---
title: Trading Software Generations: From MT4 to LLM
url: https://stratcraft.ai/insights/trading-software-generations
type: insight
---

# Trading Software Generations: From MT4 to LLM

Trading software evolved through four practical generations: indicator components, packaged strategies, system-level research pipelines, and LLM prompt-to-code workflows.

## Generations Comparison

| Dimension | Gen 1: Indicator Era | Gen 2: Strategy Era | Gen 3: System Era | Gen 4: LLM Era |
|---|---|---|---|---|
| Time period | ~2005-2012 | ~2012-2018 | ~2018-2023 | 2023-present |
| Representative products | MT4 / MQL4 | Backtrader, freqtrade, vnpy | QuantConnect, WorldQuant BRAIN | ChatGPT + brokerage API |
| Core abstraction | Modular indicators | Packaged strategy logic: entry, exit, sizing | Feature engineering pipeline | Natural language to code |
| Typical workflow | Drag indicators, set conditions, backtest | Write strategy code, optimize parameters, backtest | Build feature pipeline, train, validate, execute | Prompt, generate strategy, backtest |
| Main progress | Indicators moved from books and forums into reusable components | Complete strategies became packageable, shareable, and reproducible | The pipeline became the product, with standardized validation | Generation is fast and the entry barrier is low |
| Typical trap | Indicator worship and holy-grail thinking | Parameter overfitting: genetic search finds coincidence, not robustness | Data leakage and survivorship bias | Strategy hallucination and backtest overfitting blindness |
| Failure root cause | Rules are manual and lack systematic validation | The optimization target is backtest return, not robustness | Model complexity hides data-quality problems | The model skips infrastructure accumulated across the first three generations |

## LLM Failure Modes

### Strategy Hallucination

The model emits plausible trading logic that looks quantitative but has no market rationale, no statistical grounding, or invalid assumptions. It creates code that passes syntax checks while smuggling in false causality.

### Backtest Overfitting Blindness

The model treats a profitable backtest as validation and misses leakage, parameter mining, unstable regimes, or survivorship bias. It accelerates curve fitting because generation speed multiplies untested variants.

## Z-Score Pattern Example

```python
spread = asset_a.close - hedge_ratio * asset_b.close
z_score = (spread - spread.mean()) / spread.std()

if z_score > 2:
    short(asset_a)
    long(asset_b)
elif z_score < -2:
    long(asset_a)
    short(asset_b)
```

## Model + Tools

LLMs are useful when wrapped in tools. A serious trading stack is not just a model; it includes pricing engines, execution logic, risk controls, data validation, monitoring, and deployment discipline. Jane Street-style systems show the pattern: the model is one layer inside a larger toolchain.

## StratCraft Positioning

StratCraft is not a Gen 5 claim. It is infrastructure for Gen 2 and Gen 3 developers who need local C++-grade backtesting performance, plugin isolation, and an ecosystem that makes strategy generation, validation, and composition repeatable.

Source: https://stratcraft.ai/insights/trading-software-generations