Backtesting

What it does

When your agent uses it

• Tasks that involve Trading and backtesting
• Tasks that involve Event-driven systems

Example prompts

• “/backtesting”

Requirements

• Python 3

What it can do on your machine

Read from SKILL.md and the folder at commit 5156f81. It shows what the files ask for, not the result of running them.

• Tool permissions

  Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

  From allowed-tools in the SKILL.md frontmatter.

• Runs code

  Ships 1 file in scripts/, which the agent can run.

  From the folder's file list and the shell code blocks in SKILL.md.

• Network

  No URLs in SKILL.md.

  From URLs in SKILL.md, links to its own repository left out.

• Credentials

  Names no API keys, tokens, secrets or passwords.

  From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Backtesting loads about 2.4k tokens when it runs, and up to ~13k if it reads all its reference files. Until then it costs about 87 tokens; SKILL.md has 412 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it

~87

When it runs · the whole SKILL.md, loaded when a task matches

~2.4k

With references · SKILL.md plus every file in references/, read only if the agent opens them

~13k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); the scripts in this folder are not scanned.

SKILL.md

The full file from gauss314/skills at commit 5156f81, republished under its MIT licence (© gauss314). 412 words, ~2,357 tokens.

name

backtesting

description

Academic backtesting framework for quantitative research. ~30 risk and performance ratios, 10 classes of indicators, event-driven engine with 6+ strategies, MPT optimizer, forward-looking simulation with Johnson SU + t-Copula, walk-forward CV, stress testing, fundamental analysis (Altman Z, Piotroski, DuPont). All flat Python + numpy.

license

MIT

skills/backtesting/
├── SKILL.md                         ← This file
├── references/
│   ├── BACKTESTING_THEORY.md        ← Marco conceptual: GIGO, trilema, 5 etapas (ES)
│   ├── RATIOS.md                    ← Fórmulas, convención de retornos, advertencias (ES)
│   ├── FEATURES.md                  ← Taxonomía de 10 clases de indicadores con edges (ES)
│   ├── SIMULATIONS.md               ← Pipeline Johnson SU + cópula (ES)
│   ├── VALIDATION.md                ← Suite de validación de 4 niveles (ES)
│   └── OTHER_FEATURES.md            ← Fundamental, Sentimiento, Exógenos (ES)
├── assets/
│   ├── sp500_returns.csv            ← SPY benchmark daily returns (lin + log), 1980-today
│   ├── momentum_sma50_200_returns.csv ← SMA(50)/SMA(200) crossover strategy returns
│   ├── contrarian_bbands_returns.csv  ← Bollinger Band contrarian strategy returns
│   ├── sample_portfolios.json       ← Real investor portfolios (Buffett, Dalio, Ackman, 60/40)
│   ├── defaults.json                ← Default parameters (VaR alpha, windows, etc.)
│   └── validation_cases.json        ← 6 known cases for ratio validation
├── scripts/
│   ├── __init__.py
│   ├── ratios.py                    ← 30+ flat numpy functions for all risk/performance ratios
│   ├── indicators.py                ← 10 classes of technical/statistical/fundamental indicators
│   ├── engine.py                    ← Event-driven BacktestEngine with 8 built-in strategies
│   ├── backtesting.py               ← CLI: run, sweep, walkforward, montecarlo, optmpt, event, validate
│   ├── simulations.py               ← CLI: marginal, copula, run, portfolio, scenarios
│   ├── forward.py                   ← CLI: project, risk, stress, summary
│   ├── distributions.py             ← Fit + KS test for Normal/t/NCt/Laplace/JohnsonSU
│   ├── copulas.py                   ← t/Gaussian/Clayton/Gumbel/Frank copulas + sampling
│   ├── fundamental_ratios.py        ← Income/balance/cashflow metrics, DuPont, Altman Z, Piotroski
│   └── validate.py                  ← 4-level validation: CLI modes, math consistency, edge cases, regression
└── tests/
    └── test_ratios.py               ← 18 pytest tests for core ratios

# Compute all 30+ ratios on a CSV of prices
py scripts/backtesting.py run --prices assets/sp500_returns.csv

# Compute with benchmark comparison
py scripts/backtesting.py run --prices assets/momentum_sma50_200_returns.csv --benchmark assets/sp500_returns.csv

# Full validation (33 checks across 4 levels)
py scripts/validate.py

# Single level
py scripts/validate.py --nivel 1

# Load a CSV with OHLCV data and run SMA crossover
py scripts/backtesting.py event --data my_stock.csv --strategy sma_crossover --fast 50 --slow 200 --commission 0.001

# 2D sweep over fast/slow MA windows
py scripts/backtesting.py sweep --prices assets/sp500_returns.csv --p1-min 10 --p1-max 100 --p1-step 10
# 2D over 2 parameters
py scripts/backtesting.py sweep --prices assets/sp500_returns.csv --p1-min 10 --p1-max 50 --p1-step 5 --p2-min 25 --p2-max 200 --p2-step 25

py scripts/backtesting.py walkforward --prices assets/sp500_returns.csv --splits 5 --gap 21

py scripts/backtesting.py optmpt --assets assets/sp500_returns.csv --iterations 5000

py scripts/simulations.py marginal --returns assets/sp500_returns.csv
py scripts/simulations.py copula --returns assets/sp500_returns.csv --df 4
py scripts/forward.py project --returns assets/sp500_returns.csv --horizon 252 --paths 10000 --drift 0.08
py scripts/forward.py risk --returns assets/sp500_returns.csv --horizon 252 --paths 10000

py scripts/simulations.py portfolio --name warren_buffett
py scripts/simulations.py scenarios --name warren_buffett --cagr -0.3,-0.15,0,0.2,0.35,0.5

from scripts.ratios import *

prices = np.array([100, 105, 102, 110, 108, 115])
r = linear_returns(prices)    # [0.05, -0.0286, 0.0784, -0.0182, 0.0648]
lr = log_returns(prices)      # [0.0488, -0.0290, 0.0755, -0.0183, 0.0628]

sharpe_ratio(r)               # 0.847
max_drawdown(prices)          # -0.0370
kelly_fraction(lr)            # 0.0793
var_all(r, alpha=0.05)        # {'empirical': ..., 'normal': ..., 'johnsonsu': ...}
profit_factor(lr)             # 2.314
payoff_ratio(lr)              # 1.578
rachev_c(lr, alpha=0.05)      # 1.234
common_sense_ratio(lr)        # 2.856

from scripts.engine import BacktestEngine

eng = BacktestEngine(initial_capital=1.0, commission=0.001, slippage=0.0005)
eng.load_data(df_ohlcv)
result = eng.run(strategy='sma_crossover', strategy_params={'fast': 50, 'slow': 200})

print(result['metrics']['sharpe_ratio'])     # 0.847
print(result['trades'])
print(result['metrics'])

© gauss314, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 33 other files (scripts, references, assets) in skills/backtesting of gauss314/skills.

• SKILL.md
• assets/contrarian_bbands_returns.csv
• assets/defaults.json
• assets/frontier.csv
• assets/momentum_sma50_200_returns.csv
• assets/multi_asset_prices.csv
• assets/multi_ma_3_5_8_13_returns.csv
• assets/sample_portfolios.json
• assets/sp500_close.csv
• assets/sp500_prices.csv
• assets/sp500_returns.csv
• assets/trend_bbands_combo_returns.csv
• assets/validation_cases.json
• references/BACKTESTING_THEORY.md
• references/FEATURES.md
• references/OTHER_FEATURES.md
• references/RATIOS.md
• references/SIMULATIONS.md
• references/VALIDATION.md
• … and 15 more

Open the folder on GitHubat commit 5156f81