方法库
13
搜索页面、指标...
把基金经理、CTA、量化交易和期货基本面常用数据组织成可解释的机会评分框架。所有信号都要求来源、周期、反证和验证纪律。
方法库
13
论文 / 实战
136/100
数据分类
7
数据源状态
4/10
Start with price because every other signal can be early. Use 7D/30D/90D momentum, drawdown, realized volatility, breadth, and cross-asset confirmation.
Trade with the trend only when momentum, volume, and volatility are not contradicting each other.
Confirm whether new capital can support the setup: stablecoin supply, ETF/fund flows, M2, dollar liquidity, exchange reserves, and spot premiums.
Raise confidence when flow expands with price; downgrade rallies when price rises on shrinking liquidity.
Detect whether a move is spot-led or leverage-led using funding, basis, OI, liquidations, long/short ratios, and options skew.
High funding plus rising OI is not confirmation; it is a warning to reduce leverage or wait for a flush.
Use slow anchors such as MVRV, realized price, miner cost, NUPL, Puell, value/profitability, and factor valuation.
Treat valuation as zone selection; require faster signals for timing.
Track news, filings, macro calendars, Fear & Greed, social volume, search attention, whale events, and ETF reporting lags.
Use sentiment extremes contrarily only when price and leverage confirm exhaustion.
A signal is useful only after sizing. Use volatility targeting, invalidation stops, drawdown kill switches, slippage, depth, and liquidation buffers.
No signal bypasses risk: size from volatility and invalidation first, then decide direction.
Liu, Tsyvinski, Wu (2022, JoF)
CMKT、CSMB、CMOM 三因子解释加密横截面收益。
Grobys, Näsman, Sandretto (2026, RIBF)
同行评审验证 NUPL、MVRV Z-Score、CVDD;链上阈值仍需按周期衰减处理。
Gilchrist & Zakrajšek (2012, AER)
Excess Bond Premium 对衰退预测强,但不适合精确底部择时。
Lopez-Salido, Stein, Zakrajšek (2017, QJE)
过低信用利差往往预示未来增长恶化,而非即时底部信号。
所有机会判断的第一层:趋势、动量、回撤、波动率、成交量和广度。
识别拥挤、强平、逼空、carry 和波动率定价。
解释风险资产的顺风/逆风背景,避免只看单一市场信号。
把商品、利率和金融期货从价格信号扩展到库存、供需和仓位。
让股票机会不只依赖价格,加入估值、盈利、现金流、内部人和机构持仓。
补足链上活动、稳定币流动性、矿工成本、DeFi TVL 和交易所库存。
把新闻、公告、央行发言、财报电话会和社交关注度转成事件与情绪信号。
判断资产自身趋势是否延续,适合 1 周到 12 月的持仓与风控。
使用 30/90/180 日收益、均线距离、回撤恢复和波动率调整后的趋势分。
比较资产、行业、风格因子谁在获得增量资金。
比较同一资产池 1/3/6 月收益和最新变化率,生成轮动候选。
捕捉远近期价格、资金费率或利差带来的持仓补偿;资金费率 carry 在 2025 拥挤后年化转负,不能当作稳定收益。
使用永续 basis、资金费率、期货曲线、信用利差和利率曲线斜率。
识别风险溢价扩大、极端波动和低波动后的方向选择。
使用实现波动率、VIX/DVOL、收益 Z-score、影线和成交量尖峰。
判断上涨是否由健康资金推动,还是由拥挤杠杆驱动;OI 方向不唯一,爆仓热力图多为估算。
组合 OI、资金费率、大户多空比、COT、期权偏度和爆仓代理指标。
用利率、通胀、信用、美元和流动性判断风险资产顺风或逆风。
组合政策利率、2Y/10Y、美债实际利率、信用利差、M2、DXY 和 VIX。
衡量价格相对盈利、现金流、账面价值、成本曲线或历史区间是否便宜;BTC 链上峰值阈值逐周期下降,应使用历史分位。
股票使用估值和盈利修正;加密使用矿工成本、MVRV 类代理和历史分位。
确认趋势背后是否有稳定币、ETF、成交量、信用和货币供应支持。
使用稳定币市值、成交量 Z-score、M2、信用利差、ETF/基金流和链上流量。
在风险预算收缩时筛选现金流、盈利能力和低波动更强的资产。
使用 QUAL、USMV、SPLV、利润率、杠杆率、自由现金流和信用压力代理。
捕捉财报、库存、产量、供需和政策发布后的预期差。
对 SEC、EIA、COT、BLS、FRED 发布值计算 surprise、修正方向和滞后风险。
衡量新闻、社交、恐慌贪婪和投资者关注度是否过热或恐慌。
使用资讯 tone、恐慌贪婪、搜索/社交关注代理和事件标题聚类。
跟踪央行、财报、库存、监管和链上解锁等离散风险。
给每个事件绑定资产、预期方向、发布时间、历史波动和失效条件。
用矿工成本、MVRV、SOPR 和已实现价格判断 BTC 周期阶段;所有顶部阈值逐周期下降,固定触发位只能做 regime 背景。
价格 / 矿工成本比值;MVRV Z-Score Z<0 底部、Z>7 旧顶部但需分位;SOPR 1.0 是盈亏枢轴。
Markowitz · 1952 · risk
Start every signal stack with sizing, covariance, and diversification.
Fama and French · 1992 · value
Size and book-to-market remain core equity ranking variables.
Fama and French · 1993 · value / risk
Use factor exposures to separate alpha from broad risk.
Fama and French · 2015 · value / quality
Profitability and investment improve equity selection frameworks.
Jegadeesh and Titman · 1993 · momentum
Cross-sectional momentum is a durable ranking signal.
Carhart · 1997 · momentum / risk
Momentum belongs in baseline performance attribution.
Asness, Moskowitz and Pedersen · 2013 · value / momentum
Value and momentum diversify across asset classes.
Moskowitz, Ooi and Pedersen · 2012 · trend
Own-asset 1-12 month trend is a core CTA-style signal.
Hurst, Ooi and Pedersen · 2017 · trend / risk
Diversified trend can persist across long samples.
Brock, Lakonishok and LeBaron · 1992 · trend
Moving-average and breakout rules can contain signal.
Lo, Mamaysky and Wang · 2000 · trend / volatility
Pattern recognition needs formal testing, not chart folklore.
Han, Yang and Zhou · 2013 · trend / volatility
Technical timing can be stronger in high-volatility portfolios.
CryptoQuant · On-chain
Falling reserves reduce immediate sell inventory; rising reserves warn sell pressure.
CryptoQuant · On-chain
Negative netflow supports accumulation; positive netflow warns distribution.
CryptoQuant · On-chain
Large inflows often precede selling or volatility.
CryptoQuant · On-chain
Positive premium suggests stronger US spot demand.
CryptoQuant · On-chain
Price above realized price supports macro recovery; below marks capitulation risk.
Glassnode · On-chain
SOPR above 1 means profit-taking; 1.0 support retest confirms bull demand.
Glassnode · On-chain
Low MVRV marks value; high MVRV marks profit saturation.
Glassnode · On-chain
STH-MVRV near or below 1 shows short-holder capitulation/support.
Glassnode · On-chain
Euphoria zones reduce reward/risk; capitulation zones improve it.
Glassnode · On-chain
Large entities accumulating supports trend quality.
OKX/Binance/Bybit · Derivatives
Very positive funding means crowded longs; very negative means crowded shorts.
OKX · Derivatives
Rising OI plus extreme funding warns liquidation risk.
Perp candles, OI, funding, basis, long/short ratios, taker flow, and order-book signals.
Rates, inflation, credit, liquidity, dollar, and macro regime data.
Equity index, ETF, FX, commodity, and stock proxy coverage.
Stablecoin liquidity, DeFi TVL, and crypto-native flow context.
BTC price history, network activity, hashrate, fees, and miner-cost modeling.
DVOL, options OI, max-pain style expiry surfaces, and volatility regimes.
MVRV, SOPR, NUPL, HODL waves, Puell, Reserve Risk, and accumulation signals.
Exchange flows, reserves, Coinbase premium, miner flows, SSR, and cohort ratios.
Realized cap, stablecoin aggregates, supply, network activity, and market reference rates.
Bid-ask spreads, 1%-2% market depth, slippage, and venue liquidity quality.
Daily BTC ETF flows, issuer breadth, GBTC drag, and weekly fund-flow confirmation.
Filings, company facts, insider events, and futures positioning for stock/macro signals.
每个信号必须记录数据源、更新时间、lookback、换仓频率和适用市场。
回测必须避免未来函数,使用 walk-forward / out-of-sample 验证并显式扣除交易成本。
新增因子默认标记为 experimental,只有跨周期稳定后才进入 research-backed。
同方向信号必须展示反证项,例如美元走强、信用恶化、仓位拥挤或数据过期。