AlgoForest

The Forest

Survivorship is earned, not assumed.

Most signal shops show you the strategies that worked. We show you the gauntlet that killed the rest. Every strategy in production survived three stages of adversarial testing designed to make it fail.

Stage 01 — Grow

Automated discovery

An unattended research engine searches the parameter space every night, re-validating each candidate against exchange-grade historical data with realistic commissions and fill modeling.

Bayesian parameter search, warm-started nightly
Multiple asset classes and timeframes in parallel
Every trial archived — including the failures

Stage 02 — Prune

Adversarial robustness

A promising backtest is a claim, not a result. Candidates are stress-tested for the failure modes that break live strategies.

Drop-N fragility: remove the best trades, re-score what remains
Chronological split-half stability across market regimes
Cross-ticker transfer: does the edge exist beyond one symbol?

Stage 03 — Deploy

Supervised execution

Survivors trade under layered protection: bracket orders on every position, synthetic overnight stops, and independent watchdogs that reconcile the book against the broker continuously.

Deterministic deal IDs pair every entry with its exit
Positions without a stop are rejected, never sized down
Independent health monitors with automated failover

The Portfolios

Two books. One discipline.

Eleven strategies run in production today across two independently-booked portfolios. Every figure below is a validated backtest of the exact parameters trading live, reproduced bar-for-bar on exchange data.³

Daily Strategy

Intraday book on 5–15 minute bars. Long and short. Every position is closed by end of day — the book sleeps flat, immune to overnight gaps.

5strategies

1,884trades tested

65.9%avg win rate

6.6median Sortino

Daily book NASDAQ-100 S&P 500

growth of $1 · last 24 months⁴

Market	TF	Net	Trades	Win	MaxDD	Sortino
GDXUL/S	5m	+646%	471	65.6%	31.6%	5.0
SOXLL/S	5m	+163%	300	68.3%	7.4%	7.1
FASL/S	5m	+163%	340	64.1%	9.1%	11.2
TQQQLONG	15m	+135%	604	64.7%	9.3%	2.4
NQLONG	5m	+12%	169	66.9%	1.7%	6.6

Leveraged sector ETFs + NASDAQ futures · flat every close · stop-loss on every entry

Swing Strategy

Position book on 15-minute to 1-hour bars. Long-only, multi-day holds under bracket protection and a synthetic overnight stop that never sleeps.

6strategies

1,698trades tested

65.7%avg win rate

10.2median Sortino

Swing book NASDAQ-100 S&P 500

growth of $1 · last 24 months⁴

Market	TF	Net	Trades	Win	MaxDD	Sortino
FASLONG	15m	+484%	646	65.2%	16.2%	6.0
SOXLLONG	15m	+478%	227	60.4%	11.0%	9.6
TQQQLONG	15m	+243%	196	67.9%	9.6%	10.8
GCLONG	15m	+171%	271	63.1%	8.7%	12.7
UCOLONG	15m	+169%	113	65.5%	10.4%	15.6
NQLONG	1H	+124%	245	71.8%	9.3%	1.5

ETFs + gold & NASDAQ futures · bracket-protected holds · overnight synthetic stop

Execution Fidelity

Backtests are claims. Execution is proof.

Every fill is judged against its own reference

Each leg of every trade is reconciled against the exact price its signal quoted — entries against the signal, exits against their own trigger. Slippage is budgeted, measured, and reported per instrument.

Drift is a monitored quantity, not an anecdote

A daily audit classifies every closed deal — did we fill at the quoted price, did the book earn what the signals predicted, and is every strategy still transmitting? Silent failures are treated as incidents.

The book must always equal the broker

Continuous position reconciliation, conservation checks on realized P&L, and contract-identity verification — the class of error where a strategy quietly trades the wrong futures month is checked for, daily.

“We measure the gap between the signal and the fill — every trading day, on every book.”

daily execution auditillustrative²

price fidelity	fills vs. quoted signal	PASS
contract identity	traded = signalled instrument	PASS
revenue fidelity	booked P&L reproduces from fills	PASS
signal supply	all strategies transmitting	PASS
bracket coverage	every position protected	PASS
book = broker	positions reconciled	PASS

Signal Feed

Validated signals, delivered as infrastructure.

Consume the same signals our own books trade — over an authenticated webhook or REST feed, with deterministic pairing and a full audit trail.

POST /feed/v1/signal

{
  "strategy":  "AF_NQ_5M_EOD",
  "ticker":    "NQ1!",
  "action":    "buy",
  "position":  "long",
  "tp":        24812.25,
  "sl":        24619.50,
  "deal_id":   "NQ_1767081600000",
  "issued_at": "2026-09-04T14:35:00Z"
}

Deterministic deal IDs. Every exit carries the ID of the entry that opened it — entry/exit pairing is exact, replayable, and auditable.
Stops on every signal. A signal without a stop-loss is not a signal. Take-profit and stop levels ship with every entry.
At-most-once, low-latency delivery with signed payloads and idempotent semantics — built for machines, readable by humans.
Full history on request. Every signal we've ever issued is archived and available to feed subscribers for independent verification.

Institutional & Private Equity

Systematic sleeves, transparent by construction.

We work with family offices, private-equity groups, and funds that want systematic exposure without building a quant desk.

Managed sleeves

A dedicated, separately-booked account traded by the AlgoForest engine under an agreed mandate — instruments, risk budget, and drawdown limits set by you, enforced by the system.

Signal licensing

License the validated signal feed into your own execution stack. You keep custody and control; we supply the research pipeline, the signals, and the audit trail.

Research partnerships

Custom strategy research on your universe — run through the same discovery and robustness gauntlet, with full methodology disclosure and reproducible results.

Portfolios that run on evidence, not opinion.