algorithmic-trading-anthology-july-2026-mjo-ag-weather-risk

Conclusion: Null

No effect at any horizon tested.

Events

6,501

occurrences

Sessions

6,501

daily

Effect

+0.00006

mean abnormal

0.415

shift null

MJO phase modulates 2-4 week ag weather risk

# daily frequency: this mechanism is invisible at monthly resolution

function test(returns, event_mask):

stat = mean(returns where event_mask) - mean(returns elsewhere)

# null: slide the event mask through the calendar, so any

# calendar structure in returns appears under the null too

null = [statistic(returns, roll(event_mask, k))

for k in all offsets, abs(k) >= 21 days]

p = (count(null beyond stat) + 1) / (len(null) + 1)

return stat, p

reject if Benjamini-Hochberg q <= 0.10 across the daily batch

+ 2 more rows− fewer rows

The null shifts the event mask through the calendar, so any calendar structure in returns appears under it too.

No fixed instrument list: the instruments are set by the rule above.

6,501 trading sessions

NOAA Climate Prediction Center. Teleconnection indices. AO, NAO, PNA, SAM. www.cpc.ncep.noaa.gov/products/precip/CWlink/daily_ao_index/ao.shtml

NOAA National Hurricane Center. HURDAT2 Atlantic hurricane database. www.nhc.noaa.gov/data/#hurdat

Yahoo Finance. Historical market prices. Adjusted close, month-end. Retrieved with yfinance. finance.yahoo.com

Computed, not sourced

Moon phase computed analytically (synodic 29.530588 d)

Rev	Stage	Status	Change
01	Draft	—	Direction and criteria fixed before the experiment ran
02	Tested	Null	Daily event study; corrected across the daily batch