Skip to main content

Why strategies often “don’t work”

Most failed strategies don’t fail because the idea is bad. They fail because something important was left implicit. This page highlights the most common missing details - so you can spot them early and iterate faster.
If a strategy behaves unexpectedly, it’s almost always due to a missing or misunderstood assumption.

1. Missing or weak exits

The problem

Traders focus heavily on entries and treat exits as an afterthought. Example:
Buy when EMA(20) crosses above EMA(50).
What’s missing:
  • when to exit
  • how much risk is acceptable
  • what happens if the trade goes nowhere
Entries define when you trade.
Exits define how you survive.

How to fix it

Always include at least:
  • one take profit
  • one stop loss

2. Implicit timeframes

The problem

You think in one timeframe - ATI can’t guess which. Example:
Trade RSI oversold conditions.
Missing detail:
  • Is this 5m, 1h, or 1D?
Even the same indicator behaves completely differently across timeframes.

How to fix it

Always state the timeframe explicitly, especially early on.

3. Direction not clearly defined

The problem

Some ideas imply buying, but never say it. Example:
Trade pullbacks using RSI.
Questions ATI must answer:
  • long only?
  • short only?
  • both?

How to fix it

Be explicit when possible:
Go long on RSI pullbacks.

4. Risk assumptions left vague

The problem

Words like:
  • “small stop”
  • “tight risk”
  • “controlled drawdown”
mean different things to different traders. ATI cannot translate intent without numbers or rules.

How to fix it

Even rough numbers are better than none:
Use a 1% stop loss.

5. Overly abstract language

The problem

Human intuition doesn’t always map to deterministic logic. Bad examples:
Enter on strong momentum.
Exit when price looks exhausted.
These cannot be backtested.

How to fix it

Translate intuition into observable conditions:
Enter when RSI(14) > 60.
Exit when RSI(14) crosses below 50.

6. Filters that silently block trades

The problem

You add a filter and forget it’s there. Examples:
  • trading only on certain days
  • session filters
  • strict confirmations
Result:
  • very few trades
  • or none at all

How to fix it

If a strategy produces no trades:
  • remove filters first
  • confirm baseline behavior
  • reintroduce filters one by one

7. Backtest range assumptions

The problem

You assume a long history - data may be shorter. Example:
Backtest from 2018 onwards.
But:
  • symbol launched later
  • timeframe data starts later

How to fix it

Always check:
  • effective backtest start date
  • number of trades generated

8. Expecting AI to “know what you meant”

The problem

ATI is powerful - but not telepathic. If multiple interpretations exist, ATI must choose or ask.

How to fix it

When results surprise you:
  • inspect inferred assumptions
  • rewrite only the unclear part
  • don’t start over

What Trinigence fills automatically

See which gaps ATI safely fills - and which require your input.

A quick self-check before submitting

Before you run a strategy, ask:
  • Is the timeframe explicit?
  • Is direction clear?
  • Is there a defined exit?
  • Is risk bounded?
  • Are filters intentional?
If yes - your strategy is structurally sound.
Most strategy problems aren’t logic problems -
they’re communication problems.