Critical formulas for Bollinger Band trading – John Bollinger – 056
John Bollinger has spent 46 years in markets. Naturally, I expected our conversation to revolve around Bollinger Bands.
It did. But that wasn’t what stayed with me.
What struck me most was how consistently John returned to a handful of principles: simplicity, deep understanding of your models, robustness, first principles and hard work. Coming from someone who has spent nearly half a century researching markets, that’s worth paying attention to.
His philosophy could almost be reduced to one rule:
Build simple and elegant systems around things that are fundamentally true about markets. Then leave them alone.
That idea ran through almost everything we discussed.
There were some gems in this interview, so I’m going to point a few of them out in this article.
Start With What the Market Gives You
John primarily trades US equities, and his starting premise is wonderfully uncomplicated: stocks have a long-term upward bias.
The odds are already tilted in your favour.
His argument isn’t therefore to discover some astonishing new market anomaly. If equities already offer a long-term positive return, the job of active management is to participate in that structural tailwind while finding intelligent ways to improve upon it, particularly by avoiding periods when you don’t want to own stocks.
I like this framing.
Too much quantitative research begins with the implicit assumption that we need to manufacture an edge from nothing. Bollinger starts with an edge the market already provides and asks what can be layered on top.
Breadth is one example.
He remains a strong believer in McClellan breadth indicators, particularly the McClellan Summation Index, for intermediate and longer-term market timing. Price tells you what the index is doing. Breadth helps tell you what the market underneath it is doing.
Then he also likes waiting for confirmation.
A Bollinger Band setup isn’t necessarily a signal. He deliberately calls many of them alerts. Price still needs to confirm the idea.
Small improvement. Large consequence.
Volatility Is a First Principle
This also explains why Bollinger Bands have survived for more than 40 years.
They aren’t a trading system. They answer a simpler question:
Are prices relatively high or relatively low?
More importantly, they adapt using volatility.
Bollinger views volatility as one of the core principles of markets, alongside things such as trend and mean reversion. His preferred measures include standard deviation, Average True Range and options-derived implied volatility.
One particularly neat idea from the interview combines two of them.
Wait until both Bollinger Bands contract inside the Keltner Channels, then look for the volatility breakout.
The logic is elegant. Bollinger Bands respond rapidly to changing standard deviation, while Keltner Channels use Average True Range. When the faster-moving Bollinger Bands contract completely inside the Keltner structure, you have an intuitive measure of unusually compressed volatility. I may have used this myself. ; )
Then wait.
Compression precedes expansion often enough to make the condition interesting. Direction still needs confirmation.
Bollinger also uses the bands to frame traditional price patterns. A classic example is a W-bottom where the momentum low occurs outside the lower band, the subsequent price low occurs inside it, and price then confirms the reversal.
Old-school chart pattern. Quantifiable volatility framework. Confirmation.
No contradiction there.
If It Takes 150 Lines, Be Suspicious
Perhaps my favourite observation from the entire conversation was this:
If a trading system takes two or three lines to describe, it might be very good. If it takes 150 lines, the odds are it’s terrible.
That captures something I’ve increasingly come to believe about quantitative trading. Full disclosure: the number of lines required depends on the coding language! Mine are not always short, but it doesn’t mean they aren’t intuitive and simple.
Complexity is easy to create. Understanding is harder.
Bollinger insists that his systems remain simple enough that he can understand intuitively how they should react when markets behave unusually. The same applies to indicators. If you’re using RSI, you should know what RSI actually calculates. If you’re using ATR, understand what goes into it.
Don’t treat indicators as coloured lines supplied by software.
Know the machinery.
Interestingly, this simplicity doesn’t mean intellectual stagnation. Bollinger was equally adamant that traders need new ideas. Markets evolve. Research has to continue. AI is changing the game. You can’t simply worship the classics and stop thinking.
And yet he finished our conversation recommending books and systems developed decades ago by Welles Wilder, Charles Patel, Norm Fosback and others.
Why?
Because old systems come with something researchers desperately want: enormous amounts of genuine out-of-sample data.
The lesson is not old beats new.
It is keep researching new ideas, while refusing to discard good old ones merely because they’re old.
Risk Is What We Get Paid For
The most thought-provoking part of the interview, for me, was Bollinger’s discussion of geometric growth.
If you genuinely possess an edge, what happens if you size positions to maximise its geometric growth rate?
Optimal f provides one framework. Bollinger isn’t suggesting everyone should actually trade at full optimal f. The resulting drawdowns and equity volatility may be intolerable.
His point is subtler.
Calculate it anyway.
If you decide to trade at half or quarter optimal f, understand precisely what return you’re giving up in exchange for that smoother ride.
Bollinger traces the intellectual debate back to the St Petersburg paradox. The argument has occupied mathematicians and economists for centuries, but it creates a very practical thought experiment for traders.
We have become slightly obsessed with making trading look like an annuity: smooth returns, tiny drawdowns, beautiful equity curves.
But markets don’t pay us for eliminating risk.
They pay us for accepting it.
There is therefore a price attached to every attempt to suppress volatility and drawdown. Diversification across genuinely uncorrelated systems can reduce that price considerably, but it doesn’t make the underlying trade-off disappear.
At the same time, Bollinger draws a very hard boundary.
His risk of ruin is zero.
He doesn’t use leverage.
Maximise growth, perhaps. Understand the cost of conservatism, certainly. But survival comes first.
Build Something That Can Survive Without You
Bollinger doesn’t want systems that require continuous optimisation.
He wants systems that work year after year.
When testing parameters, he’ll look around the chosen value. If a 20-day moving average works but 16, 17, 18, 19, 21, 22, 23 and 24 don’t produce broadly similar behaviour, the system gets thrown out.
He doesn’t optimise it.
That is a very different research mindset.
You’re not asking, How good can I make this backtest?
You’re asking, Is there actually something here?
Trend. Mean reversion. Volatility. Breadth. Relative strength.
These ideas persist because they’re connected to fundamental characteristics of markets rather than accidents in a dataset.
Bollinger even revealed what he jokingly called a “secret sauce” trading system during our conversation. I won’t spoil that one here. But fittingly, even that idea was conceptually simple.
That’s the pattern.
The Real Lesson
Trading isn’t a half-hour-a-week shortcut. Bollinger was unequivocal about that.
It’s work.
Skill takes time. Research never really stops. New ideas matter. Old ideas deserve another look. Indicators should be understood rather than merely applied. Systems should be simple enough to explain and robust enough that they don’t need perpetual repair.
And perhaps most importantly, we shouldn’t forget the gift sitting underneath all of this.
For equity traders, the market has historically placed the long-term odds in our favour.
Participate in that. It’s better than what a casino will give you.
Then carefully, patiently and intelligently try to improve upon it.
Key Takeaways
- Build around first principles such as trend, mean reversion and volatility.
- Favour simplicity and elegance over optimisation and complexity.
- Understand every indicator you use at a mechanical and intuitive level.
- Study both new research and old systems with genuine out-of-sample history.
- Understand the return you’re sacrificing when reducing risk, but keep risk of ruin at zero.
- Accept that becoming good at this requires work.
Simple systems. Robust ideas. Sensible risk. Lots of research.
After 46 years in markets, John Bollinger still seems remarkably enthusiastic about doing the work.
That may be the most useful lesson of all.
Hope you enjoy it, and please do Subscribe to the YouTube channel if you do, it really helps!
See you in the Collective, www.algoadvantage.io/collective
Simon