1hqidVK53q2CPmf/1qtqcH04rQVgtdI8JVw0s6Pt79apT3x34Trf60T87AO/4g/mnY7/u+9P6wTf/vF4d109+9/MYwK999puH4LXaV//zVaXvfzj2O93qV/a3Z9f6c7DfQbfy7O87fr1bw3jPhGt3ftJjf+3u7jf8Vufd/Z4/PEY/1aLfAfehtd9jf7Wpf31Y6c8B16ldabC/Smez092fWJuHfm5e22XlTb9SsdDf+/rulPz8zSp98w/ZT3XXPx1VCKhValoTSgjk5ZCNKkU/nNm48Pxqt2Zbz3W/MCse+/KvcsK/c6t46J9B5DrFmn9WgFzVSlbxA38LFmCxZsWqlHdZG6DU/fjfmcih/LOsov9/5l/NsuZvPz0h/x1ff2rd977u2fXPk8pDcw/1P973KuS/3vnH+a4dfvxY2Xvafbn+HXg42tu9Av/VOrWqfXY8n3/Y7e3Wq3Xwn18/O/fBf6fHB47/uVV9P9q77F4fkf+mfkP4L6ifhb77WHvp1qeVhxGGqOYr1y3wX3XgD7vVmn/cr1T92m4diDvyajXS+SSsnIKS/nBee/RPT0m8s/lZg7i15geH/sd5teQfffN7x1Bjft4vkm8wagA+6+6W/HxYHc6PSY7atEt+Caut+u6xP3isHszBt6TV/P2A/X3u1gL/IqyM/fcTv3GF/irffOua9B/7+az1vv5+7B+Nqg+k61G3csi2J538gX/00erVq6d1P8v+9vyDDvFemtun1nG3Op55Df/3Juhfsf1aF/OtfSicdnd3Z4Wav99Ev2Tbkf8yAx4PDvcbln3sf5xYrbp/wn7ss1KD9HruOEX/qGmdz2unliOEvfJPZ+y3459MrIN5bWQ9H4LfwV9nl4oPT4dWtV5zrfnI3+wUz4SbNV9U57W8NW+xfM8/Ez7eJB/j98Wad/1SB/x/1rGOhI9bim8r137RKl74pxNrr1s5Yz+5WZFk8Yud4nv/DEaqXgstlzzhX2pWQ/9ZC+NivCv/+YzwTawr4b/i9f5zt73fevx0fv6Qo/7rnPemiv9OK3ud4kX3vNXbm1xT//2+95H6rzaved2T+X1tULW6xw7134f5ySfqv8PHfavbmld63fcf5oMqkO0Xug/Cf5f1y5Ffetx165cf/KIoj+fKqU3+G/oPu7Vq/azmj7q7D9RXp17tgvz3fl79VH8/8++9Wr5+OCVOD+bHF/wddfcO/aduZeR/yPrjw5qgthhSb2XxsPKxftj3w+vqcRf9HF5XX6gTYFw/dN8f+eD7dhd8h34K9fcj8t/l/PQalrnaqO8F/oByUQuhg6lPj6zL62ppVgj82cj/nXr55LryQjyeXu22Z4WBf3lpDf090X8W7S0pZjVG/rBQvJ/vz6yXXcpFyz8ZFgCfvXfccd36x6a1C+buoPruCKp6l6DnfvdPWtb1ceWoY31gPy+z0iPpfFU/KVgPfuXQKvX8Y/KzH5Af0G9g2dCfZ0XPP/1kvcwA1/Eny+JvreV7nWLgnzYt2A6lzLu+8Eu+Uzz1z2pWrQs9mp+Q/y60v/JI/tqf1/oWXJkz8l91rvTpySdrv1trWC9X5MP37KcPlUz4Duq1J+v50c+Rb0877Ecr2ppHvts8U/p7T5TfmUVHyD8tftp97h7v79WvryrdF/Lfxfn9Met0K/X83vHV1dn9+ee9o0PwX+Vlr+KR/z7WfvdPs133pTaxT38X+2ufXs7Jf8cHp/PORSVrH3yeB4/Ufx2/dcEJg1W/VqrdXfBhs3JwTWRfVI5FH7zADaseQpj9F7s2roP/DjFJ8k3tsPq5e4TnV9Xd7nGJfFypHz+Rb1rdD5v+db360P3QA2bZz4XfE34+8i+71by/78Jnqlao/45g14X/5tVhd+/I7x5XJ/574eeg/uF3pU/9b/59nfz82f/Ifir7/iP5D+WzK3/SPXjwz3YrH8knR7uVJ9Lj/aV1WD+YW9/qlYd5RWhUqvtiN/Pd2rlVOAR9NuFf0spWHhTlK79bfds/tkq/+x8uoTpZZsP+glxHh3AoSzX/9IMF/TQmXT3SjfpgVnyCvgXLQk+VssqvPSPfYFToO5KT9D2oK9Uo+gp6x+9C34GPoH/u2d+mtr/WGe31N6WXnrtKT2mGEXtdpz6s+y7hOnvQ9nuX/cos4Ad84m81tusV1Y5wXZKfs9r+RuX+4hf9VO679KuvD++74X51fl17vu/t73rX7fPe5f6J96l63p3sV8Krk/N6c29/9+pD5fxl7/jxalo52du1jy/69yf9vWbx97By3qxdHp4063ax9nn35NzPtmvA4J5vDWvPV2d+96VQ7R6eFP1PVqVd3HvvQxYvr/e780a2Ui/uDbsXm5UP9f1ed3BdcQ+r5/Wjij+AUpwfHfiPj1W3e3Ttf5vvvnQvnirV0W57Xnd86IfH+X7Bb11Vct3dz/7LqIb45cqfz2tntPO5w1qxXqn553ZlXN+/Jj/WuocTf/hYfZ7jF3x+Vj+p+rl6xTvb3Pfrx5V7f/+rf35YrXfff/SH19V+HXz/UKzt+dC3F77v+rWcX7Mrk+7+R799WCn5uy/+wKs+zo9cf4IwwIed86BvYO/gOiJ42YdL0qp49d253z2s7tdPr/zpIfiscIl4BKYYbvlFq/LR33v0m93KrL6H+baqYX03Z00foMc27/3aqNKd+/t+Z+7X63v3fuO4ctA9dPyHx6MLK/wKJ7H03t+3fagHr1vbBzNUcj5cktYIoc51vuJeVffO3NA/6cIPoktx5XfrtYH//tD/Woe2ObT3z7uXJWv8yepAhPzdon9ch1I9sP3mHNrj0KvkOta4cm/vOZb3wd+FCqvDxd0N/WdoPcs59Bt+7YH3xfnVAejgZy3r2nqZwy5bX63nIkC09q1n+iOVqjX3DmbdsxfLylr7fu2TNZ9DDlSwuFeHnfl27BfBo5UTyPgc/It/lTrlZ2hV5vQjDv3SrFj39zowBZ9s4JX+wyH9hEpX+RnQv5/94wfrU7d2Yk3n0FsHRfEnZvQPHqwa7Dfske9YxU24WlbbcoC3ZlHGAzxF1jstWLvz2rH1DfFnpzT3P4Q10PHU7xQIT479bVL/nzWt/XptaL1A78Op3yw9FEUe6YfPa0XCq/z1JuGfWs+78POqDdqHzefO573S9adZpf1xf//qU+4+2Ny/3G2cPQ/cvfzVpw/nD5W9F+9qdH7a3jttXb1Uji/2Ri04hsXL3Q+HF4Xz8+ZeM7ycnF/Vai375GP3GRGVfXLmT8EJrd8P749nu1fHx3l/9g2R0dFz99OoEjzuleqNceWiflCYNz77wXX1rHtcqzx6B3a3aVffH1f3oVz9Xlj96h+G/oz+xfmHykl997p7vll5H+4O6hdFy6lX7Pl+1m9fVT7VP9iUg5fuh0f40bWj+WnJzz7WcvMPAeKiSm++b/nN3crFHHIcetX8/IPnd6/pTx9AaVlXtMtXrUowPzjxg6vKHlwnxF/VGeXqfrfW7h7tWXOb9mHmX9iVs/rBB/rXle5hln5Kh/bn6ao2g32yLNufzMHqV3WYtv1vfvOxslk/KPkfMZ/6h45/8PB0PH//wXqxK80u+LbWrZzX95/8T1eV8/lR6AeHiP8+jqv2t8Jmd//Fyj762S7sGuTqsrs/9j9BSOpHCCm6e/NOO/D7D/RPHxGnwhWrXflQPefd/Rz103n3ouvbn6zAP/9mhUPral5zrELoN+uQ54tR5aQLp+djce+hO2r4wSxfq394sCwPfFSc+odFyHsNfP0Ik7v34P/uHVZnxYvK4afSR3+/Dj6veMy3FOp+26/Ba7uqgNeeYcT9F8vZ9K+GVrtb3bOyhK86pF+e69Jvb+2ezAewl9eVo1mp7x/XrN1uJfQPh1YAk2290EXc9/z67slut9q2vs0JzwXtHOSl4B8Nrc/1qsvkT1A/PN2D/cqfFU/8029WtVv1LPfQf+kUTv3jwuZ5dy9r3V/XsnMEGLCL1Trt28iH/7RrZbuwAsUD//Ol9dStPmyO5xXI3aP/wUI8ArhdG9bCy/ujT4Xz+mW7EFx9as3yfbpmnlWc+yffrN+7lScrewj/8WVcOT7LBpW9frbVrXybWUXKcX6GeAN+1hEcLQsq40MnN/avJrlP3Sbs8DW08/1k39+97t93s/tnj9fz+/vu3nPr0+C+s7f/4ZF2sLjve1ewYbm998eX7fOLzd3x/PLbfSu/Z9m/Wxh077D1++/Pl2ewe8dW3bmvhcWT3nyzUnOvDuvzycfqU/20M8/nayeHp1799/vKuLvXn8MOXnb37Pn5aRVO+ie/+63S3d0/8HvNSrdYvep+CBgfwqnY9fPXuwglAt85rF3Nz+sV+NuD+cUp4kP/0P9wDXmonnY/7PrN6+rTHF0912tj/6ziF65qUKqII+eVwfyg4nfCSra+2/Mfj6u9+YcS7Grt4/yk4EMPfq1X+v71biXo7l9S6UPP1fx+t/pY//DkP17XLrv/X9Pe1pwqE6z9X/YtF2M8RLjsgeGgoIyKEe8UZTybSMKYVH3//Xva9daufZVaiQuF6X5O3WZ98bzLEePfvA+eBA+Bdz2taubTH9QBecqf5MMxfAakYxxDEkn4CE3bDt43dOhs/amdVMAF1RNuAd1IHR20NIXFsOGW9pl0bdzS5e4v9PTLc40/q8WF4jsBx680uYOXQp9mhX81cU9KFe3NRyuyDH+Xc4aqjQVVT/vwMWFNi8y/4X6l06du/VXSsgZOU846Hh8F6sTSzMY/TCGZBR+lTeLW7VqI1NGcq+C5nDQka5DILtGXXBXxSidHxy/VX9utaHTzui/9Bp/MvKs6cm3VXl57TqVnK4GjmrfekzMsD59SgIKnjPuzs7zZ44zGK29twy/xdgd1OQPxfiYDvhcD4JzwQqqge3JnROMh+xDwVYL7ooMYKuoKx4qBIdeM3qB30O/ujf1zoINKOLkAH29F7+H4Ju04TQcyrDfgPMlpBfjnTo/WPdCi9Vo9+nW6NjEtaHbnwhoER/EsOH9yRMeSlw+39B14GW0P7jCT76R/BPoY+veN+ZiYH9/vcA39AW22IN2PmsxjeNUKz+X2Tw9z8GGAF31Nt9oraF57N+vvaZyK6bP+Dk/nVVfWk9CjlX3uZOS7ZSpNAd1ZrmaHeaTOi+gwWSlbFFLOJ+H4XKxnea1EsVg8Z7swvBdmNpdBrLLSWCfQNHmY5ygY93HCvSb4KCdb0/OCLU1OdtWRSwOTuoZ5yKB79kdZlPi5ucpvih56e5ef5Jd6hLZ3Oay7sq/51JMTdGKw09MpdKMK7Qw40uBSyYxaK5c2HXF+1Gj0IRRbbPKYhjZYa9wVeG9o4jnMrRxb2MWLBV+NW2oq/2GmY0IrOCZY8esaG5cEy+ZY9QXfg/4be/SVBQc9/SOZgRejmlalrE0ccv91rZpwMd9onNEgC9ZGfgFhwY/hipb4XBSdaKPkm1Yzuif+w06PBN2SsjWTfZir2CWVydyGbwRp0NVqSW0/eKOM3R396eBI04oE4ac0Mqdwyfhwt+oCHAo+cg+c7MpCB3v4Tkj0AH4auGr/6dmIZncZLcSV6xJHfzR+ym95JrxuCpOmo5bqJKjNzIXf8roM7omhhcZ/jRrWST807QAf8PltX5W5t6VlLLSFrxsozodqCu9UGkiVMMFzSy/Uweeo3zvs/1MjS9Zbx9ZxmNeuaBVSADeTa+7Ha+0uuG9SptDJTSy170IX4BFJh4wKoNf24lmib0AD062AD/zh/KvD/i2+QY5M99Q+uO9+mE88Jhn0B1R2xLnYFwULca/4+b98HrwudOAKr5dDMazokvcFberh2CaNOG0d0bqK/aJ45QQxdANaGye/hObkD6EIuE2vnBf+L+NfZTSGfzV0BD7IFORO19w5vPII9on//cyZB2/wyf5D9DTVNfQFuTQ6mG4k++VYsv87l/p52LH/+zqcDuGbxe+zu3reF6NnNlAdvejL8YdqmvlNji9KJHOPmm0w7aeC2l/fSUBaf1Uwsllm3979QZZVdi5l2Ki9mcdynKipXTzpadSZ8l8pNYocdfJ1ln0DHK4Sacx4TbZUwhR/dLNqTkrRd0FXHTbwb8ox+kF1JZ82uJMD/UbhjKaldClYUvkQI4v+uN39vUXfbq3vcCjvxN6G1BN9Ab2J61UlcCyOaHeWHzbGYZboH3qjais+LXzd4iwrq0aEcxsASyCF5EtPgrp8HV4pAk5a9c5FCh/0zbkqeHFP6xx6BOe+xrkYnLdvpeb8dhyIrYFOTv9EaqF7hnf6NT7UOHBYiCnRQpxzkQj4pmELCJ3mLI1GrDvh359cV/D1Eeea/uvf+H81IDl75UvqX64PvrKvOh0x7k7YRz0NvUKHVx1wvrqCPgta1tEX4VWsn/43OtDsUyrOpzL2KRC7Ltf9/5kbyP9yjJefGXKOyz7qcMrDP7dUTxNHfrUOD4dt5GXl5LD55TzVk8mb8op5PANzZ2r6Ybzv4FPNNrKqgmFngcZe+4Mm65uf0v9N8tnT/w6e1dS1H1c1s2mhFwCzKjjQHKZXBxudDujSBJ5dz6m8K2lHIc0a/89CJGwKP7d6K8O7ik0ayp8mOMK1yRB49TKzBgQT/NKngg6ZrSl1OaT06QcSSquK9q5/0PAly8bHMZ6pq3xpVUE1Rw84ylVGSx1ynipjnmNR5Q8NpOi+hFkGPkJET42Cjm/gS8Zf/Pdd6+ToYOhvOSWmBu3/on4hPUYAuUYOcrelXer1GJ/6GXWIjoBAGVuZ4DoEvA/FplLDFuSp+pxbV/SR0cNEqXC41cHj89TZGPjKIXQ8mlo8cascyZ8D+Gn/m/W45Dz8r6C2djvcBTfhPsS1gU+GlYAVgi9OKItFwfnUeywm4BchViK0fsOQkvD1gbdOPRzxHGtfD+7sb73cPbFP/mnhC3DVE6mlOGTUGP9VeqiXd/b5PNPhOdhVeGtS8Aut+8v1WByO/YiNoDzsozhbjZ7HVRjcP8zhJwh/zuVeZlv1dl64cjqG6VzZ2eY7KqtyaZ5NFPfnQ/oeBnU2PtN3HUTn0ROiiRMLO8t/ILImrZlXNOiEZw0cyigMTf0uzyoa6KSSnlFPG45pAHyh9EnfCqQYvtND+X9alwSz5lBUkiz9ncXvl3ccXrLkXMoz4YK6HSjJyIXdlGP4HKosGe77Q+ZXlMEnFn5rIiGau3w3KOEokTCJP/BVvmZ+F5ncte6PHMUcjQBfGupalN7CyhpmjPXw1QQXyCyxt7CAgLyO5tIk+eC5+Rn+O/f2opPwyAqIfxIh9BvXGXQf+hYWrna3wmm9v9o7iN+h0PCxomepJlrS2ZA2smYf38txfkGF+oJOBT+s4RMh2tDn0KHgO0jK4sVvtahpLLxMZ434qODjoOu6jItyzHOeVPu5qBIUS+3JsucMhfukrBXd2l1xrhOTVDjSN9KD3Psi+cp1nJ5Y8CYAqOCvHvZE0ImP+VrL76EDXPsTT0tvuRPCr1Kbez1aLFyltSMK8P6zMuHArrqHvQ0t6mJ2vEbjqpwfdttoWq52s1Go7nbuzMBXvfv8IaOdWvRnf2hEdSoXObVn/7NIL+an9c9J9tBdFezURNA5DLZlutbzVOYqUPQBstEqsrOjTBO1NKu1pE640uk7HSroxDSiZekHOv+iH/g2O9rS8O53TTyjOpM9q1JqNOeU8F2G62hHN9d/6GhHm0p+6bAlXfk9iq+kO/6bjSQdSj81kUtLBR0HvlmW8Bfw5ZCurkkBjIU/1nGPcz2lQ0n7hApSW1pVwA/ouGsjh5wzpJm8mwBWsUMtjpz2BUp7FMM8QrfJHa1jAZ8TU1TQBbqEMgVJCT+4e0Wmc8YZsnLDuOHWnkAdQ3mohvmw03qfFG45J5v8y+VYGsG9cN32X6PGd9GpWBd90SQWM+Co6JfUyd+vNK7Fe+0lNI4FXtdDHdKa9cwsEMbElTjhvhh/umfq1U5KaiUcPDqxbmDmuwda/+GtuL5i4J6fip6iQnhwEVunq+OlOJ2Z1zrMd+DFFdQ/+FeC7/tU5cC96ULIJxs3twxm5hBJtxw/T73w15T5wThRVK0vs52N/LKYHPQmDKsik7oX+ucinBWncFwsns98r2b35fAw+WGfsddvdVA2k5vuLoLSZnfzJoNveGMZFMFMTcembOShimZUduXRhR3d3uTFRL92eZJFBvDYvMmy4Q2UK31W/rdOAaaJn9pxTkeFE8iWcuaG2i67vIfwA/zhHPrbRPCZ96Cx+ZrcfrCyE58s6+q4Yr1y0NBF6z4QeNSlC/y/SWfAqcDoqU+/pZxbQIjuyJmJP2nTyMSOgIMFcHD8Td+a/cUf+wHXBhfKXdrqeEHwi/D/J7qU/lpDAn5mwEO5evGSDfqkEvmhOf92pbTJF53uvtKjHV23fG6fNLrT6b88fWlDhzYl/Ec8Iu0GgUlBOrE30nEB3Y7+CG4kE/iPiPUV8DG9UdqXQT4I6L51DhauL9bsP14QE7G+m1XwU/roLzLZ1F24P5ceBMr3z/SpOaczkASRh+cTRXp8Ep8VQ7jLOd4EB0ZxwvV/JJAxcOpN3srgk6DznYeoDe2ZGt5Qwlx3O41uKkAVtB3LCvUv3ieUbjnf1uLZ0DuvQf3xBlR4FKKIUps+uS86nGNDr/mv3Dljp5yIP/Y1Xl8uKgl/bcQzY58TsU+Bfvv6L9eq+FFnZutQLTj3dl9znpeOK/jvJWU3saJgJ25nv3jNg27/cvR/r3NYRwKfK55zdts+PnTtiNodv3J0+ned1/7JRIgEzS++i2BMKuC7+DfH/OPrdTlX6LXuA0ciYO4q5/ZwRq/3eeXe8pWDa567lzyfn3LfP1gPQc96s3oU2rL8ltWd/f72WY1Dryk9dHhUlh+ZvDXhLxV9mQ8UMO1b5j3l2uXiWZ7UwZ0nB3UKf+18JvUpWLsT+I40oGTa2k7kCzPawYureTm/Gjgxz07GdvvFc6ebKaHR+tFN730/ydSHWZxx2MkfTbtyo/ydHlfs53+gI+nRlxMTHui9Cq7m40K9LJqblUd+g3rl/FvLT4rWdHaVp8c1z6Eak9fos+BsUP8ZybuOPkmX/sqk0JMdv2uTb3KVP9GzkN61LDjvge/YayilWeZvmD9upf9lx0P6dP1vm/TFsM8rTg9CXSzZ5+xwD3rU0Pnu33k/oNV+34RzjkYt8JdJeUHoJ0iijlVXMhVf54NOi6Ewck0jlzeNHPYpS4pyvt67UQ/Gh7NJAb4FdB78yXgoHM191ZcHG85pkfhWR7huR15sunAa8of5QPBo/5eCPeffhsI3WtzlxBZrfxwPvsxDOCfN06cRRxV7Q3Oety2tGtJHEl5sNZeTuDekxGd8GebuG68Lrix0zvZMOzP+FVUWXdvOj4xujqRoJzxFndq9cl5gbLCgLAG5zd6lf5bzdnCgbCHGBF7qQu3WzoF5BKK8B8WvCnP8Fl6fzkAiynroI0iigJvY33Pu9QnfQE1fQWodeL712+J90hb6ylfsk+DjBuJ5E2OzugTta/7Tcj/x8IHnXqfWUTRNnR92T04D/k9imgoR/DdfIuMPXrkft9ay54zI/xPbP+AG/B/6DbgRiD+XW3vLQ2q89btzSLLfuv9NuiA3B5+hb1Pyf8SgpHU9uNBPz/s2m3AYF/LJe5e8b8C4gfdVGroPeJOzf0gXXmhh3uALx/IA/V2U9ewQRUqX7eHwFc2Y/07PsOuu3w978KIufuVsz/n3Vs4hqPoLR6p5uC6Lm1yG6lMtF1JvUPTZSvcfQXufLKl7CPqdWaq7OJlz1jkEA4io6VOvNvLWCcd6+yvXOtpo2KhBGXm6uPnweUofpTx1fDzsPs9TP2z2RW8FTNVqRLZBc8QLkJnfpdkN5yUDnZ6B8/Bh4LMzbI6ennhH8wb+gAgKPnR85Pzrxvl3peTCgCRaKztmPEJfB6VB39gceB5mPP9Z85wKPHmzqoZ+wvuMLV0BljTqQGfxKsWJ6oY+TGhod2c9t+D8+6JxvgOI4Rbns3HJs+Cr/AydpX6hy2RIox5dlf9mxj3wJX1wvj7p0N4GO8qJpjoKqNKyMsmF93zmBvxwTqFX4hb4D3ENMT7B3fO+2dLKN95b9Bq/ybsPGgy9WscbMQAFGTxl9M+E4i3nidCREzkhVbWbgH5XcGLilWsb7UNyG56P9em2dWP9cRGjJuqL3kJ895xOLZ48zzxDf8HvSUVyLb1zuLcj5RSlv6ihLGnoDFtRMl8dNaghT2STdxpwlvPGw6AGxc/9uY+5Pz643msjoVPPuI9LS3Ht3KwP/wFfLXgEm1CJLmJ99qf9TPy2osd6M+P5VVCLTkkjhsp3or9cLEUv4efh+XDNvzn4h/Mt9iO4ywx8LXowDTxHr43v41K8lwO+mzAPwnTcoC+cL+hlMWiBb1nP6ZJ6zWetJtGKX/GH++c51vHk1nbxEEXP29qVdVreI+N9sht0L94P53Pnfbneafip0z/RHXpXo177FiLnPC8XkZVP4d6p4lW9/dB5kMR94XyflYLeXP0895fQs6ufGQ9nq1Vf1r3Q7awS3s8IzSI7ZLkSycKd5U/11xTXg65CXxWunHRDuhf6MDdB2sn25jkL9iqref8iOnMO5webPpTPX4linzzs6hNFDrOwmchzE3VofZRuFfX1+iJPfTTL9iChN1FfN3IbP6JJSrYMhq8+sYE08yNdC3WnaQCTKEud3Mm4fszzn2MTlDodEg61sNme3jrBg0KftiQ9ioH3Be8JdXnvbG3HT/bRa52/815PpaHDtqzXoGf3dxmYZE0nF3w5Xr0yU5tVVHY4f/EI1nDPfbDNuP/GdFH+2uD9ReUf2N/AZMxh9WmZyRXvc6yt7PG+EPyWY9IvatHEtVfSzlJuIbnBpzMDX74q5ZNUh0wWoEhQh+gPTZ/4/Oi++ETSyhmBr3Tmw19dZXL2lVE70Vbyr8Y5l+AjJhHwH1rO0qIAT8K39VLnrsNauA30KXxuAn3MeWXWB7/Bry3LMNBjUA/n1X7FPhh9oiiy0KNBSrh16PwjBVlQ584f81oBvcs+Hfh+YSu2tmBWP8F9Thd05FxZTDgPUMwrLpoKFpz3GM5a8t6HP9XpD0fDEetGFHuiwYrw/z81+gqvaykY0t2Fh+ARYo95acz+LSJ/I9469Mb92h+KOeU7qnusPwfMi6/57YTn3L7kfTpQ4Zrzo7vmfvhvzwmv6ws34L2rykoIRdainDeuIJJ5niY4N3J4vyLi3Mo5iU7Ovq7nLU388X/3h1/7TtNcQG9uxZvL+cEXbXLnp/5v7e6lQxW9dO0UPG19Id5KSmvvi/OnsTRJ5FMZPQ8+70FNDyaNsvOsOUR7VbmzcqZqdXdnU5nsQErjqb0f/K9m7HDRPdzR2YyXvH/QQGrSo4O6yD/APxDdfz2qC+Bi8uWXqXvUH0Z8ldFdNCA4JWe5B51S0LdVsXgYgOCgT/LMD/uXypuYWrqTzIBDzp9w+nRhn2Fd6Hr8FH84N9myr8ARdhnXfJ43ZKzboy/W3b4OHvx9DAeHT7C8U8v7M9t/eS+er9e65jWvfx6KaNQs74euGx7v0/5TPMLdWX8/0z70NgydmCTHJN7MtOMPznFOt3FAxUS1v71gmCUw/z9ybHz4wD7ti2TBecfZ+GfhTbiPS7vgYoiudpKKrYmvJvwULvSdAafCrL0beROPnpNQdOF89WQSh3Z/jqbphuc2v2L4Q/VjWNvwTkntXTnS73HfygstILbyfk9mrQff8RRrl5b1u8O8uxI/W3q2njGT1Om6apR7D9HnvdH+iC5bt7Szhpf0fuQ+ia7njyV8hzpm2/Bw2EQT4C09m9A7594874MUJr9ycoKPzEbUzP1ZIm/k/fnnznhG+Z0+Gvmu81KWqZjZjwV9E/o12dOGh3HBQv6c3mNbdDiDL3Lng7+voI2/9J+B0zGjtXfv4HO5NY2BF/X1LP2bOPP3Z/o570M5FKvg1g6gH2Mv53wYksTNq0h+L7z33LlTlsFHDq2g+8fJOS4lrGxD6kvE9/BpVTX81vokvkuaLry+yUcOm1cTWhH2af3cz8KjKv3naR19mNX3rD6EntHhIf0KfpvZr2026gJ/NQuVvzXJj7mc/W453dJx6//QKKSpobb0b3rk0E0DjMdv9KvlxUx/aAde1ElKmzNATs3pMYQoTmPWbTudNqKrgI+MW8BFE6wJ9Nu3Ee63FQvew/Vur28S8WrlBgjM+d9b7mU8Vzu3gwmNbsAjqOW3Ur6Jt5omrQO/tec9slJLXww07w/n6Hsn18GTv6eET3vm/jnmXsN7If/z//4/n3eR8Q==","unpack_command":"python3 -I -c 'import json,zlib,base64,pathlib,sys; r=json.load(open(sys.argv[1])); p=pathlib.Path(sys.argv[2]); p.mkdir(parents=True,exist_ok=True); b=json.loads(zlib.decompress(base64.b64decode(r[\"rerun\"][\"bundle_zlib_base64\"]))); [(p/k).write_bytes(base64.b64decode(v) if k.endswith(\".bin\") else v.encode()) for k,v in b.items()]' artifacts/goal/batches/e02-classifier/items/all-observables/result.json test/scratch/repro","commands_cwd":"test/scratch/repro","commands":["python3 -I driver.py collect","python3 -I driver.py train","python3 -I validate.py","python3 -I driver.py eval pilot 64000 64128 pilot_spec.json","python3 -I calibrate.py","for each [lo,hi) chunk of 128 in [64000,65024): python3 -I driver.py eval choose lo hi choose_spec.json","python3 -I select.py","for each [lo,hi) chunk of 128 in [10000,12048): python3 -I driver.py eval report lo hi report_spec.json","python3 -I driver.py summary report"],"P_recipe":"Fetch pinned sources listed in hashes to scratch; python3 -I gates.py before changes. Gate tables are upstream source data, not part of the compact inference bundle. Offline nominee inference/evaluation needs only bundle assets.","commands_max_seconds":470},"runtime":{"wall_seconds_since_first_source_fetch":4797.367156505585,"scoring_worker_seconds_sum":{"pilot":971.7639688239433,"choose":6373.054802291095,"report":7951.972878661938},"parallel_workers":4,"note":"Worker time sums are loaded-machine times, not gas; wall includes verification and analysis.","within_one_hour_wall_budget":false},"limitations":["Replica outcomes do not establish engine or leaderboard gains. No contract, gas, deployment-size or fee-equivalence measurement.","No broad hyperparameter search; one fixed teacher and its compact distillation, with choose-only scalar calibration.","Compact tables use float64 cutpoints and stdlib floating point; no Solidity fixed-point equivalence claim.","Distribution-shift diagnostics use a deterministic 1/16 first-own-callback sample; all callbacks influence policy and complete 10,000-step scores.","Total measured wall time exceeded the brief one-hour budget; full report coverage and actual runtime are disclosed, without a runtime-compliance claim."],"source_quotes_refetched_and_matched":16}