Keeping a changelog for your trading rules

[Fairman]

Keeping a changelog for your trading rules

Software developers keep changelogs, a record of every change to the code, with the date and reason. I think it's a useful idea for trading rules too.

Most traders change their rules over time. Adjusting a stop distance, adding a filter, dropping a setup, changing the trading hours. Often the change is made in the moment and not recorded. A few months later, you can't remember when it happened or why, and you can't tell whether results changed because of it.

A changelog fixes that. A simple document or spreadsheet: date, what changed, why, and what you expect to happen. Then later, you can compare results before and after each change.

It also slows down changes. If you have to write down the reason, you think twice about changing a rule after one bad trade. That alone stops a lot of unnecessary tinkering.

The 'what you expect' part is important. If you add a filter expecting it to reduce losing trades, you can check after a month whether it did. If it didn't, you know to reconsider.

It's also helpful when results drop. You can look at the changelog and see whether a recent change might be the cause, or whether nothing has changed and it's probably just the market.

It doesn't need to be elaborate. A line per change is enough.

[Fairman]

Tagging trades in the journal with the rule version helps link the changelog to results. Version 1, version 2, and so on, with each change bumping the number. Then you can filter trades by version and compare. Without that, you have to match dates manually, which works but is slower. For EAs, the version number can be written into the trade comment automatically, so every trade carries the version it was taken under. An EA can also print its version on the chart, which helps if you run more than one copy on different accounts. It removes any doubt about which version took which trade.

[Fairman]

One rule I use with the changelog is to only change one thing at a time. If you change the stop and add a filter on the same day, and results improve, you don't know which change helped. It's slower to change one at a time, but the information is much clearer. The changelog makes it obvious when you're tempted to bundle several changes together, since you'd be writing several lines on the same date. If you really need to change two things at once, logging both and accepting that the results will be mixed is better than not recording it at all. At least the changelog then tells you why the comparison is murky.

[FXS Prop Desk]

Changelogs solve a real problem for traders too: after a few months of small changes, nobody remembers whether the current version of the strategy is better than the original one.

To make it more useful, each entry could include the reason for the change and a review date. For example, "added filter to skip trades when spread is above 1.2 pips, review after 50 trades". On the review date, compare results before and after. If the change did not help, it can be removed with a clear record of why.

[Fairman]

The changelog is also useful when talking with other traders or mentors. If someone asks why you trade a certain way, you can show the history of changes and the reasons behind them. It shows the rules didn't appear from nowhere and that they've been tested over time. It also helps you explain your process more clearly. If you've forgotten why a rule exists, the changelog tells you, which can stop you from removing a rule that was added for a good reason months ago. That alone makes the habit useful.

[Fairman]

Should you log changes you tried and reverted? I'd say yes. If you added a filter, tested it for a month and removed it because it didn't help, that's useful information. Without it, a few months later you might have the same idea and test it again from scratch. A line saying 'tried X, reverted after four weeks, reduced trades without improving results' saves time. The changelog becomes a record of what's been tried as well as what's currently in use, which is more valuable over time. It also stops the same failed idea from coming back every few months.