The Full Strategy Lab Scoreboard: All 36 Strategies, Ranked After Real Costs

Thirty-six strategies, one engine, one year. This is the scoreboard the whole series has been building toward — every rule from posts 01–36 in one ranked table, net of the same cost model (taker 0.05%/leg + hourly funding). The rank column decides. Strategy Lab #44.

The scoreboard

Buy & hold benchmark: +154.94% (maxDD -21.7%). Note posts 11 and 17 run on AAPL daily 2018–2025, not BTC hourly.

What the ranking says

Top of the table is all slow, all trend-shaped. Linear regression aside (a different dataset and a different lookback philosophy), the top six net performers are the six slowest entries in the lab, and five of them are crossovers. EMA envelope's +80.99% on 22 trades is the single best risk-adjusted entry; TEMA's +83.49% on 113 trades is the gross winner among BTC rules.

Middle of the table is the honest majority. 26 of 36 finished net positive. The 141–600 trade band is a coin flip — MACD, CCI, RSI(2) make it; Ichimoku and Aroon barely do; and every one of them loses at 25bp.

Bottom of the table is the fee machine. Stochastic (-71.31%), k-NN (-78.53%), and VWAP (-37.36%) turned triple-digit gross years into losses on 1,000+ trades. CMF's -0.00% on 557 trades is the most honest number in the lab.

The two-lever summary: (1) Trade less — the ≤140-trade group was 18/18 positive; (2) ride the trend in a trend year — the trend family took 7 of the top 10 net slots. On 2023 BTC, that's the entire playbook, and it survived every scenario I threw at it.

Code

import pandas as pd
from backtest_base import backtest_signal, metrics
from strategy import STRATEGIES

df = pd.read_pickle("assets/data/binance_BTCUSDT_1h_2023-01-01_2023-12-31.pkl")
rows = []
for name, spec in STRATEGIES.items():
    signal = spec["func"](df, **spec["params"])
    res = backtest_signal(df, signal, 0.0005, 1.25e-05)
    m = metrics(res, 8760)
    rows.append((m["total_return"], name, m["trades"]))

rows.sort(reverse=True)
for ret, name, trades in rows:
    print(f"{name:<16} {ret:+9.2%}  trades {trades}")

Reproduce it

cd blog-drafts/scripts
python gen_special_assets.py   # regenerates this chart + all group tables
python backtest_base.py --batch --out output --batch-plot output/batch.png

Data: Binance public API (BTCUSDT 1h, 8,735 bars) and Yahoo Finance (AAPL daily, posts 11/17). Every number above reproduces from assets/posts/{NN}/post.json.

This is a backtest on historical data, not investment advice. Past performance does not predict future results.