TrueBeats

TrueBeats predicts EPS & Revenue surprises for over 35,000 global equities since 2000, outperforming simpler models by incorporating other significant sources of information about a company’s likelihood to beat or miss consensus forecasts, including: individual analyst data; metrics of earnings management; and time-series trends. The result is a surprise forecast which reduces errors and is consistently more accurate than simpler methods – it predicts the direction of surprises up to 80% of the time, and is especially accurate for larger, higher-coverage stocks. TrueBeats can be an early indicator not just of earnings and revenue surprises but also returns. Stocks with the largest positive current-quarter EPS TrueBeats have outperformed those with the largest negative TrueBeats by 15% per annum, and a simple dollar-neutral portfolio on liquid U.S. equities generated a historical Sharpe ratio of 1.6 with modest turnover. Table includes the following fields: - Date (date on which a user could trade on the data) - TrueBeat_ITEM (TrueBeat value as float data type, which indicates the expected surprise relative to the consensus) - Fiscal_Period (annual or quarterly fiscal period, for example “2020” for fiscal year 2020 and “2020 Q1” for the first fiscal quarter in 2020) - Number_Of_analysts (number of analyst estimates underlying the TrueBeat computation) - Ticker (most recent ticker for the security when historical data was cut) - Ticker_PointInTime (point in time ticker on Date) - CUSIP (most recent CUSIP for the security when historical data was cut) - CUSIP_PointInTime (point in time CUSIP on Date) - ISIN (most recent ISIN for the security when historical data was cut) - ISIN_PointInTime (point in time ISIN on Date)TrueBeats 自2000年起便可针对全球超35000只股票预测每股收益（Earnings Per Share，以下简称EPS）与营收超预期、不及预期的情形。该模型通过纳入更多能够反映公司超出或未达市场一致预期概率的有效信息维度，表现优于传统简易模型，所纳入的信息维度包括：个股分析师预估数据、盈余管理指标，以及时间序列趋势。最终生成的超预期预测结果可降低预测误差，且精度始终优于简易方法：其预测收益与营收超预期方向的准确率可达80%，尤其在市值较大、分析师覆盖度较高的个股上表现更为精准。 TrueBeats不仅可作为收益与营收超预期的先行指标，还可用于预测股票回报。当前季度每股收益TrueBeats得分最高的个股，其年化超额收益较得分最低的个股高出15%；而针对流动性良好的美国股票构建的简单美元中性投资组合，历史夏普比率可达1.6，且换手率较低。 该数据集包含以下字段： - 交易日期（Date）：用户可基于该数据开展交易的对应日期 - TrueBeat_ITEM：浮点型数据类型的TrueBeat值，用于表征相对于市场一致预期的预期超预期幅度 - 会计期间（Fiscal_Period）：年度或季度会计期间，例如“2020”代表2020财年，“2020 Q1”代表2020年首个财季 - 分析师数量（Number_Of_analysts）：计算TrueBeat值所基于的分析师预估样本总量 - 最新交易代码（Ticker）：历史数据截取时对应证券的最新交易代码 - 时点交易代码（Ticker_PointInTime）：交易日期时点对应的证券交易代码 - 最新CUSIP代码（CUSIP）：历史数据截取时对应证券的最新CUSIP（统一证券识别程序委员会）代码 - 时点CUSIP代码（CUSIP_PointInTime）：交易日期时点对应的CUSIP代码 - 最新ISIN代码（ISIN）：历史数据截取时对应证券的最新ISIN（国际证券识别码）代码 - 时点ISIN代码（ISIN_PointInTime）：交易日期时点对应的ISIN代码

搜集汇总

背景与挑战

背景概述

TrueBeats是一个全球股票EPS和营收意外预测数据集，自2000年起覆盖35,000余只股票，其创新模型整合多维度信息使预测准确率高达80%。数据集包含预测数值、财务周期、证券标识等结构化字段，并验证了其预测结果与股票超额收益间的显著相关性。