tradingarbor — Navigate Markets. Make Informed Decisions.

What it does

Three engines run in parallel; you choose which to enable.

Options — long calls / long puts

90% allocation. Breakout + pullback scanner on equities, expressed as 30-DTE 0.70-delta options. Regime-Kelly sizing, swing-low stops, partial-take at 2R/3R/4R, eager pyramid into winners.

Crypto — trend-follow

10% allocation, 24/7. 20-day breakouts with SMA 20/50 alignment, BTC SMA-100 bear filter, 8% hard stop, 30-day time stop. Pyramid + vol-target + conviction tiers.

How it works

One dashboard URL. Your data on your own dedicated VPS.

1. Sign in once

Every customer signs in at dashboard.tradingarbor.com. One URL to bookmark, one URL to share when troubleshooting, one URL on your password manager.

2. Routed to your VPS

Behind the scenes, the dashboard transparently routes you to your own dedicated DigitalOcean droplet — the trading engine, Postgres, encryption keys, and broker creds all live there. No shared infrastructure, no shared keys, no shared blast radius.

3. Direct access if you need it

Your instance is also reachable directly at its per-customer subdomain for admin / debug / router-bypass scenarios. The unified URL is the default; the direct URL is the fallback.

Backtest baseline

5-year walk-forward on pinned Alpaca bars. Deterministic, reproducible.

Variant	Sharpe	Avg/yr	Worst yr
Portfolio blend (90/10)	~1.26	~+182%	+10%
Swing (current shipped)	~1.10	~+195%	~+5%
Crypto trend (with stack)	—	~+79%	-14% (2022)

Honest live expectation. Live trading typically compresses to 50-70% of backtest figures. The dashboard's Performance Expectations card surfaces both raw and theta-adjusted projections (avg × 0.5, worst × 1.5, sharpe × 0.7). Past performance is not indicative of future results. You may lose money.

How we verify the numbers

A 4-test battery on every shipped change.

1. Parameter stability

Knob ±20% sweep — neighbors must be within 6% of peak. Rejects knife-edge optimums.

2. Intra-year drawdown

Higher returns can't come from looser risk controls. Profile must not be worse than baseline.

3. Leave-one-year-out

No single year drives the aggregate Sharpe. Rejects regime-luck strategies.

4. Trade attribution

Top-10% trade share must be better than baseline. Rejects fragile "needs the home runs" strategies.

Security posture

Broker credentials live here — security is treated as load-bearing.

Per-tenant envelope encryption (AES-256-GCM data keys wrapped by master KEK)
TOTP MFA required for admin actions; step-up re-auth on every sensitive write
PBKDF2-HMAC-SHA256 @ 600k rounds for password storage
Per-IP rate limiting (5 fails / 15 min) on login AND MFA verify
CSRF defense-in-depth: SameSite=Strict + synchronizer tokens on state-changing requests
Open-redirect protection, security headers (CSP, HSTS, Permissions-Policy), sanitized error responses
Hard order bounds — pre-broker validation rejects out-of-spec orders before reaching Alpaca
Idempotent order submission via client_order_id — network retries can't double-fill
Risk-disclosure click-through gate; pre-trade pause until accepted
Per-tenant kill switches with conservative fallback
Audit log of every login, credential change, role change, paper/live flip
Container runs as non-root uid 10001; CI hardened with default-deny permissions and timeouts

Operations

Designed to be left alone.

~5 min/ day operator time

Glance at the daily summary email + dashboard. Engines self-heal dead threads; certs renew automatically; backtest snapshots refresh on launcher boot.

~$6/ month infra cost

1 vCPU, 2 GB RAM is plenty. Hetzner CX22 or DigitalOcean basic droplet. Domain registration optional but recommended for Let's Encrypt TLS.

30 minfirst-deploy time

One docker compose --profile prod up brings up engines, dashboard, Caddy reverse proxy, optional Postgres for multi-tenancy. Full runbook in DEPLOY.md.

What's NOT being claimed

That this beats the market in all conditions
That backtest = live (it doesn't; expect 50-70% compression)
That any prior result is a forecast of the next year
That it's a substitute for your own judgment about strategy fit
That it's a registered investment adviser unless explicitly stated in the dashboard's About section

Pricing

Three tiers. Every plan includes a dedicated VPS with your own encryption keys, broker creds, and isolated trading state — accessed through one unified dashboard.tradingarbor.com URL.

Starter

__PRICE_STARTER__/mo

For paper-validating the engines on a small live account.

Dedicated 1 vCPU / 1 GB VPS
All three engines (options + crypto + swing)
Five brokers supported
MFA + per-tenant encryption
Email + community support

Subscribe →

RECOMMENDED

Pro

__PRICE_PRO__/mo

Faster droplet, headroom for higher-frequency engines + future features.

Dedicated 1 vCPU / 2 GB VPS
Everything in Starter
Daily backup retention (7 days)
Priority email support (under 24h)
Beta access to new engines

Subscribe →

Elite

__PRICE_ELITE__/mo

For larger accounts running all three engines plus extras.

Dedicated 2 vCPU / 4 GB VPS
Everything in Pro
Extended backup retention (30 days)
Direct support channel
Quarterly strategy review

Subscribe →

Cancel anytime via the Sign in billing portal. Your droplet, broker keys, and trading state are deleted within 7 days of cancellation. Trading involves substantial risk of loss; see the legal notice at the bottom of this page before subscribing.

Before you subscribe

This is an automated trading platform — make sure these are true:

You have an account with one of the supported brokers (Alpaca, Schwab, Tradestation, Tastytrade, Tradier — paper at minimum)
You can paper-trade for 2-3 weeks before going live to confirm the engines fit your account size and risk tolerance
You're comfortable with the capital you're allocating being treated as risk capital
You understand that backtest figures typically compress 50-70% in live trading and past performance does not predict future results (see the legal notice at the bottom of this page)

Have questions before subscribing? Email hello@tradingarbor.com.