Tagging Setups So the Data Can Speak
If every trade is labeled simply as a buy or a sell, your journal cannot tell you what works. Consistent setup tags turn a list of trades into a research dataset.
Define a short list of named setups, each with written rules, such as "pullback to support in uptrend" or "range breakout at London open." Give each a fixed tag, and use only those tags. If a trade fits none, tag it "other" and look at those separately, since they often reveal rule-breaking.
Add secondary tags for conditions: trending or ranging, high or low volatility, news nearby or not. These allow cross-analysis, such as how the pullback setup performs in trending versus ranging markets.
Keep the number of tags manageable. Five to eight setup tags are plenty. Too many tags split your sample into pieces too small to analyze.
After thirty or forty trades per setup, compute results by tag: win rate, average R, and expectancy. You may find that one setup carries your profits while another quietly loses. That information directs your focus toward what works.
Resist the urge to redefine tags after seeing results, since that creates hindsight bias. Define them first, tag at entry, and analyze later.
Over time, the dataset becomes an asset that no course can give you, because it describes your own behavior in real conditions.
Practical step: write your list of setup tags with one-sentence definitions, and tag your last twenty trades retroactively as a test.
Tagging Setups So the Data Can Speak
Tagging Setups So the Data Can Speak
It’s Fairman 
Re: Tagging Setups So the Data Can Speak
The other tag is the most honest one in the journal. When it starts filling up, it usually means I'm taking trades that aren't in my plan. I'd review that bucket on its own every month. Look at the results, but also at when those trades happened. Very often they cluster after a loss, late in the session, or on slow days when nothing in the plan showed up. If one kind of trade keeps appearing in the other column and actually does well over a decent sample, it might deserve its own rules and tag. Most of the time though, the bucket is just a record of impatience, and the results show it.
It’s Fairman 
Re: Tagging Setups So the Data Can Speak
Tag at entry, not after the result. If you tag after, it's way too easy to file winners under your favourite setup and losers somewhere else. Your brain does this without you noticing, which is the frustrating part. The simplest fix is to make the tag a required field in the journal before the trade is placed, along with the stop and target. If you're using a screenshot, add the tag to the filename at that moment too. Once the trade is closed, the tag stays fixed. You can add a note later saying it didn't quite fit, but don't change the original. That's what keeps the setup statistics trustworthy.
It’s Fairman 
-
Shadow Trader
- Posts: 5594
- Joined: Wed Sep 09, 2026 5:19 pm
Re: Tagging Setups So the Data Can Speak
Thirty or forty trades per setup sounds modest until the secondary tags come in. Five setups crossed with trending or ranging and news or no news makes 20 cells, so 200 trades leave only about 10 in each, which is far too thin to trust.
The way around it is to slice by one condition at a time. Look at each setup split by trend state first, then separately by news, and only cross the two once a setup has a few hundred trades behind it.
One more tag rarely gets mentioned, and that is entry quality. A simple A, B or C grade given at entry, before the result is known, often shows that the C grade trades of a good setup lose money while the A grades carry it.
That turns the decision from dropping a setup into tightening its entry rules, which is a far cheaper fix. It also connects with the other bucket from the second post, because plenty of C grades are really unplanned trades wearing a familiar tag.
How do people here handle a setup whose definition changes partway through? Starting a fresh tag version seems cleaner than mixing old and new rules, but it resets the count to zero.
The way around it is to slice by one condition at a time. Look at each setup split by trend state first, then separately by news, and only cross the two once a setup has a few hundred trades behind it.
One more tag rarely gets mentioned, and that is entry quality. A simple A, B or C grade given at entry, before the result is known, often shows that the C grade trades of a good setup lose money while the A grades carry it.
That turns the decision from dropping a setup into tightening its entry rules, which is a far cheaper fix. It also connects with the other bucket from the second post, because plenty of C grades are really unplanned trades wearing a familiar tag.
How do people here handle a setup whose definition changes partway through? Starting a fresh tag version seems cleaner than mixing old and new rules, but it resets the count to zero.
-
FXS Prop Desk
- Posts: 1689
- Joined: Wed Oct 07, 2026 9:52 am
Re: Tagging Setups So the Data Can Speak
Limit the number of setup tags to a small number, such as five or fewer, at least at first. Too many tags spread trades thinly, and each tag needs enough trades (30–50 or more) before its statistics mean anything.
Reviewing the "other" tag regularly often reveals new setups worth defining or habits that need attention.
Once a tag passes that threshold, splitting it further by session or volatility becomes meaningful; before that, the extra detail mostly adds noise.
Reviewing the "other" tag regularly often reveals new setups worth defining or habits that need attention.
Once a tag passes that threshold, splitting it further by session or volatility becomes meaningful; before that, the extra detail mostly adds noise.
FXS Official Team | The Idea Lab & Strategy
-
Shadow Trader
- Posts: 5594
- Joined: Wed Sep 09, 2026 5:19 pm
Re: Tagging Setups So the Data Can Speak
Tags on taken trades only tell half the story, because setups that qualified and were skipped never reach the journal. Say a setup appears 40 times in a month and 28 get taken, then the statistics describe the 28 picked by mood as much as the setup itself.
A quick screenshot tagged as missed, with a note on why it was skipped, fixes that. Over a couple of months it often turns out the skipped ones did as well as the taken ones, which means hesitation is costing more than any filter saves.
Exit tags are the other piece worth adding before the list grows. Marking each close as stop, target, time or manual lets a setup be judged separately from how it was handled.
Manual exits are usually the revealing bucket. If a setup is positive on trades left alone and negative where it was closed early by hand, the setup is fine and the management is the problem.
On FXS Prop Desk's limit of five, does anyone count exit tags against it? They feel like a separate dimension, but they still split the sample.
A quick screenshot tagged as missed, with a note on why it was skipped, fixes that. Over a couple of months it often turns out the skipped ones did as well as the taken ones, which means hesitation is costing more than any filter saves.
Exit tags are the other piece worth adding before the list grows. Marking each close as stop, target, time or manual lets a setup be judged separately from how it was handled.
Manual exits are usually the revealing bucket. If a setup is positive on trades left alone and negative where it was closed early by hand, the setup is fine and the management is the problem.
On FXS Prop Desk's limit of five, does anyone count exit tags against it? They feel like a separate dimension, but they still split the sample.
-
Shadow Trader
- Posts: 5594
- Joined: Wed Sep 09, 2026 5:19 pm
Re: Tagging Setups So the Data Can Speak
A journal listing entry, exit and result tells what happened. Tags tell why, and why is where improvement lives.
Useful tags include the setup name, the session, the pair, the market condition such as trend, range or news, and a quality rating assigned before the trade. A few emotional tags help too, such as rushed, bored, or revenge.
After a couple of hundred trades, slicing the data by tag reveals patterns no amount of memory could. Perhaps the range setups pay and the breakout setups bleed. Perhaps trades tagged bored lose more than anything else.
Keep the tag list short. Too many options mean inconsistent tagging, and inconsistent tags make the analysis meaningless. Six to ten categories done reliably beat thirty done randomly.
Tagging should happen at entry, not at the end of the day. Memory rewrites history, and a trade that felt rushed in the moment often gets recalled as a well-planned one once the result is known.
Which tags would show the biggest gap between best and worst, if they were being recorded?
Useful tags include the setup name, the session, the pair, the market condition such as trend, range or news, and a quality rating assigned before the trade. A few emotional tags help too, such as rushed, bored, or revenge.
After a couple of hundred trades, slicing the data by tag reveals patterns no amount of memory could. Perhaps the range setups pay and the breakout setups bleed. Perhaps trades tagged bored lose more than anything else.
Keep the tag list short. Too many options mean inconsistent tagging, and inconsistent tags make the analysis meaningless. Six to ten categories done reliably beat thirty done randomly.
Tagging should happen at entry, not at the end of the day. Memory rewrites history, and a trade that felt rushed in the moment often gets recalled as a well-planned one once the result is known.
Which tags would show the biggest gap between best and worst, if they were being recorded?