面向即時市場、經得起驗證的交易系統研究。https://marketmaker.cc/zh-Hanthttps://marketmaker.cc/zh-Hant/blog/arbitrage-prereq-6-rust-basics/https://marketmaker.cc/zh-Hant/blog/arbitrage-prereq-6-rust-basics/套利課程的從零開始短課。透過一個讀取四個價格的小程式，講解 Rust 的變數、迴圈、向量和函式，以及為什麼發布版本更快。程式碼可執行，附有輸出。Thu, 12 Nov 2026 00:00:00 GMT套利課程預備課rust性能https://marketmaker.cc/zh-Hant/blog/arbitrage-prereq-5-probability/https://marketmaker.cc/zh-Hant/blog/arbitrage-prereq-5-probability/套利課程的從零開始短課。什麼是機率，怎樣透過計數來測量它，什麼是鐘形曲線，以及為什麼 z 分數大於 2 很少見。程式碼可執行，附有輸出。Thu, 12 Nov 2026 00:00:00 GMT套利課程預備課機率正態分布Pythonhttps://marketmaker.cc/zh-Hant/blog/arbitrage-prereq-4-vectors-matrices/https://marketmaker.cc/zh-Hant/blog/arbitrage-prereq-4-vectors-matrices/套利課程的從零開始短課。把一列數看作向量，把表格看作矩陣，用 numpy 對所有數一次完成同一個運算，以及什麼是廣播。程式碼可執行，附有輸出。Thu, 12 Nov 2026 00:00:00 GMT套利課程預備課numpy向量矩陣Pythonhttps://marketmaker.cc/zh-Hant/blog/arbitrage-prereq-3-mean-stdev/https://marketmaker.cc/zh-Hant/blog/arbitrage-prereq-3-mean-stdev/套利課程的短篇零基礎課。在十個數字上講解平均數、標準差和 z-score，說明為什麼一個異常值會改變一切，以及 z-score 如何告訴你何時出現異常。Mon, 09 Nov 2026 00:00:00 GMT套利課程先備知識統計z-scorePythonhttps://marketmaker.cc/zh-Hant/blog/arbitrage-prereq-2-graphs/https://marketmaker.cc/zh-Hant/blog/arbitrage-prereq-2-graphs/套利課程的短篇零基礎課。什麼是圖，如何用 Python 儲存圖，如何列出兩點之間的每一條路徑，以及為什麼地圖變大後，逐條嘗試路徑會失效。Mon, 09 Nov 2026 00:00:00 GMT套利課程先備知識圖最短路徑Pythonhttps://marketmaker.cc/zh-Hant/blog/arbitrage-prereq-1-logarithms/https://marketmaker.cc/zh-Hant/blog/arbitrage-prereq-1-logarithms/套利課程的短篇零基礎課。對數如何處理一串匯率，為什麼對數總和會取代乘積，以及如何理解負號。附有可手動核對輸出的簡短 Python 程式碼。Mon, 09 Nov 2026 00:00:00 GMT套利課程先備知識對數數學Pythonhttps://marketmaker.cc/zh-Hant/blog/arbitrage-course-16-release/https://marketmaker.cc/zh-Hant/blog/arbitrage-course-16-release/套利課程第 16 課，也是最後一課。將模擬交易機器人變成可以展示的成果：交易日誌、報告頁面、README、沒有機密資訊的乾淨公開儲存庫，以及誠實介紹結果的方法。Mon, 09 Nov 2026 00:00:00 GMT套利課程初學者發布READMEgithub模擬交易pythonhttps://marketmaker.cc/zh-Hant/blog/arbitrage-course-15-pairs-trading/https://marketmaker.cc/zh-Hant/blog/arbitrage-course-15-pairs-trading/套利課程第 15 課。為什麼比較兩個原始價格沒有意義，價差及其 z-score 如何顯示兩枚關聯幣何時偏離，以及一個展示手續費如何吃掉優勢的小型回測。附有可重現輸出的 Python 程式碼。Mon, 09 Nov 2026 00:00:00 GMT套利課程初學者配對交易統計套利z-score均值回歸pythonhttps://marketmaker.cc/zh-Hant/blog/arbitrage-course-14-futures-spot-funding/https://marketmaker.cc/zh-Hant/blog/arbitrage-course-14-futures-spot-funding/套利課程第 14 課。什麼是永續期貨及其資金費率，如何在現貨市場持有一枚幣、在期貨上持有相反倉位，以及手續費、負費率和強制平倉如何影響結果。附有可手動核對數字的簡短 Python 程式碼。Mon, 09 Nov 2026 00:00:00 GMT套利課程初學者期貨資金費率永續合約現貨持有套利pythonhttps://marketmaker.cc/zh-Hant/blog/arbitrage-course-13-speed/https://marketmaker.cc/zh-Hant/blog/arbitrage-course-13-speed/套利課程第 13 課。用迴圈檢查每個三幣種環路，再用 numpy 矩陣技巧和一段簡短的 Rust 程式檢查，並測量三種方法。附有可執行的程式碼和實際計時結果。Mon, 09 Nov 2026 00:00:00 GMT套利課程初學者numpy向量化rust效能pythonhttps://marketmaker.cc/zh-Hant/blog/arbitrage-course-12-paper-trading-risk/https://marketmaker.cc/zh-Hant/blog/arbitrage-course-12-paper-trading-risk/套利課程第 12 課。讓機器人以即時價格和虛擬資金運作，把每個數字移入設定檔，並加入風險限額，在糟糕的一小時造成損失前停止機器人。附可核對輸出的簡短 Python 程式碼。Mon, 09 Nov 2026 00:00:00 GMT套利課程初學者模擬交易風險管理設定Pythonhttps://marketmaker.cc/zh-Hant/blog/arbitrage-course-11-reality-check/https://marketmaker.cc/zh-Hant/blog/arbitrage-course-11-reality-check/套利課程第 11 課。延遲、部分成交、失敗的第二腿和過時價格，搭配一個小型模擬，展示獲利計畫如何變成平均虧損。附有可執行和修改的簡短 Python 程式碼。Mon, 09 Nov 2026 00:00:00 GMT套利課程初學者延遲執行風險模擬Pythonhttps://marketmaker.cc/zh-Hant/blog/arbitrage-course-10-size-fees-slippage/https://marketmaker.cc/zh-Hant/blog/arbitrage-course-10-size-fees-slippage/套利課程第 10 課。把手續費和走訪委託簿結合起來，計算利潤如何隨交易數量變化，並找出利潤最高的數量。附有可手算檢查數字的簡短 Python 程式碼。Mon, 09 Nov 2026 00:00:00 GMT套利課程初學者手續費滑價部位大小Pythonhttps://marketmaker.cc/zh-Hant/blog/arbitrage-course-09-order-book/https://marketmaker.cc/zh-Hant/blog/arbitrage-course-09-order-book/套利課程第 9 課。什麼是委託簿、如何閱讀附有數量的買賣價、如何走訪委託簿以找出交易的真實價格，以及一個可手算檢查的簡短 Python 函式。Mon, 09 Nov 2026 00:00:00 GMT套利課程初學者委託簿滑價Pythonhttps://marketmaker.cc/zh-Hant/blog/arbitrage-course-08-rich/https://marketmaker.cc/zh-Hant/blog/arbitrage-course-08-rich/套利課程第 8 課。為什麼機器人想找步數有限的循環、RICH 演算法（VLDB 2025）背後的色彩編碼技巧、一個小型可執行 Python 版本，以及誠實說明哪些部分被簡化了。Mon, 09 Nov 2026 00:00:00 GMT套利課程初學者RICH色彩編碼負環k 跳Pythonhttps://marketmaker.cc/zh-Hant/blog/arbitrage-course-07-spfa/https://marketmaker.cc/zh-Hant/blog/arbitrage-course-07-spfa/套利課程第 7 課。SPFA 如何只重新檢查可能改變的道路、一步步的佇列範例、含負環警報的可執行 Python 程式，以及和 Bellman-Ford 的速度比較。Mon, 09 Nov 2026 00:00:00 GMT套利課程初學者SPFABellman-Ford佇列負環Pythonhttps://marketmaker.cc/zh-Hant/blog/arbitrage-course-06-bellman-ford/https://marketmaker.cc/zh-Hant/blog/arbitrage-course-06-bellman-ford/套利課程第 6 課。用一個很小的例子說明 Bellman-Ford 演算法、為什麼多做一輪能發現負環、會回傳獲利循環的 Python 程式，以及如何手算檢查。Mon, 09 Nov 2026 00:00:00 GMT套利課程初學者Bellman-Ford負環圖論演算法Pythonhttps://marketmaker.cc/zh-Hant/blog/arbitrage-course-05-rates-as-a-graph/https://marketmaker.cc/zh-Hant/blog/arbitrage-course-05-rates-as-a-graph/套利課程第 5 課。貨幣作為節點、匯率作為箭頭、乘法為何能用對數變成加法，以及獲利的一圈為何會變成總和為負的一圈。Mon, 09 Nov 2026 00:00:00 GMT套利課程初學者圖對數三角套利https://marketmaker.cc/zh-Hant/blog/arbitrage-course-04-compare-best-prices/https://marketmaker.cc/zh-Hant/blog/arbitrage-course-04-compare-best-prices/套利課程第 4 課。買賣價差、手續費、報價貨幣和單位，以及一個掃描五間交易所並回報哪些價差扣完手續費後仍存在的短 Python 程式。Mon, 09 Nov 2026 00:00:00 GMT套利課程初學者買賣價差手續費Pythonhttps://marketmaker.cc/zh-Hant/blog/arbitrage-course-03-stages-and-design/https://marketmaker.cc/zh-Hant/blog/arbitrage-course-03-stages-and-design/套利課程第 3 課。為什麼機器人要分階段運作、每個階段做什麼，以及如何請 AI 代理人提供你看得懂也能檢查的設計。Mon, 09 Nov 2026 00:00:00 GMT套利課程初學者架構委託簿模擬交易https://marketmaker.cc/zh-Hant/blog/arbitrage-course-02-kinds-of-arbitrage/https://marketmaker.cc/zh-Hant/blog/arbitrage-course-02-kinds-of-arbitrage/套利課程第 2 課。四種主要套利（跨交易所、循環、統計、期貨現貨），各有一個小例子，以及如何選擇你的機器人要尋找哪一種。Mon, 09 Nov 2026 00:00:00 GMT套利課程初學者三角套利統計套利資金費率https://marketmaker.cc/zh-Hant/blog/arbitrage-course-01-what-is-arbitrage/https://marketmaker.cc/zh-Hant/blog/arbitrage-course-01-what-is-arbitrage/套利課程第 1 課。認識套利的意思、價格落差為何存在，以及用十行 Python 寫出的掃描器，比較同一枚幣在兩間交易所的最佳買價和最佳賣價。Mon, 09 Nov 2026 00:00:00 GMT套利課程初學者買價賣價Pythonhttps://marketmaker.cc/zh-Hant/blog/xgboost-gradient-boosting-trading/https://marketmaker.cc/zh-Hant/blog/xgboost-gradient-boosting-trading/報酬方向分類器是不平衡問題，而 0.5 是錯誤的決策閾值。比較加密貨幣資料上的 scale_pos_weight、焦點損失與精確率約束閾值最佳化，以及 XGBoost、LightGBM 和 CatBoost 之間的工程差異。Tue, 25 Aug 2026 00:00:00 GMTmachine-learningxgboostgradient-boostingquantpredictionhttps://marketmaker.cc/zh-Hant/blog/transfer-entropy-crypto-information-flow/https://marketmaker.cc/zh-Hant/blog/transfer-entropy-crypto-information-flow/DCC-GARCH tells you when crypto dependence tightens. Transfer entropy tells you which way it points. A directed information-flow measure, its null calibration, and an honest account of what it does and does not add over average pairwise correlation.Mon, 24 Aug 2026 00:00:00 GMTcausal-inferencetransfer-entropyinformation-theorycryptonetworkhttps://marketmaker.cc/zh-Hant/blog/trade-classification-algorithms/https://marketmaker.cc/zh-Hant/blog/trade-classification-algorithms/從 OHLCV、棒歷史資料或沒有攻擊方標記的交易場所重建交易方向 - - 經典規則、過時報價問題、批量成交量分類，以及如何用加密市場免費提供的真實標籤測量它們。Sun, 23 Aug 2026 00:00:00 GMTmicrostructuretrade-classificationLee-Readyorder-flowdeep-learninghttps://marketmaker.cc/zh-Hant/blog/tick-data-forecasting-deep-learning/https://marketmaker.cc/zh-Hant/blog/tick-data-forecasting-deep-learning/餵入逐筆資料的序列模型仍假設間隔規則。本文介紹三種告訴 Transformer 某筆成交實際發生時間的方法 - - 可學習時間尺度的連續編碼、ODE-RNN 潛在狀態，以及把 delta_t 作為一般特徵 - - 並以消融實驗決定它們之間的取捨。Sat, 22 Aug 2026 00:00:00 GMTHFTtick-datadeep-learninghigh-frequencypredictionhttps://marketmaker.cc/zh-Hant/blog/synthetic-control-trading-evaluation/https://marketmaker.cc/zh-Hant/blog/synthetic-control-trading-evaluation/This series has priced the selection route to a false edge - DSR prices the winner, PBO prices the search. Neither touches confounding: the strategy that made money because volatility doubled the week you deployed it. The Synthetic Control Method builds a weighted counterfactual from a donor pool of untouched instruments and gives you a falsification criterion and a placebo p-value.Fri, 21 Aug 2026 00:00:00 GMTcausal-inferencesynthetic-controlcounterfactualevaluationeconometricshttps://marketmaker.cc/zh-Hant/blog/probabilistic-forecasting-trading/https://marketmaker.cc/zh-Hant/blog/probabilistic-forecasting-trading/How to evaluate a predictive distribution honestly - CRPS as a proper scoring rule, the PIT histogram as a calibration diagnostic, and DeepAR sampling in GluonTS.Thu, 20 Aug 2026 00:00:00 GMTforecastingprobabilisticquantile-regressionDeepARuncertaintyhttps://marketmaker.cc/zh-Hant/blog/pinn-options-pricing/https://marketmaker.cc/zh-Hant/blog/pinn-options-pricing/Putting the Heston, free-boundary, and jump-diffusion pricing PDEs into a neural network loss - the log-price residual, the mixed-partial autograd trick, and what still needs measuring.Wed, 19 Aug 2026 00:00:00 GMTdeep-learningPINNoptionsBlack-ScholesPDEhttps://marketmaker.cc/zh-Hant/blog/pcmci-causal-discovery-crypto/https://marketmaker.cc/zh-Hant/blog/pcmci-causal-discovery-crypto/How PCMCI's two-stage MCI test recovers directed causal links between crypto assets where correlation and bivariate Granger cannot - the construction, the tigramite pipeline, and the real-data study it still needs.Tue, 18 Aug 2026 00:00:00 GMTcausal-inferencePCMCIcausal-discoverytime-seriescryptohttps://marketmaker.cc/zh-Hant/blog/order-flow-imbalance-prediction/https://marketmaker.cc/zh-Hant/blog/order-flow-imbalance-prediction/Turning raw book updates into a signed flow quantity: the CKS event decomposition, multi-level OFI with PCA reduction, and Lee-Ready trade classification - plus an honest accounting of what the headline R-squared actually measures.Mon, 17 Aug 2026 00:00:00 GMTmicrostructureorder-flowimbalancepredictionHFThttps://marketmaker.cc/zh-Hant/blog/neural-ode-continuous-finance/https://marketmaker.cc/zh-Hant/blog/neural-ode-continuous-finance/Neural ODEs, Neural SDEs and continuous normalizing flows for irregularly-sampled market data - and the one ablation that decides whether continuous dynamics are worth their solver cost.Sun, 16 Aug 2026 00:00:00 GMTdeep-learningneural-ODEcontinuous-timeSDEdynamicshttps://marketmaker.cc/zh-Hant/blog/multi-task-learning-trading/https://marketmaker.cc/zh-Hant/blog/multi-task-learning-trading/Does jointly predicting return, volume, and volatility actually help? Measuring loss-balancing schemes and diagnosing negative transfer through gradient cosine similarity - with a classical baseline and purged walk-forward folds.Sat, 15 Aug 2026 00:00:00 GMTdeep-learningmulti-taskshared-representationpredictionauxiliaryhttps://marketmaker.cc/zh-Hant/blog/model-pruning-low-latency-trading/https://marketmaker.cc/zh-Hant/blog/model-pruning-low-latency-trading/Magnitude and structured pruning, the Lottery Ticket Hypothesis, movement pruning, distillation and 2:4 sparsity - the methods behind shrinking a trading model, and what still has to be measured before any of it ships.Fri, 14 Aug 2026 00:00:00 GMTmodel-compressionpruninglottery-ticketlatencydeploymenthttps://marketmaker.cc/zh-Hant/blog/liquidity-prediction-execution/https://marketmaker.cc/zh-Hant/blog/liquidity-prediction-execution/Our VWAP article shipped a static, weekly-refit volume curve and called the forecaster the weakest link. This is the follow-up: a Bayesian intraday updater run against the same 500-parent BTCUSDT harness, with the IS delta conditioned on realized curve error.Thu, 13 Aug 2026 00:00:00 GMTmicrostructureliquidityexecutionpredictionorder-bookhttps://marketmaker.cc/zh-Hant/blog/koopman-operator-market-dynamics/https://marketmaker.cc/zh-Hant/blog/koopman-operator-market-dynamics/Dynamic Mode Decomposition fits a linear operator to nonlinear market dynamics. The only question that matters: do the fitted modes persist from one window to the next, and does the rolling spectral radius lead realised volatility? Here is the measurement protocol and the code to run it.Wed, 12 Aug 2026 00:00:00 GMTmathematicsKoopmandynamical-systemsspectralpredictionhttps://marketmaker.cc/zh-Hant/blog/knowledge-distillation-trading-models/https://marketmaker.cc/zh-Hant/blog/knowledge-distillation-trading-models/The blog's standing answer to the accuracy-vs-latency tension is a two-stage fast/slow split. Distillation is a different answer: train one small model to mimic the ensemble. The KD loss, temperature, born-again nets, early exits for a variable latency budget, and the distill-to-FPGA pipeline - plus the measurements that would decide whether it beats the two-stage split.Tue, 11 Aug 2026 00:00:00 GMTmodel-compressiondistillationlatencyHFTdeploymenthttps://marketmaker.cc/zh-Hant/blog/instrumental-variables-quantitative-finance/https://marketmaker.cc/zh-Hant/blog/instrumental-variables-quantitative-finance/Selection bias in the search is not the only way a regression fools you. When the regressor is correlated with the error, more data makes the estimate more confidently wrong. Instrumental variables applied to the one endogeneity problem this blog has already left open: permanent impact from net taker flow.Mon, 10 Aug 2026 00:00:00 GMTcausal-inferenceinstrumental-variables2SLSeconometricsendogeneityhttps://marketmaker.cc/zh-Hant/blog/hawkes-process-order-arrival/https://marketmaker.cc/zh-Hant/blog/hawkes-process-order-arrival/Fitting a self-exciting point process to real crypto trade tape: where the three numbers (mu, alpha, beta) come from, how to estimate the branching ratio n, whether the exponential kernel survives a goodness-of-fit test, and how much n moves when you change the estimation window.Sun, 09 Aug 2026 00:00:00 GMTmicrostructureHawkes-processpoint-processorder-arrivalcryptoHFThttps://marketmaker.cc/zh-Hant/blog/hamiltonian-neural-networks-finance/https://marketmaker.cc/zh-Hant/blog/hamiltonian-neural-networks-finance/Hamiltonian Neural Networks are provably stable - but stability is worthless if the conserved quantity does not exist. Learning a scalar H via autograd, symplectic integration, and the falsification test that decides whether a financial (q, p) pair is canonical at all.Sat, 08 Aug 2026 00:00:00 GMTdeep-learningHamiltonianphysics-informeddynamicsconservationhttps://marketmaker.cc/zh-Hant/blog/granger-causality-lead-lag-crypto/https://marketmaker.cc/zh-Hant/blog/granger-causality-lead-lag-crypto/Granger causality on crypto prices done two ways - differenced returns and Toda-Yamamoto on levels - with a correct Wald implementation, an effective-N corrected causality matrix, and a rolling-stability test of whether the lag is tradeable at all.Fri, 07 Aug 2026 00:00:00 GMTcausal-inferenceGranger-causalitylead-lagcryptoVARhttps://marketmaker.cc/zh-Hant/blog/gaussian-processes-trading/https://marketmaker.cc/zh-Hant/blog/gaussian-processes-trading/金融時間序列的核設計 - - 母體粗糙度、局部週期成分、光譜混合 - - 以及作為不需要驗證集的正則化器的邊際似然。加上在任何東西可以交易之前仍然需要衡量的誠實清單。Thu, 06 Aug 2026 00:00:00 GMTbayesiangaussian-processkerneluncertaintynon-parametrichttps://marketmaker.cc/zh-Hant/blog/fourier-neural-operator-finance/https://marketmaker.cc/zh-Hant/blog/fourier-neural-operator-finance/算子學習映射整個函數空間，而不是點 - - 傅立葉神經算子如何在頻率空間中參數化 PDE 解算子，為選擇權定價購買什麼，以及它的哪些主張仍然需要在真實硬體上測量。Wed, 05 Aug 2026 00:00:00 GMTdeep-learningFNOPDEoperator-learningoptionshttps://marketmaker.cc/zh-Hant/blog/flash-attention-trading-models/https://marketmaker.cc/zh-Hant/blog/flash-attention-trading-models/Flash Attention 讓 23,400 步驟的交易上下文無需計算 - - 平鋪、線上 softmax 和 N^2 d^2 / M 的 IO 界限。更長的上下文是否使模型更好是一個單獨的問題，它必須被測量，而不是斷言。Tue, 04 Aug 2026 00:00:00 GMTdeep-learningattentionFlash-AttentionGPUoptimizationhttps://marketmaker.cc/zh-Hant/blog/ensemble-methods-trading/https://marketmaker.cc/zh-Hant/blog/ensemble-methods-trading/本部落格未涵蓋有關交易整合的三件事：為什麼協方差項（而不是模型計數）在整合誤差中占主導地位，堆疊如何作為具有折疊元特徵的訊號組合層工作，以及跨多個訊號的周轉淨額是否真正降低了執行成本。Mon, 03 Aug 2026 00:00:00 GMTmachine-learningensemblebaggingstackingalphahttps://marketmaker.cc/zh-Hant/blog/double-ml-causal-trading/https://marketmaker.cc/zh-Hant/blog/double-ml-causal-trading/本博客至今所有模型回答「什麼預測什麼」。Double ML 回答「什麼導致什麼」 - - 帶有可辯護的標準誤。部分線性模型、Neyman 正交性、order book 數據上的清理交叉擬合，以及為何有效 DML 信賴區間僅對一個預先指定的問題成立的誠實說明。Sun, 02 Aug 2026 00:00:00 GMTcausal-inferencedouble-MLtreatment-effecteconometricsquanthttps://marketmaker.cc/zh-Hant/blog/causal-forests-treatment-effects-trading/https://marketmaker.cc/zh-Hant/blog/causal-forests-treatment-effects-trading/本博客每個回測估計條件平均值。因果森林估計條件治療效果 - tau(x) 而非 mu(x) - 具有誠實分裂、自適應核權重表示以及告訴您發現的異質性是否真實的校準測試。Sat, 01 Aug 2026 00:00:00 GMTcausal-inferencecausal-forestheterogeneous-effectseconometricstreatmenthttps://marketmaker.cc/zh-Hant/blog/bayesian-neural-networks-trading/https://marketmaker.cc/zh-Hant/blog/bayesian-neural-networks-trading/本博客的每個位置 sizing 規則都把不確定性坍縮成單一純量。MC Dropout 和深度集成將其拆分為模型無知與市場噪聲 - - 這兩者值得不同的 position size。Fri, 31 Jul 2026 00:00:00 GMTbayesianuncertaintyneural-networkriskposition-sizinghttps://marketmaker.cc/zh-Hant/blog/automl-systematic-trading/https://marketmaker.cc/zh-Hant/blog/automl-systematic-trading/自動特徵生成（tsfresh、Featuretools）、具預算感知的模型搜尋（FLAML 的節省成本最佳化器），以及 WorldQuant 公式化 alpha 工廠 - - 本部落格搜尋與過度擬合系列尚未涵蓋的研究流程部分，以及該系列自身結果對它們的啟示。Thu, 30 Jul 2026 00:00:00 GMTAutoMLNASfeature-engineeringautomationquanthttps://marketmaker.cc/zh-Hant/blog/liquidation-cascades-trading-signal/https://marketmaker.cc/zh-Hant/blog/liquidation-cascades-trading-signal/Перевод статьи на zh-HantWed, 29 Jul 2026 00:00:00 GMTliquidationscascadetrading signaldefiperpetualsorder bookmarket microstructurecrypto