algorithmic-trading-anthology-july-2026-random-sign-control

Conclusion: Control

A benchmark, not a candidate. Held continuously it returned 0.3% a year at 12.1% volatility, with a worst fall of 68.5% over 558 months. Strategies are measured against this.

Figure 1 · Growth of $1

1980–2026 · net of 15bp round-trip costs

What one dollar became, after costs: $1.15 by 2026. Shown against S&P 500, rebased to the same starting dollar.

Return for risk

0.03

range (0.23) to 0.28

Ann. return

0.3%

volatility 12.1%

Max drawdown

68.5%

over 558 months

Beats the search

0.0%

chance, after correction

Method

Hypothesis

Calibration. Should score zero.

Rule

UNIVERSE = the high-conviction set

function positions(bar):

# signs drawn once from a fixed seed and never revisited

for symbol in UNIVERSE:

w[symbol] = coin_flip(seed, symbol) / vol_36m(symbol, bar.prev)

size = clip(zscore(feature(soi_3m), expanding) / 2, -1, 1)

return w * size

# ── execution ────────────────────────────────────────────────────────

COST_PER_TRADE = 15 bps of the notional that changes hands

function on_bar(bar, book):

if not bar.is_month_end:

return # decisions are made monthly only

target = positions(bar) # the rule above

target = scale_to_gross(target, 1.0)

rebalance(book, target, bar)

function scale_to_gross(w, limit):

gross = sum(abs(w)) # total capital at work

if gross == 0:

return w # flat is a valid target

return w * limit / gross # leverage fixed, never implicit

function rebalance(book, target, bar):

for symbol in union(book.symbols, target.symbols):

delta = target[symbol] - book.weight(symbol)

if delta > 0:

buy(symbol, delta, fill = next_bar.close)

else if delta < 0:

sell(symbol, -delta, fill = next_bar.close)

book.charge(COST_PER_TRADE * abs(delta))

# Orders are filled at the next month's close, never the one the decision

# was made on. There is no stop_loss(), take_profit(), limit order or

# position cap: a holding changes only when positions() returns a

# different target at the next month end.

Every claim in this anthology depends on the evaluation harness being capable of returning nothing. This strategy runs the identical construction — same universe, same inverse-volatility weighting, same costs, same execution lag — with exposure signs drawn once at random.

It returns 0.025. That is what licenses reading any other number here.

Evaluation

Table 1 · Performance

1980-01 – 2026-06 · 558 months, 558 invested

Measure	Value
Annualised return	0.3%
Annualised volatility	12.1%
Return for risk taken	0.03
95% range	(0.23) to 0.28
t-statistic	0.17
Maximum drawdown	68.5%
Hit rate	50.2%
Annual turnover	3.60×
Return skew	0.27
Return kurtosis	6.67

+ 6 more rows− fewer rows

The range is measured by resampling the history in blocks, so runs of good and bad months stay intact.

Table 2 · Robustness

Diagnostic	Value	Reads as	Bad	Okay	Yay
Design half (pre-2015)	0.04	return for risk	≤ 0	0 – 0.40	≥ 0.40
Holdout half (2015+)	(0.11)	return for risk	≤ 0	0 – 0.40	≥ 0.40
Top-5-month share of profit	506.1%	how much rode on a few months — lower is better	≥ 40%	25 – 40%	< 25%
Top-2-year share of profit	311.4%	concentration	—	—	—
Distinct trading episodes	1	how many independent runs this really is	< 50	50 – 100	≥ 100
Chance the edge is real	56.7%	non-normality corrected	—	—	—
Chance it beats the whole search	0.0%	chance of beating the whole search by luck	—	—	—

bad/ok boundary is the pre-registered criterion · ok/good is a margin above it, not a criterion · blank where none was registered

+ 3 more rows− fewer rows

The last row asks whether the result would stand out from the hundred-odd ideas tried here, rather than being judged alone.

Table 3 · Net return by calendar year

after costs

Year	Net return
1980	1.9%
1981	(3.4%)
1982	(5.8%)
1983	6.6%
1984	(15.5%)
1985	(20.1%)
1986	6.6%
1987	18.4%
1988	24.8%
1989	0.6%
1990	13.6%
1991	18.5%
1992	(8.2%)
1993	18.1%
1994	(15.9%)
1995	(3.0%)
1996	(1.9%)
1997	(4.2%)
1998	(14.8%)
1999	(7.2%)
2000	(16.3%)
2001	13.4%
2002	2.3%
2003	(5.6%)
2004	13.4%
2005	1.6%
2006	0.9%
2007	0.1%
2008	8.1%
2009	1.9%
2010	(4.9%)
2011	(0.7%)
2012	2.1%
2013	0.3%
2014	(7.1%)
2015	(5.8%)
2016	2.2%
2017	(1.4%)
2018	(0.3%)
2019	(0.3%)
2020	(2.9%)
2021	(1.2%)
2022	(1.0%)
2023	4.3%
2024	(0.0%)
2025	(1.1%)
2026	2.3%

+ 43 more rows− fewer rows

Table 4 · Concentration detail

Measure	Value
Best years	1988:+25%, 1991:+19%, 1987:+18%
Worst year	1985:-20%

Reference

Instruments

last price and one-month change, live from Yahoo Finance

Symbol	Notes
GNC.AX	GrainCorp — listed on ASX, in AUD.
ELD.AX	Elders — listed on ASX, in AUD.
NUF.AX	Nufarm — listed on ASX, in AUD.
AAC.AX	Australian Agricultural Company — listed on ASX, in AUD.
BALRAMCHIN.NS	Not in the instrument table.
TRIVENI.NS	Triveni Engineering & Industries — listed on NSE, in INR.
DHAMPURSUG.NS	Not in the instrument table.
COROMANDEL.NS	Not in the instrument table.
CHAMBLFERT.NS	Not in the instrument table.
M&M.NS	Mahindra & Mahindra — listed on NSE, in INR.
ESCORTS.NS	Escorts Kubota — listed on NSE, in INR.
ZW=F	Chicago SRW Wheat — continuous futures on CBOT, quoted in US cents.
ZM=F	Soybean Meal — continuous futures on CBOT, in USD.
ZR=F	Rough Rice — continuous futures on CBOT, in USD.

list truncated in the source record

Sample

1980-01 – 2026-06 · 558 months

Sources

NOAA Physical Sciences Laboratory. Climate indices. ONI, Niño 3.4, MEI v2, DMI, PDO, AMO, TNA, QBO, solar flux. psl.noaa.gov/data/climateindices/list

Queensland Department of Agriculture and Fisheries. Southern Oscillation Index. The Long Paddock. Monthly, 1876–. www.longpaddock.qld.gov.au/soi/soi-data-files

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

Revision history

Rev	Stage	Status	Change
01	Draft	—	Direction and criteria fixed before the experiment ran
02	Tested	Control	Benchmark, not scored against the criteria

Random-sign control