What a strategy does
A strategy is a fixed set of rules that decides, every five days, whether to hold one market or sit in cash. The app replays those rules on past prices and shows you what came of it.
Think of a strategy as a recipe written down in advance. It looks at a market's price, say BTC (Bitcoin), and answers one question: should I own this right now, or not? The answer comes from rules, not from anyone's mood, so the same prices always give the same answer.
Golden Path Strategies replays each recipe over years of recorded prices, as if someone had followed it to the letter since 2015. That replay is called a backtest. Nothing is guessed and nothing is traded: it's a record of what the rules would have done. The app never places orders for you and doesn't tell you what to buy. It shows you how a set of rules behaved, and what you do with that is up to you.
Long or Flat, nothing in between
At any moment a strategy is in one of two states, shown as a small pill next to its name.
- Long means it holds the market. If BTC rises 10%, the strategy's account rises 10% (and falls with it, too).
- Flat means it holds cash. Whatever the market does, the account stays where it is.
For most strategies there's no third option. They never bet on prices falling (that would be "short"), never borrow money, and never hold half a position: a buy puts the whole account in, a sell takes all of it out. When the docs say a strategy "steps aside" or "goes flat", this is what they mean.
Two kinds work differently, and each has its own folder, DCA and Multi-Ticker. DCA books keep a share of the account in BTC and turn that share up or down in steps, so they're often partly invested. Two-market strategies compare two markets, and in their dollar view Flat means holding the second market rather than cash. Neither ever goes short or borrows.
So what does the account hold while it's Flat? In the backtest it's simply cash in dollars, earning no interest, which is a slightly pessimistic assumption.
A small example shows why this matters. Say the market rises 20%, then falls 30%, then rises 50%, and the strategy happens to be Flat during the fall:
| Market move | Strategy | Strategy account | Simply holding | |
|---|---|---|---|---|
| Start | 100 | 100 | ||
| Step 1 | +20% | Long | 120 | 120 |
| Step 2 | −30% | Flat | 120 | 84 |
| Step 3 | +50% | Long | 180 | 126 |
Real rules are never this neat. They buy a bit after a rise has started and sell a bit after a fall has begun, and now and then they sell right before a rally. But that's the whole idea: miss the worst of the long falls, catch most of the long rises.
[Where the gold line runs level the strategy was Flat, in cash, while the blue Buy & Hold line kept falling. Both lines start at 100.]
What 5D means
You'll see 5D on every card. It means each strategy looks at the market once per five-day candle (a "bar": the open, high, low and close of that stretch), and decides at the close. Most strategies trade at that same close, a few at the start of the next bar. You don't choose a timeframe; every strategy here is built and tested on 5-day bars, and that's the only way it's published.
Five days was chosen after years of testing shorter and longer bars, and About tells that story. The upshot is that most strategies trade only a few times a year: one with 16 trades since 2015 isn't broken, it's patient. How a bar is cut for crypto and for stock funds is on How the backtests are built.
The account that starts at 100
Every backtest starts its account at 100. It isn't dollars or euros, just a starting point that makes strategies easy to compare: if the line ends at 250, the rules turned 100 into 250 over that period. Next to it you'll usually see Buy & Hold, an account that bought the same market on day one and never sold. That's the plain alternative a strategy should be measured against.
Most backtests charge no trading fee, and none include spreads, slippage or taxes; How the backtests are built lists every assumption. Real results are likely to come in below the backtest, since these rules were chosen with the past in view.
Why the names sound like science fiction
Names like Von Lohengramm T7, God Emperor T9 or Lisan al-Gaib C4.2 are just labels, borrowed from Dune and Legend of the Galactic Heroes. They aren't people or funds. The part after the dash tells you the market (- BTCUSD is BTC priced in dollars). A letter and number, where there is one, tell you the strategy is a vote among smaller strategies and how that vote works:
- T stands for threshold: it waits until a minimum number of votes agree before it buys. God Emperor T9 has nine strategies voting.
- C stands for crossover: it's in the market as long as there are more bullish votes than bearish ones. Lisan al-Gaib C4.2 has four voters; the ".2" marks a later version.
- Names without that letter, such as RMSE Envelope Trail or Regime Momentum Oscillator, are mostly single strategies that follow one idea.
Neither family is safer by design. Judge each one by its own numbers, especially its worst fall. Choosing strategies to follow shows how.
How votes work
Many strategies are a vote among smaller strategies, and their card's category reads consensus. Take the description of Von Lohengramm T7 - BTCUSD: "Seven strategies share nine votes." Two of the seven count double (candle pressure and swing momentum get two votes each), so seven voters cast nine votes in total. The strategy buys when five of the nine say "up", and steps aside when five say "flat".
Why bother? Each small strategy is right in some markets and wrong in others. A vote smooths out their individual mistakes: one nervous indicator can't throw the strategy out of a good trend on its own. On a strategy page you can watch the vote count bar by bar; see The strategy page.
Published runs don't change
Each strategy is a finished, stored run, and everyone sees the same one. Every night the newest prices are added at the end; the past part stays as it was. When you pick a shorter window, such as the last 3 years, the app cuts the stored run to those dates and restarts both lines at 100. It never reruns the rules to find a better result.
A backtest is history. Rules that worked well on past prices can stop working, and a beautiful line says nothing certain about the next year. Overfitting and how we test for it explains how to look at a result critically.