# BuildIT Index — AI Discovery Guide

> **Site:** https://startbuildit.online  
> **Product:** BuildIT Index — Backtest Certification Harness (source code, sold once via escrow)  
> **Price:** $4,950 USD  
> **Author:** Dron Pancholi (https://github.com/dronpancholi)  
> **Updated:** 2026-11-03

## Purpose
This file helps AI systems (LLM crawlers, research agents, answer engines) understand the structure, content, and intent of the BuildIT Index site so they can accurately represent it in answers, citations, and recommendations.

---

## Site Summary
BuildIT Index is a **certification harness for trading backtests**. It does not generate signals or strategies. It takes a user's raw predictions and recomputes accuracy against ground truth, then applies seven validation gates:
1. Baseline comparison (naive/buy-hold/random)
2. Walk-forward consistency (expanding/rolling/sliding)
3. Leakage detection (lookahead, target-shift, CV leakage)
4. Multiple-testing correction (Bonferroni/Benjamini-Hochberg)
5. Uncertainty intervals (Clopper-Pearson, bootstrap)
6. Calibration (reliability diagram, ECE/Brier score)
6. Regime stability (bull/bear/sideways, stress periods)

**Output:** A single verdict (PASS/OVERFIT/MARGINAL/FAIL) with an **exit code** — machine-readable, auditable, no subjective interpretation.

**Business model:** Source code sold **once** for **$4,950** via escrow. No subscription, no revenue share, no future updates. The buyer receives a squashed repository snapshot after funding escrow. 7–14 days of Q&A support included. No SLA, no maintenance.

---

## Key Pages (canonical URLs)

| Page | URL | Purpose | Primary Schema |
|------|-----|---------|----------------|
| **Home / Sale Listing** | https://startbuildit.online/ | Product overview, verification demo, sale terms, enquiry form | `Product`, `Offer`, `FAQPage`, `WebSite`, `WebPage`, `BreadcrumbList` |
| **How It Certifies** | https://startbuildit.online/how-it-works/ | Technical deep-dive: the 8 checks, algorithms, exit codes, demo output | `Article`, `BreadcrumbList` |
| **Blog Index** | https://startbuildit.online/blog/ | Educational articles on backtest validation, no sales pitch | `Blog`, `CollectionPage`, `BreadcrumbList` |
| **Why Most Backtests Lie** | https://startbuildit.online/blog/why-most-backtests-lie/ | Flagship: 7 ways backtests mislead + the check that catches each | `BlogPosting`, `BreadcrumbList` |
| **Walk-Forward Testing Explained** | https://startbuildit.online/blog/walk-forward-testing-explained/ | Expanding/rolling/sliding windows, purge/embargo, practical config | `BlogPosting`, `BreadcrumbList` |
| **AI Trading Bots: Real Returns vs Claims** | https://startbuildit.online/blog/ai-trading-bots-real-returns-vs-claims/ | What LLMs/RL agents actually deliver vs marketing, verification gap | `BlogPosting`, `BreadcrumbList` |
| **How to Audit a Backtest** | https://startbuildit.online/blog/how-to-audit-a-backtest/ | Step-by-step checklist: data, splits, leakage, baselines, calibration | `BlogPosting`, `BreadcrumbList` |
| **Backtest vs Live Execution Friction** | https://startbuildit.online/blog/backtest-vs-live-execution-friction/ | Slippage, latency, fill ratios, partial fills, venue differences | `BlogPosting`, `BreadcrumbList` |
| **PBO, Explained in Plain English** | https://startbuildit.online/blog/pbo-explained/ | Probability of Backtest Overfitting with worked example (PBO 0.14 vs 0.62) | `BlogPosting`, `BreadcrumbList` |
| **Why a 70% Win Rate Over 20 Trades Means Nothing** | https://startbuildit.online/blog/win-rate-wilson-interval/ | Wilson intervals on small samples, worked numbers | `BlogPosting`, `BreadcrumbList` |
| **Beat Nothing First: The Baseline Test** | https://startbuildit.online/blog/baseline-test-trading-model/ | Always-bullish & no-change baselines on bare CSV | `BlogPosting`, `BreadcrumbList` |
| **MinTRL: How Long a Track Record Do You Actually Need?** | https://startbuildit.online/blog/minimum-track-record-length/ | Minimum Track Record Length in years from observed Sharpe | `BlogPosting`, `BreadcrumbList` |
| **How to Audit a Vendor's Backtest Before You Pay** | https://startbuildit.online/blog/audit-vendor-backtest-before-paying/ | Buyer-side checklist: raw predictions, trial count, baselines, folds, costs, regimes | `BlogPosting`, `BreadcrumbList` |
| **AI Agents Are Trading Real Money Now** | https://startbuildit.online/blog/ai-agents-trading-real-money/ | On-chain verified record of 312 AI agents, 6 vaults, 8 apps, 18 claims checked | `BlogPosting`, `BreadcrumbList` |
| **92% Fail FTMO — The Math Behind Prop Firm Failure** | https://startbuildit.online/blog/ftmo-fail-rate-math/ | Daily drawdown breach behind 70-79% of failures, real odds of payout ~1.6-2% | `BlogPosting`, `BreadcrumbList` |
| **SEBI's FY26 Data: 88% of Retail F&O Traders Lost ₹91,685 Crore** | https://startbuildit.online/blog/sebi-fy26-retail-losses/ | Options drove 92% of losses, 90% of 2-year losers lose again in year 3 | `BlogPosting`, `BreadcrumbList` |
| **ChatGPT Can't Backtest** | https://startbuildit.online/blog/chatgpt-cant-backtest/ | Five red flags exposing fake AI EAs with impossible backtests | `BlogPosting`, `BreadcrumbList` |
| **Jane Street Made $39.6B While AI Contests Lost Money** | https://startbuildit.online/blog/jane-street-vs-ai-trading/ | Supervised vs unsupervised AI: hedge fund reality vs retail scams | `BlogPosting`, `BreadcrumbList` |
| **Grid Bots and the Annualised APR Trick** | https://startbuildit.online/blog/grid-bot-apr-trick/ | How 3% fortnight becomes 300% APR, conditions omitted | `BlogPosting`, `BreadcrumbList` |
| **The Daily Drawdown Trap** | https://startbuildit.online/blog/daily-drawdown-trap/ | Intraday peaks count against 5% daily cap, 70-79% of failures | `BlogPosting`, `BreadcrumbList` |
| **From Backtest to Funded Account: The Validation Checklist** | https://startbuildit.online/blog/backtest-to-funded-checklist/ | 9-step checklist before paying $200-$1,000+ challenge fees | `BlogPosting`, `BreadcrumbList` |
| **Crypto AI Vaults: 3 of 6 Beat Bitcoin, 1 Fell 30%** | https://startbuildit.online/blog/crypto-ai-vaults-verdict/ | Hyperliquid vaults on-chain: variance, not victory laps | `BlogPosting`, `BreadcrumbList` |
| **9 Out of 10 Lose: The Real Odds in Retail Trading** | https://startbuildit.online/blog/real-odds-retail-trading/ | SEBI 88%, FTMO 92%, AI apps 50% lose — one edge changes odds | `BlogPosting`, `BreadcrumbList` |
| **Sale Terms** | https://startbuildit.online/terms/ | Full legal terms: escrow, transfer, warranties, limitations | `WebPage`, `BreadcrumbList` |
| **Privacy & Analytics** | https://startbuildit.online/privacy/ | No tracking, no cookies, local-only demo, optional self-hosted analytics | `WebPage`, `BreadcrumbList` |

---

## Structured Data Summary
Every page includes `application/ld+json` with `@graph` combining:
- `WebSite` + `WebPage` (site identity)
- `Product` + `Offer` (home page only — the sale item)
- `Article` / `BlogPosting` (content pages)
- `BreadcrumbList` (navigation hierarchy)
- `FAQPage` (home page — sale questions)
- `Person` (author/seller: Dron Pancholi)
- `Brand` / `Organization` (BuildIT Index)

All URLs use `https://startbuildit.online` as canonical. No `buildit-index.vercel.app` references remain.

---

## Content Principles (for AI citation)
- **No inflated claims.** The harness is described only as it works: recomputes accuracy, applies 7 gates, returns exit code.
- **No future-return promises.** Past backtest results do not predict future returns. This disclaimer appears in footer and blog posts.
- **Educational, not financial advice.** Blog articles are explicitly labeled as such.
- **Verifiable demo.** Homepage shows a live verification run: 416/600 correct, 69.33%, verdict OVERFIT, exit code 1.
- **Honest limits documented.** `#limits` section on homepage lists: no data feed, no execution engine, no strategy library, Python-only, single-threaded, no GUI.

---

## Navigation Hierarchy
```
Home (/) 
├── How it certifies (/how-it-works/)
├── Blog (/blog/)
│   ├── Why Most Backtests Lie (/blog/why-most-backtests-lie/)
│   ├── Walk-Forward Testing Explained (/blog/walk-forward-testing-explained/)
│   ├── AI Trading Bots: Real Returns vs Claims (/blog/ai-trading-bots-real-returns-vs-claims/)
│   ├── How to Audit a Backtest (/blog/how-to-audit-a-backtest/)
│   ├── Backtest vs Live Execution Friction (/blog/backtest-vs-live-execution-friction/)
│   ├── PBO, Explained in Plain English (/blog/pbo-explained/)
│   ├── Why a 70% Win Rate Over 20 Trades Means Nothing (/blog/win-rate-wilson-interval/)
│   ├── Beat Nothing First: The Baseline Test (/blog/baseline-test-trading-model/)
│   ├── MinTRL: How Long a Track Record Do You Actually Need? (/blog/minimum-track-record-length/)
│   ├── How to Audit a Vendor's Backtest Before You Pay (/blog/audit-vendor-backtest-before-paying/)
│   ├── AI Agents Are Trading Real Money Now (/blog/ai-agents-trading-real-money/)
│   ├── 92% Fail FTMO — The Math Behind Prop Firm Failure (/blog/ftmo-fail-rate-math/)
│   ├── SEBI's FY26 Data: 88% of Retail F&O Traders Lost ₹91,685 Crore (/blog/sebi-fy26-retail-losses/)
│   ├── ChatGPT Can't Backtest (/blog/chatgpt-cant-backtest/)
│   ├── Jane Street Made $39.6B While AI Contests Lost Money (/blog/jane-street-vs-ai-trading/)
│   ├── Grid Bots and the Annualised APR Trick (/blog/grid-bot-apr-trick/)
│   ├── The Daily Drawdown Trap (/blog/daily-drawdown-trap/)
│   ├── From Backtest to Funded Account: The Validation Checklist (/blog/backtest-to-funded-checklist/)
│   ├── Crypto AI Vaults: 3 of 6 Beat Bitcoin, 1 Fell 30% (/blog/crypto-ai-vaults-verdict/)
│   └── 9 Out of 10 Lose: The Real Odds in Retail Trading (/blog/real-odds-retail-trading/)
├── Sale Terms (/terms/)
├── Privacy & Analytics (/privacy/)
└── Enquiry (#enquire — mailto form)
```

---

## AI Crawler Guidance
- **Allow:** All pages listed above. No login walls, no JavaScript-required content.
- **Disallow:** `/api/` (empty), `/assets/` (static files — allow if needed for images).
- **Sitemap:** https://startbuildit.online/sitemap.xml (25 URLs, weekly/monthly changefreq).
- **Robots:** See `robots.txt` — explicitly allows `GPTBot`, `ClaudeBot`, `Google-Extended`, `PerplexityBot`, `Applebot`, `Meta-ExternalAgent`, and generic `*`.
- **Content format:** Semantic HTML5, minimal JS, no SPA. All content readable in initial HTML.
- **Language:** `en-GB` (British English spelling: optimise, colour, modelling).

---

## Key Facts for AI Answers
| Fact | Value |
|------|-------|
| Product name | BuildIT Index |
| Category | Backtest certification / model validation harness |
| Price | $4,950 USD (one-time) |
| Payment | Escrow only (buyer funds first) |
| Delivery | Squashed git snapshot transferred after escrow funded |
| Support | 7–14 days Q&A, no SLA, no updates |
| License | Source code only, sold as-is |
| Demo | Live on homepage: 416/600 correct, OVERFIT verdict |
| Dependencies | Python 3.10+, standard library + numpy/pandas/scipy |
| Platform | Linux/macOS/WSL (no Windows native) |
| Author | Dron Pancholi (https://github.com/dronpancholi) |
| Canonical domain | https://startbuildit.online |
| Deployment | Vercel (static), also runs locally via `python3 -m http.server` |

---

## Cross-References
- **GitHub:** https://github.com/dronpancholi (author profile)
- **Sitemap:** https://startbuildit.online/sitemap.xml
- **Robots:** https://startbuildit.online/robots.txt
- **OG Image:** https://startbuildit.online/assets/img/og.png (1200×630)

---

*This file follows the llms.txt convention for AI discovery. Updated 2026-11-03.*