Step 11: Summary

This page consolidates the main research outcome into a compact executive view: key counts, top candidates, best-by-category tables, benchmark checks, warnings, and step-by-step findings.

InputArtifacts from every previous step: optimization, SL/TP, robustness, regimes, scoring, benchmark, leverage, slippage sensitivity, and walk-forward diagnostics.
OutputPipeline statistics, top final candidates, category winners, benchmark reality checks, warnings, and narrative conclusions.
Used laterUse this page as the first review page and as the final checklist before deeper inspection of individual steps.

Research Conclusion

Best composite score is 0.705 (strong). Best robustness score is 0.767. Best regime score is 0.678. 2 automatic warning condition(s) require review.

Pipeline Statistics

Description
A simple EMA crossover strategy that generates buy signals when a short-term EMA crosses above a long-term EMA. The strategy is designed to capture trends in the market. It can be applied to various timeframes and assets, and its performance can be sensitive to the choice of EMA periods.
Strategy
EmaCrossStrategy
Trend following
Symbol / Timeframe
BTC/USDT / 4h
longonly
Baseline Candidates
9
Step 01
Grid Candidates
1520
Step 02
Metric Winners
15
Step 02
SL/TP Candidates
277
Step 03
Robustness Windows
9
Step 04
Shortlist
36
Step 04
Regimes / Intervals
5 / 15
Step 05
Scored Candidates
36
Step 06
Best Strategy Composite Score
0.705
Strong
Leverage Grid Safe
17 / 36
Step 08
Slippage Survivors
30 / 36
Step 09
WFA Windows
12
Step 10
Best WFA Adaptability
0.497
Step 10

Top Final Candidates

Candidate IdBuy Ema ShortBuy Ema LongFull Profit TotaliFull Max DrawdowniFull TradesiFull Win RateiMax Safe LeverageiFirst Unsafe LeverageiBest Grid Leverage ScoreiLeverage At Best Grid ScoreiReference Leverage ScoreiSLTPShortlist ReasonStrategy Composite ScoreiRobustness ScoreiRegime ScoreiParameter Stability ScoreiBenchmark Profit TotaliAlpha vs BenchmarkiStatus At Configured LeverageiReference Leverage StatusiRolling Robustness ScoreiOOS Robustness Scorei

Best By Category

Candidate IdCategoryMetricValue

Best By Category (Base Candidates Only)

Same category winners as above, but SL/TP-modified candidates are excluded. This keeps the base strategy parameter sets visible separately from SL/TP enhancements.

Candidate IdCategoryMetricValue

Key Warnings

SeverityCheckDetail

Benchmark Winners

Candidate IdStrategy Full ProfitBenchmark ProfitItemAlpha vs BenchmarkiBeta to BenchmarkiCorrelation to BenchmarkiDownside Capturei

Benchmark Positive Alpha Candidates

Candidate IdBuy Ema ShortBuy Ema LongSLTPCandidate TypeiTradesiProfit TotaliWin rateiBenchmark Profit TotaliAlpha vs BenchmarkiProfit DrawdowniMax DrawdowniDownside Capturei

Risk-Adjusted Benchmark Comparison

Candidates that trail buy-and-hold by profit but improve drawdown versus the benchmark. These may still be useful when the goal is smoother risk-adjusted exposure rather than maximum raw return.

No data found.

Step Findings

Step 01

Baseline Backtesting

9 static parameter sets backtested on the full period as baseline reference. This establishes a starting point before grid search and lets you verify the strategy logic works as expected.

Step 02

Parameter Optimization

Full parameter grid of 1,520 candidates evaluated. 15 metric winners extracted across different optimization objectives. Best efficiency score in the grid: 0.599 (cand_8_167).

Step 03

SL/TP Optimization

277 SL/TP combinations generated and tested across four exit modes (SL+TP only, exit+SL, exit+TP, exit+SL+TP).

Step 04

Robustness Analysis

Rolling robustness evaluated across 9 windows. Robustness scores ranged from 0.462 to 0.745 (sltp_cand_8_161_exit_tp_topEfficiency) - best candidate is strong. Shortlist of 36 candidates built from top robustness performers, metric winners, and static params.

Step 05

Regime Analysis

36 candidates evaluated across 5 regimes (bear, bull, dump, pump, sideway) covering 15 total intervals. Best regime score: 0.678 (sltp_cand_8_161_exit_tp_topEfficiency) (acceptable).

Step 06

Scoring

36 candidates ranked by strategy composite score combining robustness, regime performance, and parameter stability. Best strategy composite score: 0.705 (sltp_cand_8_161_exit_tp_topEfficiency) (strong). Best robustness score: 0.767 (sltp_cand_8_161_exit_tp_topEfficiency).

Step 07

Shortlist Benchmark

Shortlist of 36 candidates backtested across full, yearly, monthly, and configured segment periods. Provides equity curves, trade breakdowns, and benchmark comparison for each candidate.

Step 08

Leverage Suitability

Leverage grid tested levels 1x to 50x with step 1 for 36 candidates. 17 candidates had at least one safe leverage level. Best grid leverage score is 0.784 (sltp_cand_8_161_exit_tp_topEfficiency); max safe leverage observed is 3x. Reference leverage 3x status: 1 ok, 29 unsafe, 6 insufficient data. This step is diagnostic and does not affect final scoring.

Step 09

Slippage Sensitivity

Slippage sensitivity tested 36 candidates across the configured execution-cost grid. 30 candidates stayed profitable through the maximum tested slippage (0.5000% per order). Most resilient candidate in the summary: sltp_cand_8_167_exit_sl_topProfit.

Step 10

Walk-Forward Adaptability

Walk-forward analysis across 12 windows, independently for each selection metric.
- Efficiency Score: adaptability 0.497 (weak), positive windows 41.7%
- Balanced Score: adaptability 0.234 (poor), positive windows 33.3%
- Profit Quality Score: adaptability 0.330 (poor), positive windows 41.7%