What algo trading can and cannot do

You have now built, in miniature, a complete automated trading system: data, strategy, orders, risk controls, monitoring, tests. It would be a poor teacher who let you finish without saying plainly what such a system can and cannot do for you. This chapter is the honest reckoning the whole course has pointed toward. It will not tell you algo trading is easy money, because that would be a lie, and the data says so.

Automation does not change the odds

Return to the one idea this course opened with: automation is not an edge. A program trades faster, more consistently, and without emotion, but it does not trade with a better idea than the one you gave it. If your strategy has no real edge, and most do not, automating it does not create one. It just applies the same losing edge more times. The arithmetic is unforgiving.

# Automation does not change a strategy's edge per trade. It changes how many
# trades you make. So automation multiplies whatever edge you have: a positive
# edge compounds in your favour, a negative one against you, faster. Here is the
# same per-trade edge (after costs) at three frequencies. Figures illustrative.
for edge_per_trade, label in [(-0.001, "a small NEGATIVE edge after costs"),
                              (+0.001, "a small POSITIVE edge after costs")]:
    print(f"{label}: {edge_per_trade * 100:+.2f}% per trade")
    for trades_per_year in [50, 500, 2000]:
        annual = edge_per_trade * trades_per_year * 100
        print(f"   {trades_per_year:>5} trades/year -> {annual:+7.1f}% per year")
    print()

print("Automation multiplies the edge you already have.")
print("Most retail strategies, tested honestly, have a NEGATIVE edge after costs,")
print("so automating them trades toward zero faster, not slower.")

a small NEGATIVE edge after costs: -0.10% per trade
      50 trades/year ->    -5.0% per year
     500 trades/year ->   -50.0% per year
    2000 trades/year ->  -200.0% per year

a small POSITIVE edge after costs: +0.10% per trade
      50 trades/year ->    +5.0% per year
     500 trades/year ->   +50.0% per year
    2000 trades/year ->  +200.0% per year

Automation multiplies the edge you already have.
Most retail strategies, tested honestly, have a NEGATIVE edge after costs,
so automating them trades toward zero faster, not slower.

The two columns say everything. A strategy with a small positive edge per trade, after costs, compounds that edge as it trades more: pleasant. But a strategy with a small negative edge, which is what most retail strategies have once costs and slippage are honestly counted, loses more the more it trades: 5% a year at fifty trades becomes 200% at two thousand. Automation multiplies whatever edge you actually have. Since most retail edges are negative after costs, automation most often multiplies a loss. This is the same lesson as the SEBI statistic from the start of your trading journey, that most individual traders lose, seen now from the coder's chair.

The regulated reality

Recall the framework from the second chapter, because it shapes what you are actually undertaking. In India, retail algo trading now sits inside a SEBI framework: cross the order-rate threshold or use a third party's algo, and your activity must be registered through your broker; every algo order is tagged and traceable; your broker is the responsible principal; and you must meet the security rules. This is not a reason to be afraid, but it is a reason to be serious. Running an algo is a regulated activity with real obligations, not a private hobby the market cannot see. Treat it with the formality that framing implies, and confirm the current rules with your broker and SEBI before you begin, because they change. This is education, not legal or financial advice.

Is algo trading for you

So ask yourself honestly, as the Risk and Psychology course taught you to. Do you have a genuine, tested edge, validated out of sample, that survives costs and slippage? Almost no one does, and there is no shame in concluding you do not. Do you have the temperament to run a system that will have losing months, the discipline to keep it inside its limits, the patience to size up only slowly? Do you have money you can genuinely afford to lose, because the honest base case is that you might? For many people the right answer is that automated trading is a fascinating skill to understand and a poor way to try to make money, and that their savings belong in a low-cost index fund, held for years, not in a bot. That is not a failure of this course; it is one of its most valuable possible conclusions.

What to carry forward

Automation changes the speed and scale of trading, not its odds: it multiplies the edge you already have, and since most retail edges are negative after honest costs, it most often multiplies a loss, exactly as the expectancy table showed. Algo trading in India is a regulated activity, registered and traceable through your broker, to be taken seriously and checked against the current rules. And the honest question, whether you have a real, tested edge and the temperament and money to run it, has for most people the answer no, with a low-cost index fund the wiser path. None of that makes the skill worthless. The final chapter turns it into the thing it is genuinely good for: a disciplined practice, rehearsed safely.