Go looking for a journal that will help you find your edge and you usually land in one of two places.
The first place assumes you trade a specific way. Order blocks, fair value gaps, liquidity sweeps, the whole Smart Money Concepts / ICT dictionary baked straight into the tag list. Great if that's your world. Useless if you trade a 20/50 moving average cross, or a volume profile value area, or a simple gap fill on the open. You end up mapping your own ideas onto someone else's words and hating it.
The second place gives up entirely. A free-text notes box and a tag field where you can type anything. Total freedom. And after 300 trades you have london, London, LDN session, good setup, clean, nice entry, revenge, and no way on earth to compute anything from that pile.
I think both options are a mistake, and I think they're the same mistake: assuming structure and neutrality can't live in the same journal. They can. You just have to be strict about the shape of the data and completely relaxed about the content.
Structure is about slots, not about vocabulary
Here's the distinction that took me too long to see.
A framework can tell you what kind of thing goes in each box without telling you which thing goes in it. "Market environment" is a slot. Whether you fill it with "trending", "high VIX", "post-CPI", or "Asian range day" is entirely your business.
That's the unclaimed middle ground. Strict enough that a computer can group and count your trades. Neutral enough that a supply-and-demand trader and a moving-average trader can both use it without either one learning the other's language.
So the question becomes: how many slots, and which ones?
The four places an edge can actually come from
If you strip away all the strategy-specific noise, a trade only has four moving parts that you control or observe. Every real edge lives in one of them, or in a combination of them.
1. Technical setup. What you saw on the chart that made you take the trade. The pattern, the level, the indicator reading, the structure. This is the part everyone obsesses over, and it's one of four places the answer can be hiding. Not necessarily the biggest one, and for a lot of traders not even close.
2. Execution. How you actually got in. Entry timing, where the stop went, whether you took the trade at the level or chased it 15 pips late, whether you followed the plan at all.
3. Environment. What kind of day, week or regime it was. Trending or chopping. Volatile or dead. Session. News on the calendar or not. The setup didn't happen in a vacuum.
4. Management. Everything after entry. Break-even moves, partials, trailing, cutting early, letting it run to target. This one changes your results more than most traders expect, which is why management deserves to be measured as part of your edge rather than treated as a personality trait.
That's it. There is no fifth place. Psychology feels like a fifth place, but when psychology hurts you it shows up as an execution problem or a management problem, and that's exactly where you want it recorded, because that's where you can count it.
Four layers is not a number I picked because it looks tidy on a landing page. It's four because those are the four parts that can vary separately on the same trade. Collapse execution and management into one "how I traded it" bucket and you can no longer tell a late entry from a panicked exit — two problems with completely different fixes. Keep them apart and the journal can finally answer whether the setup is bad or you are, which is covered in more depth in the post on whether you have an edge problem or an execution problem.
What this looks like for two very different traders
Let me make this concrete, because the whole claim rests on it working for people who'd never agree on anything.
One bit of vocabulary first, because it's about to appear in both lists. R is just the money you're risking on a trade, counted as one unit. Risk $100 and a 1R win makes that $100 back. It lets you put a small trade and a big one on the same scale, which is why the rest of this post talks in R instead of dollars.
Trader A trades supply and demand zones on the 4H. Same trade, tagged across four layers:
- Technical setup:
4H demand zone,first touch,bullish engulfing on 15m - Execution:
limit fill at zone,stop below zone low,1.0R risk - Environment:
4H uptrend,London open,no red news - Management:
partial at 1R,runner to next supply
Trader B trades a 20/50 EMA cross on the 1H and has never used the word "liquidity" in her life:
- Technical setup:
20/50 cross up,pullback to 20 EMA,RSI above 50 - Execution:
market entry on candle close,stop 1.5 ATR,1.0R risk - Environment:
ADX above 25,New York session,earnings week - Management:
break-even at 1R,trail on 20 EMA
Nothing is shared between those two lists. Not one word. But the shape is identical, and that means the same analysis runs over both. Both traders can ask: what's my expectancy, meaning the average result per trade across a decent sample, when the environment layer says trend versus range? Both can ask: do my losers cluster in the execution layer? Neither had to adopt the other's vocabulary to get an answer.
This is the principle we built EdgeFlow around, honestly more out of stubbornness than strategy. The category examples that ship with it are starting points you can delete. They are not a methodology you're signing up for.
Once you have four layers, the good questions appear
Structure isn't valuable because it's tidy. It's valuable because of what it lets you ask.
Say you've got 200 trades in a demand-zone strategy. Overall it's +0.11R per trade. Fine, slightly profitable, nothing exciting. Now split by the environment layer:
- 4H uptrend, 120 trades: +0.31R
- 4H ranging, 80 trades: −0.19R
Read that in plain terms: in an uptrend the average trade earns about a third of what it put at risk, and in a range the average trade loses about a fifth of it. Suddenly you have something real. Not "my strategy works", but "my strategy works in trend and bleeds in range". That's an actionable rule, and it came from a layer most traders never tag at all.
But be careful here, because this is exactly where people fool themselves. Two splits on 200 trades is fine. Twenty splits on 200 trades and you'll find a "winning" combination by pure chance. The 80-trade ranging bucket above is already thin enough that you'd want to see it hold up on the next 50 trades before you rewrite your plan. Slicing your journal honestly is a skill in itself, and testing condition combinations has more on where the traps sit.
The four layers also give you a natural way to sort conditions by how much they matter. Some conditions are non-negotiable, some help, some kill the trade. Sorting your tags into required, supporting and avoid conditions turns a flat tag list into something that resembles an actual trading plan.
Rules that keep the framework usable
A few things I'd insist on, learned the boring way.
Tag what you saw, not what happened. If you only mark a day as "ranging" after you lose in it, your data is hindsight, not evidence. Decide before or at entry.
Keep the tag list small and stable. Ten to fifteen tags per layer is plenty. Every new tag splits your sample smaller. If you have 300 trades and 40 technical tags, most tags have five or six trades behind them and five or six trades tell you nothing.
Never let the same word live in two layers. "London" is environment. It is not a technical condition. Sounds obvious, gets violated constantly, and it quietly wrecks your grouping.
Write down what each tag means. One line each. Future you will not remember whether "late entry" meant more than 5 pips or more than 15.
Where this pays off later
The other thing four layers buys you is diagnosis when performance drops. And it will drop.
Without layers, a losing month is just a losing month, and the honest answer is "I don't know". With layers, you can look at whether your environment mix changed (maybe the market stopped trending), whether your execution drifted (maybe you started entering later), or whether the setup itself genuinely stopped producing. Those three have completely different fixes, and telling them apart is most of what figuring out why a setup stopped working comes down to.
Nobody can promise you that tagging four layers will find you an edge. There might not be one there. Structure doesn't manufacture an advantage, it just makes an existing one visible, and makes the absence of one visible too, which is the less fun half of the deal.
But if you don't have the layers, you're not even in a position to look. You're just collecting screenshots.
Start where you are
You don't need to rebuild your journal this weekend. Take your existing tags, whatever mess they're in, and sort them into the four buckets. Most traders find their technical bucket is overflowing and the other three are nearly empty. That imbalance is the finding. It's telling you that three quarters of your edge has never been measured.
If the vocabulary itself still feels shaky, the piece on what a trading edge actually is is a reasonable place to back up to first.
And if you want the four-layer structure already wired into the statistics, that's what the edge-building side of EdgeFlow does — in your own words, not ours.