One percent at a time: what a Japanese manufacturing principle can teach your trading
Continuous improvement in trading does not require dramatic changes; small, consistent improvements can compound over time.
When traders hit a difficult period, there is often a temptation to do something dramatic.
Rewrite the strategy. Add new filters. Change the timeframe. Replace the entry. Tighten the stop. Start again.
It can feel productive because something is being changed.
But large changes made in response to a short run of disappointing results can make it harder to determine whether the original strategy had a genuine problem at all.
A useful alternative comes from a very different field: the idea of Kaizen, commonly associated with continuous improvement in Japanese manufacturing and management.
The trading lesson is simple: instead of constantly rebuilding the entire process, identify one area that may be improved, change it deliberately, measure what happens and decide from the evidence whether the adjustment deserves to remain.
What Kaizen Actually Means
Kaizen is commonly translated as “change for the better” and is closely associated with the idea of continuous improvement.
It became particularly influential within Japanese manufacturing and quality-management systems in the decades following the Second World War.
The broader philosophy emphasises improving processes through repeated, deliberate changes rather than relying only on occasional large-scale transformation.
Importantly, Kaizen does not literally mean that every improvement must be exactly 1%.
“One percent at a time” is a useful way of expressing the philosophy, not a mathematical rule.
The more important principle is:
Make improvements small enough to understand, measure and integrate.
The Plan-Do-Check-Act Logic
Continuous-improvement systems are often connected with the Plan-Do-Check-Act cycle:
Plan: identify a specific change worth testing.
Do: implement the change in a controlled way.
Check: compare the result against the previous process.
Act: keep, modify or reject the change according to the evidence.
Then the cycle begins again.
The value is not simply that something changes.
The value is that the change has to survive review before becoming part of the permanent process.
Why the Same Logic Fits Trading
A trading strategy should not be rebuilt every time a losing sequence appears.
Even a strategy with positive expectancy can experience consecutive losses.
The first question should therefore be:
“Is this performance outside what the strategy could reasonably produce, or am I reacting to normal variation?”
If the drawdown remains within previously observed or reasonably expected parameters, changing the strategy immediately may be premature.
If the evidence does suggest a genuine weakness, the next step does not necessarily need to be a complete redesign.
A smaller, testable adjustment may provide much more useful information.
Do Not Confuse Improvement With Constant Modification
There is an important tension here.
Kaizen encourages improvement.
Trading requires enough stability for results to be measured.
If a trader changes something after every few trades, the strategy never remains unchanged long enough to establish what the original version actually does.
Continuous improvement therefore does not mean continuous interference.
A better process is:
Trade the defined version.
Collect enough evidence.
Identify one possible weakness.
Test one modification.
Compare it with the original.
Only then decide whether the strategy should change.
Change One Variable at a Time
This is one of the most useful lessons traders can take from continuous-improvement thinking.
Suppose a trader simultaneously changes:
The entry timing
The stop-loss distance
The profit target
The trading session
Performance subsequently improves.
What caused the improvement?
There is no clean answer.
Perhaps one change helped significantly.
Perhaps another made performance worse.
Perhaps the market environment simply became more favourable.
By changing one meaningful variable at a time, the trader has a much better chance of understanding what actually affected the result.
A Trading Example
Imagine a trader believes entries may be occurring too early.
Instead of redesigning the entire strategy, they create a specific hypothesis:
“Would requiring an additional price-action condition before entry improve the strategy’s overall results?”
The important point is not that delaying the entry is automatically better.
A later entry could:
Improve the entry price in some circumstances.
Reduce false entries.
But it could also:
Cause profitable moves to be missed.
Reduce trade frequency.
Produce a worse price.
Change the achievable risk-reward profile.
The modification therefore needs testing.
If the trader also wants to test a tighter stop, that should ideally be treated as a separate experiment rather than automatically combining both changes.
Define What “Better” Actually Means
A modification should not be judged simply because the new equity curve looks attractive.
Before testing, define what improvement means.
Possible measures include:
Expectancy
Maximum drawdown
Average win
Average loss
Win rate
Profit factor
Trade frequency
Transaction costs
Consistency of execution
A modification could improve one metric while damaging another.
For example, an extra entry filter may increase win rate while reducing trade frequency so substantially that overall expectancy does not improve.
That is why the entire system needs to be evaluated rather than one attractive number.
Use a Baseline
Improvement cannot be measured without knowing what is being improved.
Before changing anything, record the performance of the current version.
Your baseline might include:
Number of trades
Win rate
Average R multiple
Expectancy
Maximum drawdown
Average holding period
Trading costs
Rule-adherence rate
Then the revised version can be compared against something concrete.
Give the Change Enough Time to Produce Evidence
One of the hardest parts of incremental improvement is patience.
A change cannot be judged reliably from two or three trades.
A small sample may make an improvement look exceptional simply because the first few outcomes happened to be favourable.
The reverse can also happen.
A useful modification may initially experience several losses.
The sample required will depend on the strategy, its trade frequency and the size of the effect being measured.
The important principle is to define the review point in advance rather than deciding after each individual outcome whether the change worked.
Do Not Test Only on the Data That Suggested the Change
There is another important danger.
Suppose your historical review reveals that adding a particular filter would have removed ten losing trades.
You add the filter and then test it on exactly the same historical period.
Unsurprisingly, the result looks better.
But the same data helped create the rule and evaluate it.
Where practical, a stronger process is to develop the modification on one sample and then evaluate it on data that was not used to create the change.
This does not eliminate overfitting, but it provides a more meaningful challenge than repeatedly improving performance against the same historical sample.
Not Every Improvement Needs to Change the Strategy
Kaizen can also be applied to the trader’s operating process rather than the market rules themselves.
Improvements might include:
A clearer pre-trade checklist.
A faster position-sizing process.
Better screenshot organisation.
More complete journal entries.
A clearer daily loss rule.
A more consistent post-market review.
Reducing distractions during trading hours.
These changes may improve execution without altering the underlying strategy at all.
Often that distinction is worth investigating before changing the setup itself.
Improvement Can Also Mean Removing Something
Continuous improvement does not always mean adding another rule.
Sometimes improvement comes from simplifying.
A trader may discover that:
An indicator adds little useful information.
A filter reduces opportunities without improving expectancy.
A complicated journal field is never used during review.
An unnecessary trading session produces weaker results.
Removing waste is entirely consistent with the continuous-improvement philosophy.
The objective is a better process, not a more complicated one.
Know When Not to Change Anything
This may be the most important Kaizen lesson for trading.
Sometimes the correct result of the review is:
No change.
A losing streak can be uncomfortable while remaining statistically ordinary.
A difficult month can occur without demonstrating that the trading rules are defective.
The trader should therefore distinguish between:
A problem requiring improvement
and:
Normal variation requiring patience.
Continuous improvement works only when changes respond to evidence rather than discomfort.
A Simple Kaizen Trading Cycle
The entire approach can be reduced to six steps:
1. Measure. Establish how the current process is actually performing.
2. Identify. Choose one specific weakness or inefficiency.
3. Hypothesise. Define one modification and what improvement should look like.
4. Test. Apply the change under controlled conditions.
5. Compare. Measure the modified process against the original baseline.
6. Decide. Keep, reject or refine the change according to the evidence.
Then repeat when another genuine improvement opportunity appears.
What This Comes Down To
Do not rebuild a strategy because a few trades went badly. First establish whether the results are outside normal expectations.
Continuous improvement does not mean constant tweaking. The strategy needs periods of stability so its behaviour can actually be measured.
Change one meaningful variable at a time. Otherwise it becomes difficult to identify what caused the result.
Define improvement before testing. Decide which metrics matter rather than simply choosing whichever result looks best afterwards.
Use enough data. A handful of trades rarely provides strong evidence that a modification genuinely improved the process.
Validate changes beyond the data that created them. This helps reduce the risk of designing a strategy around historical accidents.
Improve execution as well as strategy. Checklists, journals, risk procedures and routines may sometimes deserve attention before the setup itself.
Sometimes improvement means doing nothing. Normal variance does not require a strategic response.
Conclusion
Kaizen offers traders a useful alternative to the cycle of frustration, overhaul and reinvention.
The objective is not to find one dramatic change that suddenly transforms the strategy.
It is to create a process in which weaknesses can be identified, changes can be tested and improvements can be incorporated without destroying the stability needed to measure what is happening.
Measure first.
Change one thing.
Test it.
Keep it only if the evidence supports it.
Then repeat.
That approach is slower than rebuilding everything after every difficult period.
But in trading, slower and measurable is often far more useful than dramatic and impossible to evaluate.