Evidence-based investing research
Value Investing Strategy (Strategy Overview)
Allocations for November 2026 (Final)
Cash TLT LQD SPY
Momentum Investing Strategy (Strategy Overview)
Allocations for November 2026 (Final)
1st ETF 2nd ETF 3rd ETF

Predicting Stock Index Reversals with Amareos Sentiment Indicators

Steve LeCompte | | Posted in: Sentiment Indicators

Guide Market Indicators Sentiment

Key Insight

Amareos sentiment indicators using random forest algorithms achieved 75.1% accuracy identifying market bottoms but missed 53.3%, with strategy outperforming buy-and-hold on 4 hold-out indexes (2005-2016; 6-month forecast horizons).

Do data-intensive, high-frequency investor sentiment measurements usefully predict stock index performance? In his May 2016 paper entitled “Can Sentiment Indicators Signal Market Reversals?”, Arnaud Lagarde applies a random forest machine learning algorithm to test the power of Amareos sentiment indications to predict stock index reversals. Algorithm training data relates sentiment to known stock index return for the next 182 days (six months). If this return is -20% or lower (+10% or higher), he designates the condition at the time of forecast as a market top (bottom). Otherwise, he designates the condition as neutral. He starts with 20 global equity indexes. He holds out four indexes (CAC40, CSI300, Nikkei and S&P500) for out-of-sample testing. He then randomly selects 80% of daily observations on the other 16 indexes for algorithm training, with the remaining 20% reserved for additional out-of-sample testing. Out-of-sample testing includes tabulation of raw top/bottom identification accuracy and a simple trading strategy that is long (in cash) after a bottom (top) indication and does not react to a neutral indications. He focuses trading strategy testing on: (1) the four hold-out indexes over the entire sample period; and, (2) the last six weeks of data for all indexes, which cannot be used for training. Using daily Amareos market sentiment readings and returns for the 20 equity indexes during January 2005 through mid-April 2016, he finds that:

  • Regarding accuracy of out-of-sample top/bottom signals:
    • Among all 20 indexes, 75.1% (92.4%) of bottom (top) signals are correct. However, 53.3% (59.6%) of bottom (top) signals are missed.
    • Among four hold-out indexes, 58.3% (37.2%) of bottom (top) signals are correct. However, 86.3% (96.5%) of bottom (tops) signals are missed.
  • Regarding trading strategy performance:
    • Based on both terminal value and maximum drawdown, the strategy outperforms buy-and-hold over the full sample period for all four hold-out indexes.
    • During the last six months of the sample period, the strategy outperforms buy-and-hold based on cumulative return for 11 of 20 indexes and matches buy-and-hold for 7 of 20. Average outperformance is 6.2%.

In summary, evidence from simple tests over a short sample period suggests that the Amareos sentiment signals may be useful for equity index timing.

Cautions regarding findings include:

  • The large day-to-day overlap (181 out of 182 days) in forecast intervals results in very high autocorrelation in forecasted returns and therefore probably much clustering of signals. Since sampling for training is random, it is very likely that training includes days from any given signal cluster, confounding discrimination between in-sample and out-of-sample data. Said differently, the sample period is very short in terms of independent 6-month return intervals, and the sampling methodology does not relieve this weakness.
  • Regarding section “3.3.2 Performances on CAC40, CSI300, Nikkei and S&P500”: The success of the strategy as applied to each of the four hold-out indexes derives essentially from avoiding one large drawdown (crash) during the sample period. Global equity indexes may have elevated return correlations during major crashes. Per the preceding caution, behaviors of the four hold-out indexes during crashes may be somewhat in-sample. [The author responds that “most of the ‘crashes’ are regional,” such as the CAC40 crash at the end of 2011 and the CSI300 crash at the end of 2015.]
  • Regarding section “3.3.3 Performance on the last six months”: As noted in the paper, this out-of-sample test is very short, with no changes in position for 7 of 20 indexes. During this subperiod, global equity markets generally decline. Even random market timing with a long-only strategy would tend to outperform buy-and-hold. [The author responds with a roughly estimated 27% probability that strategy outperformance is due to luck.]
  • As noted in the paper, strategy tests ignore trading frictions, which would lower performance. Per the paper: “However, because trading is limited – once or twice a year based on the 20 indices over the past decade – impact would definitely be minimal.” But, signals may tend to occur when liquidity is low (and trading frictions high). Also, the cost of the sentiment data reduces performance.
  • The methodology based on machine learning is likely beyond the reach of most investors, who would bear fees for delegating to an investment manager/fund.

Frequently Asked Questions

What is a market sentiment indicator and how is it used to identify stock market tops and bottoms?

A market sentiment indicator is a measure of investor mood or attitude, such as the high-frequency, data-intensive Amareos daily sentiment readings, used to anticipate stock index direction. In the study, a market top is defined as a subsequent 182-day (six-month) return of -20% or lower, and a bottom as a return of +10% or higher. Any other condition is classified as neutral.

Can sentiment indicators predict stock market reversals and beat buy-and-hold?

Amareos sentiment indicators showed some predictive power for stock index reversals in a 2005-2016 test, but the evidence is limited. Across all 20 indexes, 75.1% of bottom signals and 92.4% of top signals were correct, though 53.3% of bottoms and 59.6% of tops were missed. The timing strategy outperformed buy-and-hold on terminal value and maximum drawdown for all four hold-out indexes (CAC40, CSI300, Nikkei, S&P500).

How is a random forest machine learning model used to test sentiment-based market timing?

The random forest algorithm was trained on Amareos sentiment readings matched to each index's known return over the next 182 days, using a random 80% of daily observations from 16 of 20 global equity indexes. The remaining 20% of those observations and four hold-out indexes (CAC40, CSI300, Nikkei, S&P500) were reserved for out-of-sample testing. The trading strategy goes long after a bottom signal, moves to cash after a top signal, and does not react to neutral signals.

How does a sentiment-based timing strategy compare to buy-and-hold for stock indexes?

The sentiment-based timing strategy beat buy-and-hold on terminal value and maximum drawdown for all four hold-out indexes over January 2005 to mid-April 2016. In the last six weeks of data it outperformed buy-and-hold on cumulative return for 11 of 20 indexes and matched it for 7, with average outperformance of 6.2%. That period featured generally declining global markets, which favors a strategy that can move to cash, so the comparison may be flattered.

What are the limitations and risks of using machine learning sentiment signals to time the stock market?

The main limitations are that the 182-day forecast windows overlap by 181 days, creating heavy autocorrelation and signal clustering, and random training sampling likely blurs the line between in-sample and out-of-sample data. The strategy's success on the four hold-out indexes came essentially from avoiding one large crash each, and the short out-of-sample period of falling markets could favor any long-only timing approach. Tests also ignore trading frictions and the cost of the sentiment data, and the machine learning method is likely beyond most investors' reach without paying fund fees.

Market Indicators guide

More on sentiment

  • Testing a Proxy for the Goldman Sachs Panic Index Premium Claude-estimated Goldman Sachs Panic Index values (pixel-reconstructed from a chart) showed high readings (above 8) followed by 1.2% average next-week S&P 500 returns vs…
  • AI-simulated CFO Sentiment Premium GPT-5.4-simulated CFO optimism correlates with actual Duke-Federal Reserve CFO Survey responses (LLM mean 56.6 vs actual 60.1; regression coefficient over six standard errors from…
  • CFO U.S. Economic Sentiment and Stock Market Returns Premium CFO average U.S. economic sentiment shows little ability to predict S&P 500 returns; correlations with next-four-quarter returns are very weak and negative, while change…
  • How Are Sentiment-driven ETFs Doing? Premium An equal-weighted average of three sentiment ETFs (BUZZ, MOOD, NEWZ; monthly rebalanced) underperformed VTI (1.0% vs 10.7% CAGR; MaxDD -51% vs -25%; sample periods…
  • Goldman Sachs Panic Index Proxy Premium A VIX-percentile Panic Index Proxy (trailing 2-year rank, 0–100) shows weak lead-lag association with daily S&P 500 returns, with very high and very low…

All 309 articles in the Market Indicators guide

Guru Grades

1,206 predictions by 42 market gurus since 2022: 41% right. Always saying stocks will go higher: 62%.

As of the Nov 18, 2026 close

  1. 1 Ed Yardeni A 71%
  2. Always bullish B 62%
  3. 2 Jim Paulsen B 61%
  4. 3 Sam Stovall B 58%
  5. 4 Adam Turnquist B 58%
  6. 5 Ryan Detrick B 56%

Full leaderboard ›