Nobody knows where the next bar goes

Here is the part most trading education skips. Look at a one-minute chart, cover the right-hand side, and try to say what the next bar does. Up or down, and by how much. Do it a hundred times and count. Nobody — not you, not us, not the person selling you a course — does meaningfully better than a coin on that question.

That is not cynicism. It is the working assumption behind everything we build. Intraday price movement is close enough to a random walk that treating the next bar as predictable is the fastest way to lose money with confidence.

What that rules out

If the next bar is unknowable, then a long list of familiar activities are entertainment rather than method:

• Forecasting where price “should” go, then trading the forecast.
• Adding to a loser because the move is “overdone.”
• Moving a stop because this particular trade feels different.
• Taking profit early because it looks like it might turn — and holding a loser late because it looks like it might come back.

Every one of those is a prediction wearing a disguise. Each one takes a system with a small measurable edge and replaces it with a guess.

What is left is not nothing

Accepting that you cannot predict the next bar leaves you three things you genuinely control, and they turn out to be enough.

One: where you engage. You cannot predict the next bar, but conditions are not identical all day and all year. Ranges expand and contract. Some sessions trend and some chop. An engine that only engages in the conditions it was built for is not predicting anything — it is declining to play the hands it plays badly. That is regime, and it is a filter, not a forecast.

Two: how much you risk. Position size and stop distance are decided before the trade exists, in dollars, by rule. This is the only variable in trading that is fully under your control and known in advance. It is also the one most often abandoned at the worst moment.

Three: how you leave. Entries get all the attention and deserve very little of it. The exit decides what the trade was worth. Not a feeling about the exit — an engineered rule: this stop, this target, this much given back from the best price seen, this time of day, flat.

Why this forces the rules to be boring

If the edge is a small statistical skew rather than a prediction, then any change made to flatter recent results is almost certainly fitting noise. The tempting move — retune the parameters after a bad month, tighten the filter that would have avoided last week — converts a thin real edge into a beautiful backtest and a disappointing live account.

So the rules stay fixed. The engines on this site were run across their published window without pausing for news days, without skipping inconvenient sessions, and without re-fitting parameters to make the chart look better. When we review, we review on a schedule against a baseline, and the decision is to keep, trim or retire — not to tinker.

What it means for live results

This is also the honest answer to the question every subscriber should ask: will live match the backtest? Close, we hope, and we publish the live fills so anyone can check. But never exactly. A backtest fills at a modelled price; a live account fills where the market actually was, on your data feed, at your latency, with your broker’s commission. Those differences are small per trade and they accumulate across hundreds of trades.

Anyone who tells you their live results track their backtest perfectly is either not running live or not looking closely. What we can promise is a fixed rule set, a published record of what it did, and the same per-trade evidence in your hands that we look at in ours.

The part that is actually hard

None of the above is intellectually difficult. It is difficult in the way that keeping a diet is difficult: the rule is simple and the temptation to break it arrives daily, usually dressed as insight.

An engine does not get bored, does not get stubborn about a level, and does not decide that this one is different. That is not the same thing as an engine being smart. It is an engine being consistent about the small number of things that can actually be controlled — which, once you accept that nobody knows where the next bar goes, is the only edge on offer.