---
title: "Discussion on \"We tested 1,000 random trading rules on gold: how many look great?\" | Hashnode"
description: "Discussion on \"We tested 1,000 random trading rules on gold: how many look great?\". Before costs 40 looked significant; after spreads, 4 — none beat holding gold."
url: "https://hashnode.com/posts/we-tested-1-000-random-trading-rules-on-gold-how-many-look-great/6ade0424ba9692b180a8df15"
---

[Feed](https://hashnode.com/feed) /

Discussion

# [We tested 1,000 random trading rules on gold: how many look great?](https://forwardproof.hashnode.dev/we-tested-1-000-random-trading-rules-on-gold-how-many-look-great?utm_source=hashnode&utm_medium=feed)

Before costs 40 looked significant; after spreads, 4 — none beat holding gold.

Oct 25 · 7 min read · forwardproof.hashnode.dev  
Read the article

-   [#trading](https://hashnode.com/tag/trading)
-   [#backtesting](https://hashnode.com/tag/backtesting)
-   [#gold](https://hashnode.com/tag/gold)
-   [#quantitative-finance](https://hashnode.com/tag/quantitative-finance)
-   [#overfitting](https://hashnode.com/tag/overfitting)
-   [#xauusd](https://hashnode.com/tag/xauusd)
-   [#algorithmic-trading](https://hashnode.com/tag/algorithmic-trading)

## Written by

[ForwardProof](https://hashnode.com/@forwardproof)

@forwardproof

  
[View profile](https://hashnode.com/@forwardproof)

Transparent AI gold research & honest backtest validation

## Responses 1

Markdown supported · ⌘↵ to post

[Arnold Holm](https://hashnode.com/@stratcorealpha)

[Oct 25](https://hashnode.com/posts/we-tested-1-000-random-trading-rules-on-gold-how-many-look-great/6ade0424ba9692b180a8df15/comment/6ade112aba9692b180a91ebe)

For a reversal from +1 to -1, does the cost calculation multiply by abs(new\_position - old\_position), so it charges two units rather than one? I would add three small fixtures: flat to long, long to flat, and long to short, each with a fixed bid/ask and zero slippage. The reversal should cost exactly the same as closing the long and opening the short at that quote. Publishing that check alongside the script would make the position-change cost convention much easier to audit.
