Backtest Truth, Part 1: Why a Clean Chart Can Invent Profit
A beginner look at fake fills. One Renko test showed +929% on brick prices, then lost about 72% to 99% when real candle prices were used.
On this page
This is education only and not financial advice. Trading involves risk, and you can lose money.
This is part 1 of Backtest Truth, a short series for beginners, including OFWs who want to understand a trading test before they put salary money anywhere near it. A backtest asks a simple question: if these rules had been followed in the past, what would have happened? The answer is useful only when the test is honest.
What a fill actually is
A fill is the price at which a trade actually opens or closes. In a real market you get a real price, such as the next candle’s open or the close of a real candle. You send an order, and the market gives you a price that existed. Some chart types do not use those prices. They draw a smoother picture. A careless backtest can then “buy” and “sell” at prices that never existed for a live order.
That is the whole trap in one sentence. The picture can be calm while the test is cheating, not because someone typed a fake profit on purpose, but because the chart’s own drawing price was used as if a broker would have filled you there.
Why a Renko chart looks so clean
Renko is one of those charts. It is made of “bricks.” A new brick is drawn only when price moves a fixed amount, such as 1% or a set dollar size. Small wiggles between those steps never become a brick. The chart looks clean because it hides the noise. That calm look is useful for reading direction. It is dangerous for testing fills.
Think of it like a bus that only stops every few kilometres. The map of stops looks simple. It does not mean you could have boarded at the stop after the bus had already passed it. By the time a brick is confirmed, the real price has often already moved past the neat level where the brick was drawn. A test that fills at the brick price is stepping onto a bus that left.
The result that was too kind
We learned this the hard way, on Bitcoin rules, not on a story from a forum. One Renko backtest printed an amazing result: +929%. It used the brick price, the neat level where the brick was drawn. The test was taking a small “free gift” on trade after trade. In one check, that gift averaged about 0.35 times the size of the stop, on every entry.
Read that slowly. The stop is the planned loss if the idea is wrong. A gift of about 0.35 times that stop, on every entry, is not a small drawing error. It is a discount the live market was not offering. Stack that discount across many trades and a test can print a huge gain that a real order could not have collected.
The same rules with real prices
When the same family of rules was run with real candle prices and real fees, the story changed.
- Over 2 months, a live-style Renko rule made only about +4.6%, with a drawdown of about 32%.
- Over 2 years it lost about 72%, with a drawdown of about 79%.
- Over the full history from 2020 it lost almost everything, about 99%.
Same market. The only honest change was the fill. The brick-price engine had been too kind.
A drawdown is the drop from a high point to a later low point in the test, before any new high. It is the pain in the middle, not the final headline. A two-month result of about +4.6% that still suffered about a 32% drawdown is not the same story as +929%. Two years at about a 72% loss, with about a 79% drawdown, is a different story again. The full history from 2020, about a 99% loss, says the kind fill was not a rare lucky week. It was the engine.
What to check before you trust a number
This is not an argument against reading charts. A smooth chart can still help a person see direction. It is an argument against trusting a test that fills at a price you could not have got. If a result looks too good to be true, check the engine before you trust the number.
Ask one plain question: at the moment the test says “enter,” what price could a real order have received? If the answer is the next candle’s open, or the close of a real candle, the test is at least speaking the market’s language. If the answer is “the neat brick level, after the brick was already confirmed,” the test is using a price that was already behind you.
Nothing in this part says Bitcoin, or any other market, should be bought or sold. The lesson is about the honesty of the test. Part 2 looks at the next quiet leak: the fee on every entry and every exit.
Key takeaways
- A backtest is only as honest as the prices it uses.
- Smooth charts such as Renko can hide the real entry price and inflate results.
- Always test with prices you could actually get, such as the next candle’s open or a real candle close.
FAQ
What is a fill, in plain words?
It is the price a trade actually opens or closes at. A fair test uses a real price, such as the next candle’s open or a real candle close, not a decorative level on a smoothed chart.
Why did one Renko test show +929%?
It filled at the brick price, the neat level where the brick was drawn. By the time a brick is confirmed, price has often already moved past that level. In one check the “free gift” averaged about 0.35 times the size of the stop on every entry.
What happened with real candle prices?
Over 2 months a live-style Renko rule made only about +4.6%, with a drawdown of about 32%. Over 2 years it lost about 72%, with a drawdown of about 79%. Over the full history from 2020 it lost about 99%.
Does this mean Renko charts are useless?
No. The calm look can help you read direction. It is a bad source of fill prices in a test. Check the engine before you trust a result that looks too good.
This is education only and not financial advice. Trading involves risk, and you can lose money.