Ever wonder what a really old chart looks like? Go back far enough and the data is really spotty. In antiquity, price changes were recorded when they were large, not on a regular schedule. (This has a mirror in point and figure charting).
Weekly or monthly series are also common, and daily series, where they exist, are usually close-only. Academic researchers have done a lot of work cleaning up these old price records, and there's probably a bit of guesswork involved.
No one is suggesting quant backtests on this data (I hope). But the bones are there--enough to see that the way humans have interacted with risk and opportunity in financial markets has not really changed, in literally all of recorded history.
Working through some of these older charts in Askesis today, a testing platform I've developed for blended discretionary/quant tests. Have a look at a chart from 1714:
I'm going to publish some of the statistical work supporting the discussions I had with @WaqarAsim10 on the @TOT_podcast_ this week. Stay tuned! I'll post to my blog and also link here.
5
55
2,411
Adam Grimes retweeted
Adam on testing Fibonacci levels against random lines:
He wrote The Art & Science of Technical Analysis.
"We draw a bunch of random lines on the chart, and we test those random lines against the Fibonacci levels. They should look different.
What I have found overwhelmingly, and this is where people probably start throwing things at the screen or spiking their phone on the ground: Fibonacci levels are not special. Moving averages are not special. VWAP is not special.
If they don't look different, I don't believe that something can be really significant in the market and also invisible in the data."
His test is one any trader can run.
"The tape is full of useful information, but it's also full of deception.
I've been the guy trying to screw OrderFlow readers over, I know the tricks to do that!"
Adam Grimes (@AdamHGrimes), the man who wrote The Art & Science of Technical Analysis, on what the order-flow tape hides, why Fibonacci levels test no better than random lines, & the one level his data says is real.
Grimes has traded professionally since 1995, was a Chief Technical Strategist at SMB Capital, and he's done extensive quantitive research in price data as far back as barley prices from the 1300s.
"I don't believe that something can be really significant in the market and also invisible in the data."
We cover:
- Order flow: "there's also a tremendous amount of deception on the tape", and the tricks he says he used when he had to execute a lot of volume in thin stocks
- His test: "we draw a bunch of random lines on the chart, and we test those random lines against the Fibonacci levels"
- His verdict: "Fibonacci levels are not special. Moving averages are not special. VWAP is not special."
- Why stock indexes bounce at a moving average: "But guess what? So do our random lines"
- "The 200-day moving average is not a magical line in the sand"
- The one level he says behaves differently: "previous day's high and low, that's a real thing"
- Why a trader at home can't compete with Jane Street: "the length of a copper cable, the speed of light through that cable matters"
- Where he thinks a trader at home has the better chance: "two to five days is probably a sweet spot for a retailer"
- Stacking confluences: "it's like me calling the same person six times"
- The pattern he put first in his book: "this idea of penetrating a level and then failing"
- "I believe we are at an inflection point in history, full stop."
- The wrong way to use AI in trading, and the way he uses it: "the AI is driving my code"
- The "timeframe justification system": "risking a few hundred dollars and losing tens of thousands of dollars"
- Why "you can greatly accelerate alpha decay by publishing"
It was a pleasure @AdamHGrimes to host you on Titans of Tomorrow.
Highlights:
00:00 Intro
03:23 The pullback: a principle as old as financial markets
05:05 Why a trader at home can't compete with Jane Street
07:31 The holding period he calls a sweet spot for a retailer
11:03 An inflection point in history
15:07 Testing Fibonacci against random lines
17:30 Fibonacci, moving averages, VWAP: "not special"
21:03 The 200-day moving average is not a magical line
24:07 The previous day's high and low
30:30 The two forces he builds on: momentum and mean reversion
32:58 Order flow, and the deception on the tape
36:44 Why he walked away from day trading
44:19 Why a pretty chart is no judge of its outcome
49:58 Why stacked confluences are one signal
55:05 How he kills a setup
59:55 The failure test: the first pattern in his book
1:07:24 The timeframe justification system
1:13:20 The wrong way to use AI in trading
1:26:52 Publishing and alpha decay
Adam Grimes retweeted
@LanceBeggs book and nifty nirvana blogs were my first price action reads and till date didn't need anything else. Lance beggs book has quite Mechanical rules in case anyone is interested
This Post is from an account that no longer exists. Learn more
I spend a lot of time figuring out how to present data in a way that has the potential to spark some intuition. I find this to be a useful way to understand sector shifts. Spend some time with the options and see if something clicks for you. marketlifetrading.com/sector…
I haven't been on X much lately. My feed is full of people asking Grok for trading techniques, backtests, or market analysis. This is basically bringing a spoon to a gunfight.
Is this as common as my feed makes it seem? How did this become a thing??
13
28
4,220
Poor market breadth predicts crashes. At least, that's what you've been taught.
I tested 9 breadth measures against 35 years of market data, including subsequent returns and major declines. The conventional wisdom doesn't hold up. And something changed around 2010.
The Breadth Myth: What Market Data Really Says
adamhgrimes.com/the-breadth-…
Quiet markets tend to stay quiet.
Big volatility shocks tend to cool off—but remain unusually volatile. And downward moves in stock indexes contain information about what comes next.
I tested some very old trading assumptions against the data.
Volatility Shocks and What Follows
adamhgrimes.com/volatility-s…
Volatility has a memory.
Quiet markets tend to stay quiet. Volatile markets tend to stay volatile. Until they don’t—and those transitions are where things get interesting.
A look at ATR, volatility persistence, distributions, and what the data actually says:
adamhgrimes.com/volatility-h…
The goal isn’t to find the perfect diet, training program, or health protocol.
It’s to build a process that notices when you’re wrong and keeps adapting.
That idea came straight from trading.
Follow-up to “How Trading Saved My Life”:
adamhgrimes.com/the-experime…
4
20
1,904
How Trading Saved My Life
adamhgrimes.com/how-trading-…
I’ve never written on 9/11, and I’m not really going to start today.
This day has always been strange, heavy, difficult.
5
25
4,597
Or maybe simply because it’s a beautiful and haunting work of art that touches the heart of terror and loss.
The poem: poetryfoundation.org/poems/4…
The poet reading it: youtube.com/watch?v=H3anIr7v…
9
3,978
Maybe because, even in the brief years I spent in that industry, I saw some of the people I knew in this poem. Maybe because of the musicality of the language.
6
2,274
Years ago I stumbled across a poem I’ve returned to many times.
Maybe because it remembers people mostly forgotten. Maybe because, had I stayed in the business and built a career cooking professionally, those would have been my people.
1
7
2,623







