- Leapday Quantitative Research Team
Improving REACTION signal performance with active news analysis
In this report we qualitatively examine the largest losing and winning trades for REACTION signals to determine if extreme performance is due to accuracy of the signals or exogenous factors. We determine that for the 10 largest losing trades in 2021, 83% of losses can be attributed to information released shortly after the signal that contradicts the direction of the signal. In contrast, only 8.5% of the profits for the 10 largest winning trades can be attributed to information released shortly after the signal that affirms the direction of the signal. These results confirm the strength of the REACTION model and highlight the opportunity for investors to boost performance by optimizing exits and mitigating losses in situations where important information is released after the signal.
Read as PDF
The Leapday REACTION signals predict short-term moves after earnings announcements. The predictions range from -100 to +100 indicating the direction and magnitude of the signals. These predictions are made for fixed holding periods ranging from 2 to 15 trading days. In our previous reports, we constructed market-neutral portfolios using the 3 day signals for each entry point (t0 open + 60 min, t0 close, t1 open).
A fixed holding period, however, does not account for market moving information that becomes publicly available after the signal is generated. This report sets out to examine if the largest winning and losing trades are influenced by new information that is released after the signal is generated but before the end of the holding period. To conduct this analysis, we use signals produced by the REACTION t0 close 3 day model in 2021 and perform a qualitative analysis of the top 10 losing trades and top 10 winning trades based on P&L. The purpose of this analysis is twofold: (1) determine if positive and/or negative extreme performance in the REACTION model is due to exogenous factors that take place after the release of the signals; (2) determine if new information would have boosted performance of losing trades by providing a catalyst to exit the trade early.
We proceed by examining the top 10 losing and top 10 winning trades based on P&L in 2021 for the t0 close 3d signals and discussing the findings.
Top 10 Losers of 2021
We begin with a portfolio that trades the top one-third of t0 close 3d REACTION signals using constant capital of $50M. This is the same portfolio used in our previous reports and the full list of assumptions appears in the Appendix. We start by examining the top 10 losing trades in 2021 for this portfolio. For each of these trades, we research if any important news was released between the signal release time and the exit of the trade 3 days later. For each instance of important news we assign a direction (positive or negative) to that news based on economic intuition. If such news contradicts the REACTION signal, it could indicate that the signal itself was not necessarily incorrect and that a trader incorporating additional news into their model could be given an opportunity to exit a trade early and mitigate losses.
For each of the top 10 losing trades below, we provide details on any important news that came out after the signal. Then, we breakdown performance by day in the holding period so we can isolate how much of the P&L was impacted by any new news, and highlight (in green) how much of the P&L was impacted by the new news.
As identified above, 8 out of the top 10 losing trades of 2021 were caused by information that was released after the signal. In all of these cases, the new information contradicted the direction of the signal. While the top 10 losers of 2021 represented a total of $6,569,001 in losses (a 13.1% impact on portfolio returns), 83% of these losses equating to $5,455,114 (a 10.9% impact on portfolio returns) can be attributed to new information that was released after the signal was generated which contracted the direction of the signal.
This new information includes a mix of press releases issued by the company, sentiment from an earnings call, broad sector trends, and analyst rating changes. All of these are information sources readily available and used by investors and traders in their models and decision making processes. While it is unlikely that the full ~$5.5M in losses attributable to new information could be prevented due to immediate market reactions, using this extra information to exit trades early would undoubtedly have had a meaningful impact.
The analysis above indicates that poor performance in the largest losing trades is typically not due to bad REACTION signals. Rather, it is attributable to new information that is released after the signal is generated.
Top 10 Winners of 2021
Next we examine the top 10 winning trades of 2021 for the same portfolio as above which trades the top one-third of t0 close 3d REACTION signals using constant capital of $50M. The purpose here is to identify if there was any important news that came out after the signal which was unrelated to the earnings announcement. For each news item we establish whether it is positive or negative (relative to the signal).
For each of the top 10 winning trades below, we provide details on any important news that came out after the signal. Then, we breakdown performance by day in the holding period so we can isolate how much of the P&L was impacted by any new news, and highlight (in green) how much of the P&L was impacted by the new news.
As indicated above, only 1 out of the top 10 winning trades of 2021 was caused by important information that was released after the signal which by luck coincided with the direction of the signal. While the top 10 winners of 2021 represented a total of $5,287,152 in profits, only 8.5% of those winnings equating to $451,875 can be attributed to new information unrelated to the earnings announcement that was released after the signal was generated which affirmed the signal.
This analysis indicates the biggest winners are associated with accurate signals and are not attributable to sheer luck where new information released after the signal is the driving factor causing a trade to be very profitable.
We have demonstrated that most of the losses in the biggest losing trades can be attributed to new information that was released after the REACTION signal was generated. These results indicate that investors should incorporate additional information into their models and decision making process that would allow them to exit trades early and cut losses when major news contradicts the REACTION signal before the end of the holding period. Examples of this data include sentiment on news, press releases, earnings calls, sector trends, and analyst changes. All of these news items are readily available and already used by many investors and traders.
Additionally this report demonstrated that while almost all of the biggest losers can be attributed to new information after the signal was generated, almost none of the biggest winners were the result of luck in which new information happened to be in the same direction as the signal. These results confirm that the REACTION model works well and its performance can be boosted by utilizing additional information to optimize exits on losing trades.
In this research note we examined the top 10 winning trades and top 10 losing trades in 2021 from a portfolio that trades the top one third of signals for the REACTION t0 close 3 day model. For each extreme trade we researched whether any important news was disseminated after the release of the signal and before the end of the 3 day hold. In each case where subsequent important news existed we applied economic intuition to determine whether the news corroborated the signal or contradicted it. 8 of the top 10 losing trades had contradictory news. The impact of this contradictory news drove 83% of the underperformance in P&L from these losing trades. Meanwhile, only 1 out of the top 10 winning trades (accounting for 8.5% of the profit for these trades) had news unrelated to the earnings announcement that affirmed the signal. This result indicates that the success of the top 10 winning trades cannot be attributed to sheer luck due to new information being released subsequent to the signal. These results confirm the strength of the REACTION model and highlight the opportunity to improve performance by exiting early in situations where new news is released that contradicts the signal direction.
Appendix - Assumptions Used In Portfolio Construction
We use the strongest one-third of the t0 close 3d REACTION signals to build a simulated portfolio. Below we present all the assumptions used in the portfolio construction.
Enter a trade at the market close on the first day the market is open subsequent to a company’s earnings announcement (t0 close) and exit the trade on the close 3 days later (t3 close).
Create a market-neutral portfolio by hedging trades using an opposite side trade in SPY with a beta-adjusted notional value that offsets the beta risk of the event trade.
Starting capital of $50M, assuming constant capital throughout the portfolio (i.e. - we do not reinvest profits).
Portfolio leverage used during periods of high activity
Trade no more than 6 percent of the starting capital per position.
Trade no more than 2 percent of a company’s historical average trade value per position. (The historical average trade value is calculated using the trailing 20 day average traded value.)
Beyond the two constraints on trade size stated above, trade equal weight.
Round-trip trading costs are based on breakpoints for liquidity group as follows:
Low (<$3M historical average trade value): 20 bps
Mid ($3M-$25M historical average trade value): 10 bps
High (>$25M historical average trade value): 5 bps
Costs are allocated half on entry and half on exit
In the metrics presented below, we assume $0 in hedge costs (for the SPY hedge) and the hedge does not use any equity or margin
This material is solely for informational purposes and is not an offer or solicitation for the purchase or sale of any security, nor is it to be construed as legal or tax advice. References to securities and strategies are for illustrative purposes only and do not constitute buy or sell recommendations. The information in this report should not be used as the basis for any investment decisions. We make no representation or warranty as to the accuracy or completeness of the information contained in this report, including third-party data sources. The views expressed are as of the publication date and subject to change at any time. Hypothetical performance has many significant limitations and no representation is being made that such performance is achievable in the future. Past performance is no guarantee of future performance.