I automated a Discord Blackjack bot and the house still won

I created a Discord Blackjack bot designed to operate within OwO Discord bot guidelines without triggering moderation or depleting funds rapidly. The aim was to explore how human-like timing and core game rules could be integrated into LLM-style automation, not to develop a system guaranteed to win at gambling.

The resulting project is called GhoSty OwO BlackJack Farm, focusing on making automation appear natural rather than robotic.

Technical Setup and Challenges

Using modern discord.py libraries for a self-bot is problematic because recent versions lack the self_bot=True option due to Discord's terms of service. To bypass this, the project relies on discord.py==1.7.3, a version that allows interaction with user tokens instead of bot tokens. The project directory contains minimal files: main.py for the core engine and game logic, config.json for settings, requirements.txt, and README.md.

To deploy the bot, the following commands are executed:

pip install discord.py==1.7.3 colorama
python main.py

Implementing Natural Timing

The primary challenge in bot automation is avoiding predictable timing patterns. A script that sends commands at exact intervals will likely be detected. Two strategies were employed to address this:

1. Randomized Delays

A "human pause" function was implemented to avoid fixed timing. Each action includes a randomized delay, creating a more natural execution rhythm:

import asyncio, random

async def human_pause(base: float = 1.5, spread: float = 2.0):
    """Sleep for a randomized, human-ish interval."""
    await asyncio.sleep(base + random.uniform(0, spread))

2. Work and Rest Phases

Continuous operation would attract scrutiny, so the bot was designed to work in cycles. It runs sessions followed by extended breaks:

while running:
    await play_session() # The active work cycle
    await take_break()   # The long-term cooldown

Betting Strategy

The Martingale system was avoided, as it is mathematically unsound for long-term success. Instead, a straightforward Blackjack strategy based on the player's hand total and the dealer's visible card was used. The decision-making logic is:

def decide(total: int, dealer_up: int) -> str:
    if total >= 17:
        return "stand"
    if total <= 11:
        return "hit"
    # Simplified logic for demonstration
    return "stand" if dealer_up <= 6 else "hit"

Outcome and Takeaways

Despite the sophisticated code and betting strategy, the house edge remains a fixed probability. The project serves as a practical example of logic patterns relevant to prompt engineering or basic agent behavior, but it does not change the fundamental odds of gambling. It demonstrates effective deployment, pacing, and rule-driven automation, which is sufficient in itself.

The project is available on GitHub.