Commit graph

  • 788115970c fix(vote): unranked members voted with the stock 25/50/75/100 ladder main AnimateDread 2026-08-23 08:55:04 -04:00
  • 042f20bdb9 fix(barriers): cap the horizon at what the close-all actually grants AnimateDread 2026-08-23 08:32:13 -04:00
  • de382bbf88 diag(barriers): the horizon the geometry is sized for does not exist AnimateDread 2026-08-22 16:24:14 -04:00
  • b5e22a1e34 fix(geometry): a free zero made "never resolve" the winning geometry AnimateDread 2026-08-22 11:30:54 -04:00
  • 05f1a539c4 feat(geometry): measure per-candidate barriers against the one global pair AnimateDread 2026-08-22 08:29:59 -04:00
  • e372ce60a9 fix(vote): "0 fired" on USDJPY meant the threshold is above the highest vote the ensemble can cast AnimateDread 2026-08-22 08:16:11 -04:00
  • 04d70fb748 refactor(dry): one writer for the era-resume context, one for an alt-data row AnimateDread 2026-08-22 00:36:36 -04:00
  • b91c7b1f7a refactor(comments): box headers to stdlib length AnimateDread 2026-08-22 00:30:14 -04:00
  • 5efdb48de4 refactor(comments): stdlib comment style across the remaining in-scope files AnimateDread 2026-08-22 00:25:52 -04:00
  • 0b06f8e84b refactor(comments): ExpertSignalAIBase to stdlib comment style AnimateDread 2026-08-22 00:24:45 -04:00
  • ee4d459c6a fix(excursion): the sample gate asked for more windows than the configuration can contain AnimateDread 2026-08-21 23:06:49 -04:00
  • 15b75d0a73 fix(panel): VerboseMode dropped the era progression it was supposed to add to AnimateDread 2026-08-21 22:35:15 -04:00
  • a71d6d1d0b feat(chart): signal marks in full-brightness lime and red AnimateDread 2026-08-21 21:57:21 -04:00
  • 026078d635 fix(reset): a fresh model quoted the deleted model's timeout share AnimateDread 2026-08-21 21:31:42 -04:00
  • f8ac10c808 fix(replay): the exit replay held trades through the Friday flat that the label and the live EA both close AnimateDread 2026-08-21 21:08:26 -04:00
  • 20b7d9967e chore(test-run): defaults for the horizon-break-even measurement, and give the replay line a cadence AnimateDread 2026-08-21 20:13:03 -04:00
  • 4070c5c9bd feat(breakeven): the break-even every layer scores against prices a trade that always resolves AnimateDread 2026-08-21 19:26:25 -04:00
  • 3d81fed5e5 fix(baselines): the budget said 45 s and the log said 548.8 s, because a boundary check cannot stop a phase in flight AnimateDread 2026-08-21 18:12:14 -04:00
  • a2a207484a fix(baselines): the diagnostic held the EA thread long enough for MetaTrader to kill the process AnimateDread 2026-08-21 15:22:42 -04:00
  • 2de62241a5 fix(init): the OnInit failure the operator reads did not name the reason it failed AnimateDread 2026-08-21 15:12:23 -04:00
  • 81ad276859 docs(altdata): the as-of comparison is naive-vs-broker-time, and the publication buffer is what makes that safe AnimateDread 2026-08-21 15:01:50 -04:00
  • 9332a33e15 fix(telemetry): FEATURE HEALTH said "f30", which is a puzzle rather than an answer AnimateDread 2026-08-21 14:42:02 -04:00
  • 2c19cf8408 fix(calibration): the field reported the raw argmax, which is not what trades AnimateDread 2026-08-21 12:44:03 -04:00
  • 68dce35295 fix(imbalance): the class correction skipped the abstain class on a premise era 1 just falsified AnimateDread 2026-08-21 12:09:07 -04:00
  • 0a046db530 feat(calibration): fit the operating point on the label rate instead of on edge AnimateDread 2026-08-21 10:20:31 -04:00
  • 667f2bcb6b revert(labels): drop the one-sided exit target; measure the calibration drift instead AnimateDread 2026-08-21 10:14:26 -04:00
  • 68ef19797b fix(telemetry): one ensemble member had never printed a single era line, in any run on record AnimateDread 2026-08-20 14:48:56 -04:00
  • d201d22206 perf(baselines): the Alglib diagnostic fit the same matrix once per ensemble member AnimateDread 2026-08-20 14:04:06 -04:00
  • a86379621c feat(labels): on a one-sided book the blocked side's class is retargeted from an entry it can never take to the EXIT of the one it holds AnimateDread 2026-08-20 14:02:50 -04:00
  • c780fd3e5f fix(lstm): the CPU LSTM leaked four buffers per forward pass, and the input-gradient loop checked nothing AnimateDread 2026-08-20 13:54:08 -04:00
  • 2ba0f348c0 feat(ui): thresholds pick from a dropdown, and the finder arrows are back beside the level lines AnimateDread 2026-08-20 13:50:33 -04:00
  • 750070c2a3 fix(indicators): a dead handle and a priming one both read -1, so the repair report proved nothing AnimateDread 2026-08-20 13:38:02 -04:00
  • b563fca7ab refactor(comments): another 25 paragraph blocks in the AI base cut to stdlib style AnimateDread 2026-08-20 09:54:23 -04:00
  • 21263cfd01 refactor(comments): Inputs.mqh and the AI base header move to stdlib comment style AnimateDread 2026-08-20 09:49:33 -04:00
  • ef6d62c604 fix(ensemble): the era barrier read healthy startup work as a dead member AnimateDread 2026-08-20 09:35:25 -04:00
  • 45af5b9c1e refactor(training): delete the minority-replay machinery that stopped running in July AnimateDread 2026-08-20 09:30:51 -04:00
  • 506626b381 feat(panel): commands reach signals down the filter tree, not through a registry AnimateDread 2026-08-20 09:05:50 -04:00
  • 8c0186c850 refactor(signals): AI signal files are identity + topology, nothing else AnimateDread 2026-08-20 08:57:54 -04:00
  • c171eace78 feat(schedule): close-all defaults to market close minus 5 minutes AnimateDread 2026-08-20 08:39:26 -04:00
  • 4346dd3c24 refactor(stdlib): the vote thresholds are ints on the library's scale, not "confidence %" AnimateDread 2026-08-20 08:39:01 -04:00
  • 97f0631e28 feat(baselines): combining-weight fit, MLP cross-validation, all-lags correlation AnimateDread 2026-08-20 07:44:36 -04:00
  • 7584f119e2 feat(baselines): ALGLIB's MLP as a third baseline, not a replacement AnimateDread 2026-08-20 07:19:53 -04:00
  • 19e3595e20 fix(build): g_eta - the learning rate global no longer shadows a stdlib local AnimateDread 2026-08-20 07:00:09 -04:00
  • 561c88e6d4 feat(baselines): geometry-drift check on the derived stop AnimateDread 2026-08-20 00:34:55 -04:00
  • 90c6e26e94 feat(rng): ALGLIB's L'Ecuyer generator replaces MathRand, and a seed collision goes with it AnimateDread 2026-08-20 00:29:12 -04:00
  • 36254dcd55 feat(baselines): matrix redundancy and a declustered regime test AnimateDread 2026-08-20 00:01:34 -04:00
  • d1f08eda13 refactor(dry): one retry-and-report for OnInit's five init steps AnimateDread 2026-08-19 23:53:12 -04:00
  • 5cf706dfd6 refactor(kiss): lift the model fingerprint out of InitNeuralNetwork AnimateDread 2026-08-19 23:50:30 -04:00
  • 786aa76083 refactor(dry): one shrinkage estimator for classic ladders and AI tiers AnimateDread 2026-08-19 23:47:06 -04:00
  • ce2c324527 feat(baselines): Alglib forest + linear on the NN's own matrix AnimateDread 2026-08-19 23:45:03 -04:00
  • 38a12a240b refactor(kiss): drop the AI sub-vote early-exit route; certified == traded AnimateDread 2026-08-19 20:27:28 -04:00
  • 77594ef5fb refactor(stdlib): one quantile definition, from Math\Stat AnimateDread 2026-08-19 20:16:03 -04:00
  • 29c82ad50b refactor(dry): one binomial arithmetic for every "is this edge real" test AnimateDread 2026-08-19 19:50:24 -04:00
  • 0826f8b900 refactor(dry): one definition of "usable quote" in the pre-trade checks AnimateDread 2026-08-19 19:33:19 -04:00
  • ea2552efe2 refactor(stdlib): adopt Math\Stat for the deploy gate's normal tail; retire the b1/b2/lr/momentum macros AnimateDread 2026-08-19 19:31:36 -04:00
  • b0fec27779 merge: built-in indicator migration + YAGNI accessor pass AnimateDread 2026-08-19 19:20:00 -04:00
  • 61c0d19ca9 feat(indicators): run the built-in iMA and MetaTrader's ZigZag; add a classic-vote shift AnimateDread 2026-08-19 19:09:58 -04:00
  • 552edb5fb1 refactor(yagni): drop 13 accessors nothing called; unify the ATR trailing pair AnimateDread 2026-08-19 18:55:36 -04:00
  • 2be2970434 fix(topology): the capacity budget counted overlapping bars as independent examples AnimateDread 2026-08-19 18:43:48 -04:00
  • caad156464 fix(geometry): the ratio raise was reported against reachability, not bounded by it AnimateDread 2026-08-19 17:04:32 -04:00
  • 6f2def0605 fix(geometry): a weights reset could never change the barrier - the pair laundered itself through the wipe AnimateDread 2026-08-19 16:31:18 -04:00
  • c3daded397 feat(geometry): 1:2 becomes a FLOOR the swing legs may raise, and the scan can no longer undercut it AnimateDread 2026-08-19 16:16:33 -04:00
  • 888f32d21c fix(gate): the plateau shortcut re-ran the deploy test every era, raising its own bar AnimateDread 2026-08-19 15:40:52 -04:00
  • 9a0d063da4 fix(inputs): the Neural Networks group header was singular AnimateDread 2026-08-19 14:23:37 -04:00
  • 5e0317f09d feat(chart): signal marks become price LEVELS at the trigger, not arrows beside the candle AnimateDread 2026-08-19 14:11:49 -04:00
  • 70eaa8a498 feat(direction): Intelligent becomes the shipped default for Trade direction AnimateDread 2026-08-19 14:02:56 -04:00
  • 6717509c8b fix(build): four compile faults - one was a SILENT enum collision that inverted the direction policy AnimateDread 2026-08-19 13:16:45 -04:00
  • f64e0f8b67 feat(ensemble): per-NN inputs replace the preset selector - the meta head becomes the vote's gate AnimateDread 2026-08-19 13:01:02 -04:00
  • 1445f175ce feat(consistency): the five review flaws fixed - training wears the live constraints, the gate wears the policy AnimateDread 2026-08-19 12:14:01 -04:00
  • b63e39f026 refactor(time): broker time throughout - and the GMT DB basis was already a live bug AnimateDread 2026-08-19 11:37:44 -04:00
  • 1a46dfdad9 feat(sessions): market-hours entry gate + "Market close" close-all option, both live from the symbol's session table AnimateDread 2026-08-19 11:34:27 -04:00
  • b43b676239 feat(fingerprint): an active close-all schedule keys the model identity AnimateDread 2026-08-19 11:23:54 -04:00
  • 3e467f92e9 feat(labels): the scheduled close-all is now a vertical barrier in the label walk AnimateDread 2026-08-19 11:17:14 -04:00
  • 348492fb3b feat(direction): INTELLIGENT trade direction - the measured drift picks the side(s) AnimateDread 2026-08-19 11:10:05 -04:00
  • cf9a9395ee fix(hud): the headline word is the DECISION, not the lean AnimateDread 2026-08-19 10:24:21 -04:00
  • b55880c57d fix(hud): sign-mismatch warning in the display-forward throttle AnimateDread 2026-08-19 10:07:12 -04:00
  • f30e342a1d feat(logs): throttle the settled per-era diagnostics - measured 22MB/9.5h of confirmed-working systems AnimateDread 2026-08-19 09:36:15 -04:00
  • 83ae56cc9e feat(hud): per-member neuron lines + a vote label that moves as the nets learn AnimateDread 2026-08-19 08:48:34 -04:00
  • c85bd9f516 feat(inputs): 5-point vote-threshold steps below 50 - the consensus rungs sit between the old 10s AnimateDread 2026-08-19 08:25:49 -04:00
  • 1a900a0a35 feat(vote): CONSENSUS arithmetic - agreement is now what the threshold dials AnimateDread 2026-08-19 06:58:03 -04:00
  • c393497fd6 fix(chart): display now reads era-end SNAPSHOTS - the live cache is wiped mid-era AnimateDread 2026-08-18 22:03:20 -04:00
  • 659638e072 fix(chart): the prospective vote was four models' opinion of ONE frozen bar AnimateDread 2026-08-18 21:47:51 -04:00
  • 4f2d81a52e fix(chart): sweep hammered the news filter; peak was a fossil; neutrals invisible AnimateDread 2026-08-18 21:18:55 -04:00
  • b05b4f21d7 fix(chart): the vote readout was repainted once per bar, not once per timer tick AnimateDread 2026-08-18 19:56:06 -04:00
  • a042cb4359 feat(chart): show the PROSPECTIVE vote while the models are still training AnimateDread 2026-08-18 18:29:27 -04:00
  • 8ab8cf6b92 feat(chart): on-chart vote readout, and Min_Vote_Open 50 -> 40 AnimateDread 2026-08-18 18:22:45 -04:00
  • 155f56e0c0 fix(chart): the overlay sweep compared a series index against a bar count AnimateDread 2026-08-18 18:17:25 -04:00
  • 129a0d448d fix(vote): a leaf filter was dividing its own module weight back out AnimateDread 2026-08-18 18:02:35 -04:00
  • e59dc1629f fix(deinit): vote arrows survived the cheap sweep, and 5 long loops ignored the stop AnimateDread 2026-08-18 17:13:50 -04:00
  • 07aa01777c feat(chart): reconstruct the filtered view behind the handover point AnimateDread 2026-08-18 16:05:13 -04:00
  • b28c81eb78 feat(rank): AI models rank their own confidence tiers from held-out outcomes AnimateDread 2026-08-18 16:00:32 -04:00
  • 4858507146 feat(vote): thresholds become confidence percentages, on ONE scale everywhere AnimateDread 2026-08-18 15:52:08 -04:00
  • 282b535037 feat(chart): filtered view - one arrow per trade the bot would actually take AnimateDread 2026-08-18 15:41:28 -04:00
  • 2c443ba3ad fix(gate): the ensemble gate certified a vote the EA never casts AnimateDread 2026-08-18 15:23:33 -04:00
  • 525e92ecb2 fix(plateau): the IS-error early stop was inert for every ensemble member AnimateDread 2026-08-18 13:09:26 -04:00
  • 65a3e4e877 fix(chart): a purge that reports "zero leftovers" was only ever checking its own list AnimateDread 2026-08-17 22:07:09 -04:00
  • 64b77e4bbe fix(diag): the cache-invalidation stall message could never name the cause AnimateDread 2026-08-17 21:57:12 -04:00
  • f102a695d5 fix(geometry): the ensemble was training on TWO DIFFERENT TARGETS - propagate the adopted barrier AnimateDread 2026-08-17 19:35:45 -04:00
  • cb30360c18 fix(handles): the MA handle was SHARED, and a rejected sweep freed it for everyone else AnimateDread 2026-08-17 18:16:28 -04:00
  • 9bd7bd1b7f fix(reset): say what the reset actually did, per member and per file AnimateDread 2026-08-17 17:57:26 -04:00