Which of the 191 GTJA alphas still work in 2026?

TL;DR

Of the 191 short-horizon alphas published by GTJA in 2014, only 8 (4%) still pass our alive filter — positive mean IC above 0.02, t-stat above 2, and at least 55% of days with positive IC — on CSI 300 over 2018–2025. A further 14 (7%) have reversed sign with statistical significance and now act as contrarian signals. Another 168 (88%) no longer clear the significance filter in either direction, and one (gtja191_138) could not be tested at all: 99% of its output is NaN. Half of the 8 survivors carry the microstructure / shape-of-the-bar tag, the best survival rate of any theme, rather than the raw volume-price interaction family we expected. Exact counts below; the methodology is deliberately conservative and the caveats are non-trivial.

Numbers below are from a bench run on 2026-09-29 (Vibe-Trading 0.1.16: CSI 300 with point-in-time membership and split-adjusted prices, 2018–2025) — reproducible via the CLI snippet at the end of the post.

Background

The 2014 Guotai Junan research report — titled "191 个短周期交易型 alpha 因子" (191 Short-Period Transactional Alpha Factors) — landed at a particular moment in Chinese quant. The market had just emerged from a multi-year sideways grind, retail participation was climbing back, and the "formulaic alpha" template would be popularised internationally a year later by Kakushadze's 2015 arXiv preprint, 101 Formulaic Alphas. The GTJA team produced what is, in retrospect, one of the most systematic public catalogues of short-horizon factors ever published for the A-share universe: 191 numbered formulas, each a few lines of operator algebra over daily OHLCV plus turnover.

The formulas in that report read like first-principle hypotheses about market microstructure translated into pandas-friendly arithmetic. Some are obvious in hindsight (rank-based reversal over five days; correlation between volume and close); others are exotic (the report's SUMIF / FILTER / REGBETA compositions reach four or five operators deep). They are, in the language of modern quant, a library of priors: each one encodes someone's belief about which microstructure regularity is exploitable on a 1–5 day horizon.

The reason this matters in 2026 is simple: short-horizon alphas decay faster than any other category. A long-horizon value or quality factor can plausibly survive a decade with no adjustment; a 1–5 day formula built on volume-price interactions probably cannot. Twelve years out-of-sample is, in factor-research terms, a near-eternity. The A-share market itself has changed: T+1 settlement remains, but the institutional/retail mix has flipped, the STAR Board and ChiNext registration-system reform have re-priced small-cap risk, and high-frequency-flavoured execution by mutual funds has compressed many obvious mean-reversion windows.

So we ran the test: method first, then the numbers.

Method

The test is deliberately a cross-sectional information-coefficient (IC) study, not a full backtest. The goal is to ask: does this alpha rank stocks in a way that correlates with next-day return, on average, robustly across the window? A higher-fidelity strategy backtest (with t-cost, position limits, sector neutralisation, decay-multiplied portfolios) is a separate question and is out of scope for this post.

Universe

CSI 300: the 300 most liquid A-shares, rebalanced semi-annually by the index provider. Each day's cross-section is the index membership in force on that day, from Tushare's month-end rosters, so a stock counts only while it is in the index — 348 names were members at some point in the window. Prices are forward-adjusted for splits and dividends.

Period

2018-01-02 through 2025-12-31. Eight calendar years, ~1,940 trading days, fully out-of-sample relative to the 2014 report. The window deliberately spans 2018's bear market, the 2020–2021 liquidity-driven rally, the 2022 drawdown, and the 2024–2025 sideways regime — so the average IC is regime-averaged, not regime-cherry-picked.

Signal definition

For each alpha and each trading day t, we compute the alpha value for every stock in the universe, then cross-sectionally rank-transform it to [0, 1]. Forward return is the 1-day simple return from close t to close t+1, also cross-sectionally rank-transformed. The Spearman IC for day t is the Pearson correlation of the two ranked series. We report the mean IC across all valid days, the t-statistic of the IC series, and the fraction of days with positive IC.

Categorisation

• Alive. Mean IC > 0.02, t-stat(IC) > 2, and ≥ 55% of days with IC > 0. All three conditions must hold.
• Reversed. Mean IC < −0.02 and t-stat(IC) < −2 (sign-flipped version of the above). The original report intended the alpha to predict with one sign; we observe it predicting with the opposite sign at statistical significance.
• Dead. Everything else that could be computed: |mean IC| ≤ 0.02, or |t-stat| ≤ 2, or a positive IC on fewer than 55% of days. An alpha whose output is more than 95% NaN is not tested at all and is reported separately.

Caveats baked into the method

Three deliberate choices that constrain how the results should be interpreted:

• The IC is 1-day forward, no decay smoothing. Decay-3 or decay-5 IC (the original report's preferred horizon) will produce slightly different numbers; we'll publish those as a follow-up.
• No t-cost adjustment. A high-IC alpha with daily-rebalance turnover near 100% is, in practice, unprofitable after a realistic 5–10 bps round-trip. The bench does not measure turnover, and the survival classification here ignores it.
• No sector neutralisation. None of the GTJA 191 formulas needs sector tags, so every IC here is measured on the raw cross-section.

Tool used

$ vibe-trading alpha bench --zoo gtja191 --universe csi300 \
                            --period 2018-2025 --top 20

One command. It writes an HTML report to ~/.vibe-trading/reports/ with the top alphas by IR (mean IC, IC standard deviation, IR, share of positive-IC days, number of days) and the alphas it could not compute. Reproducibility recipe at the end of the post.

Findings

Aggregate survival

The headline counts again:

What surprised us is not the dead count — everyone expects decay — but the reversed count. A meaningful slice of alphas that worked in 2014 now act as contrarian indicators with statistically significant magnitude. The simplest reading is that a behavioural anomaly the formula was capturing (small-cap mean reversion, end-of-day flow from retail) has been crowded out by precisely the kind of systematic trading the formula represents, and what remains is the opposite trade.

Theme breakdown

Each alpha carries one or two theme tags in its registry metadata — the operator vocabulary it leans on most heavily. An alpha with two tags counts in both rows, so the counts add up to more than 190. Numbers below are survival rates within each theme:

The read from the re-run: microstructure / range alphas are the standout survivors (22%, 4 of 18). Reversal and short-horizon volatility follow at 8% (3 of 38 and 2 of 26), raw volume-price interaction at 4% (3 of 81). Momentum keeps none of its 63 — five of them now predict with the opposite sign — and the small turnover bucket keeps none of its 2. The pattern is consistent with what you would expect if a decade of systematic capital has compressed the easiest reversal/momentum trades but left intact the structural daily-bar geometry alphas that key on shape-of-the-bar effects.

Top 5 surviving alphas

The five alphas with the highest IR (mean IC divided by its standard deviation) across the window, in descending order, each with its formula as the registry stores it (__alpha_meta__["formula_latex"] of the corresponding Python module).

((RANK(DELAY(((HIGH-LOW)/(SUM(CLOSE,5)/5)),2)) * RANK(RANK(VOLUME))) / (((HIGH-LOW)/(SUM(CLOSE,5)/5))/(VWAP-CLOSE)))

-1*((l-c)*(o^5))/((c-h)*(c^5))

rank(vwap-close)/rank(vwap+close)

sma(v*((c-l)-(h-c))/(h-l),11,2)-sma(v*((c-l)-(h-c))/(h-l),4,2)

rank(((-1*ret)*mean(v,20))*vwap*(high-close))

Three alphas that now point the other way

The three alphas with the most negative mean IC. All three sit in the reversed bucket: they now predict, with statistical significance, in the direction opposite to the report's sign.

(close-vwap)/decay_linear(rank(tsmax(close,30)),2)

(c-delay(c,1))/delay(c,1)*v

16*((c-dc1+(c-o)/2+dc1-do1)/MAX_term) * MAX(abs(h-dc1), abs(l-dc1))

One paragraph of reflection

The temptation, after a survival study like this, is to over-generalise: "decay is inevitable; formulaic alphas are dead." We don't think that's the right read. What this exercise teaches is narrower and more useful: a small part of a 12-year-old short-horizon catalogue — 8 of 190 — still produces signal; the survivors cluster in an interpretable theme (bar geometry: open-close-high-low decompositions), and the losers cluster in equally interpretable ones (momentum, where nothing survives and five alphas flipped sign, and simple turnover transforms). It tells us very little about whether the Kakushadze 101 formulas, or the Qlib 158 feature set, will decay at the same rate — those zoos have different operator vocabularies and different intended universes (US equities, multi-horizon respectively). Future work, separate bench runs.

Caveats

A few constraints on how these results should be read. Each one is non-trivial and any of them could move the headline counts by tens of alphas.

1-day IC is not profitability

The whole study is at the IC level, not at the strategy-PnL level. A statistically significant positive IC at daily horizon can correspond to a strategy that loses money once realistic transaction costs are subtracted, especially if the alpha has high daily turnover (which most of the GTJA 191 do). Treat this post as a signal-quality scan, not a profitability claim. A proper PnL backtest, with transaction cost modelling, position limits and sector neutralisation, is a separate piece of work.

CSI 300 only

We benchmarked on the 300 most liquid A-shares. Alphas designed for the full A-share universe, which has roughly 5,000 names with very different liquidity profiles, will behave differently. In particular, small-cap mean-reversion alphas tend to look worse on CSI 300 than on the full universe (because CSI 300 is institutional-flow-dominated and short-horizon retail reversion is muted), and some "dead" alphas here might revive on a CSI 1000 cut.

Tushare data scope

Our data feed is Tushare end-of-day OHLCV and traded amount, forward-adjusted for splits and dividends, with VWAP derived from amount; it has no intraday tick or order-book data. Five GTJA 191 formulas reference an input the panel does not carry — the benchmark index close (gtja191_075, _149, _181, _182) or a multi-factor regression residual (gtja191_030) — and use a proxy instead: the cross-sectional mean close, or a weighted average of squared daily returns, as each alpha's notes say. Treat those five as not properly tested here, not as "dead".

8-year window is short

By academic standards an 8-year window is on the short side. The classic Fama-French papers use 30–90 year windows. Some of the alphas we label "dead" may revive on a longer or differently-positioned sample — for example, if the 2026–2030 regime returns to a more retail-dominated mix, mean-reversion alphas could recover. The label is "dead in this window", not "dead forever".

Survivorship bias in the universe

The first version of this post applied the latest CSI 300 roster to the whole 2018–2025 window, so every name was present because it survived to the end — a hindsight-selected set, which biases IC upward. The re-run masks each day's cross-section to the roster in force that day (348 names were members at some point in the window). The sp500 universe still uses today's constituent list, so the same caveat applies to any cross-zoo comparison run on it — its bench report says so — and a low alive count there should be read as "decay plus survivorship", not as a clean failure of the alpha family.

190 formulas, one test each

Testing 190 formulas on the same window makes some look good by chance. The bench reports how large the best IR would be expected to be if none of them carried any signal: 0.204 here. The best observed IR, 0.293 (gtja191_114), survives that deflation (probability 0.9999). Two more alive alphas are above 0.204 — gtja191_171 (0.230) and gtja191_120 (0.212); the other five, with IRs between 0.130 and 0.189, are below it, so read them as candidates rather than confirmed signals.

No regime conditioning

We average IC over the full 2018–2025 window. Many of these alphas are almost certainly regime-dependent — they work in trending markets but fail in choppy markets, or vice versa. A follow-up post will slice the survival counts by regime (bull / bear / sideways, as defined by 20-day index momentum) but the aggregate numbers in this post are regime-averaged. A regime-aware deployment might keep half the dead alphas as conditional signals.

Reproduce it yourself

The whole bench is one CLI command on top of an open-source install. The full source is on GitHub under the HKUDS organisation; the package is on PyPI as vibe-trading-ai.

pip install vibe-trading-ai
export TUSHARE_TOKEN=your_token_here
vibe-trading alpha bench --zoo gtja191 \
                         --universe csi300 \
                         --period 2018-2025 \
                         --top 20

You get back an HTML report saved to ~/.vibe-trading/reports/ with the top alphas by IR and the alphas that could not be computed. The same command works against --zoo alpha101 (Kakushadze 101 Formulaic Alphas, paper-faithful rewrite of the 2015 arXiv preprint) and --zoo qlib158 (the Microsoft Qlib feature library, used under Apache-2.0 with attribution). Cross-zoo comparison runs via alpha compare.

If you find an alpha whose survival classification surprises you — especially a survivor we did not flag in the top 5, or a reversed one you expected to survive — please open an issue with the alpha id and your reasoning. Community pull requests adding new zoos, new universes (CSI 1000, NASDAQ 100, crypto majors) or new validation tooling are welcome under the CONTRIBUTING.md DCO process.

Source citation: Guotai Junan Securities, "191 个短周期交易型 alpha 因子" (191 Short-Period Transactional Alpha Factors), 2014. Re-implementation agent/src/factors/zoo/gtja191/ uses only the formula content from the report; the report's narrative prose, in-sample tables and figures are not reproduced. See the directory's LICENSE.md for the full provenance note.