Event-Driven Stock Strategy Backtester

What it is

The project started as a way to investigate whether clusters of public congressional stock transactions could be turned into a testable, repeatable strategy. It has grown into a chronological portfolio simulator with explicit entry, exit, position-sizing and cash-management rules.

How it works

The backtester is deliberately chronological. The signal logic and portfolio engine are separated so that the model can distinguish between when information becomes available and when a trade can actually be executed.

Current strategy rules

The rules are intentionally explicit so that changes can be tested one at a time instead of being hidden inside discretionary decisions.

Inside the code

This code view is adapted directly from a development excerpt of the chronological portfolio test. It is one of the pieces that helped move the project away from separate buy/sell phases and toward processing portfolio events in time order, so cash and open positions reflect what would actually have been available at each point in time.

print()
print("CHRONOLOGICAL PORTFOLIO TEST")
test_cash = starting_cash
open_positions = {}

for event in portfolio_events:
    position_id = event["position_id"]

    if event["event"] == "buy":
        position = positions[position_id - 1]
        cost = position["shares"] * event["price"]

        if cost > test_cash:
            raise ValueError(f"Not enough cash for position {position_id}")

        test_cash -= cost
        open_positions[position_id] = position

        print(
            event["date"].date(),
            "BUY",
            event["ticker"],
            "position",
            position_id,
            "cash:",
            test_cash
        )

    elif event["event"] == "sell":
        if position_id not in open_positions:
            raise ValueError(f"Position {position_id} is not open")

        position = open_positions[position_id]

2026-05-21 BUY PANW position 1 cash: 95187.5
2026-06-11 BUY NVDA position 2 cash: 90217.5
2026-08-26 BUY PANW position 3 cash: 85341.5

The project has continued to grow around this test with entry candidates, planned buy/sell events, position IDs, cash reuse after exits and portfolio-accounting logic. The output above comes from a development test used to verify chronology and state handling; it is not a historical performance result. The code window is intentionally an excerpt rather than a full source dump.

What I am learning

Python is a tool for the project rather than the end goal. The bigger challenge is learning to turn an idea into a system whose logic can be inspected, tested and improved without losing track of how the pieces fit together.

Current status

The project is actively being developed. The current milestone is completing and validating the end-to-end pipeline for the first full real-data portfolio backtest. Until that validation is finished, I intentionally avoid presenting performance statistics and focus instead on making the chronology, portfolio accounting and assumptions easy to inspect and verify.

About me

Hi, I'm Sverker Dilts Lindblom. This is a relatively new personal project that grew out of two things I enjoy: markets and learning how technical systems work by building them myself.

I started the project in late August 2026. During its first week and a half I worked on it across several focused sessions, going from setting up and learning the workflow around Python, Git and VS Code to building and testing core pieces of the backtester, including signal rules, chronological trade events, position tracking and cash handling.

What I enjoy most is the learning process: breaking an unfamiliar problem into smaller parts, understanding why the code behaves the way it does, and then improving the system one step at a time. The project has become a practical way for me to learn quickly without treating the code as a black box.