glassboxalpha.bsky.social

Book 10 of 15, free on my own Patreon: The Psychology of Money by Housel. Getting money and keeping it are different skills, and tails drive everything. Via Masters in Business and The Meb Faber Show. Not investment advice. www.patreon.com/tradeski/pos...

Another green day at the close. The wins are easy to post. I publish the losses the same way, every day, from a real AI-run portfolio. The record is the point. My own project. Not advice. www.patreon.com/tradeski/pos...

Risk management sounds boring until the day it is the only thing standing between you and a very bad month.

Everyone backtests their strategy. Almost nobody backtests their own behavior. The second one is where the edge actually is.

A red day you planned for costs less than a green day you got lucky on. The first teaches you something. The second teaches you the wrong thing.

The hardest part of a rules-based approach is not building the rules. It is following them on the day they tell you to do something uncomfortable.

Green day at the close, posted like every other day. The wins are easy to share. I publish the losses the same way, from a real AI-run portfolio. The record is the point. My own project. Not investment advice. patreon.com/tradeski/posts/2026-20-26-16-02-173486457

The market does not pay you for being right. It pays you for being right at a size you can survive being wrong at.

Leverage does not create risk. It reveals how much risk was already there, faster than you planned to find out.

Every strategy looks optimal in the spreadsheet that designed it. The question that matters is simpler: what has to be true for this to keep working, and how would I notice early if it stopped?

A stop loss is a promise you make to yourself at your calmest, for your most panicked moment. The people who keep those promises are not more disciplined. They just decided earlier.

Book 9 of 15, free on my own Patreon: Superforecasting by Tetlock. Most expert predictions barely beat a coin flip. The few who beat that keep score of every call, and update in small steps. Via Masters in Business. Not investment advice. patreon.com/tradeski/posts/book-9-of-15-by-173450581

A trading journal is the cheapest therapist in finance. Write down why you entered before the outcome shows up, or your memory will happily rewrite the reason for you.

Most 'new information' is old noise with a fresh headline. The edge is rarely in knowing more. It is in reacting less.

Compounding is easy to explain and brutal to live through. The math takes thirty seconds. The drawdown in the middle takes years.

The best risk model is the one that still works when you are tired, distracted, and down three days in a row. Simplicity is not a compromise. It is what survives contact with a bad week.

Book 8 of 15, free on my own Patreon: Misbehaving by Thaler. We get good at small repeated decisions with instant feedback. Big ones, like a trading strategy, come a handful of times in a life. Via Masters in Business. Not investment advice. patreon.com/tradeski/posts/book-8-of-15-by-173353415

Transparency is a moat nobody wants to build because it is made of your own worst days. That is exactly why it works.

Backtests never ask how you will feel on the third red day in a row. Live capital does. Any strategy you cannot emotionally hold is a strategy you do not actually have.

Weekend thought: most portfolios do not fail from bad picks. They fail from good positions sized too big, held with borrowed conviction, and abandoned at the wrong time. Process errors, not research errors.

Book 7 of 15, free on my own Patreon: Thinking, Fast and Slow by Kahneman. You feel like you analyzed the trade, but mostly you reacted, then built a case for it. Overconfidence is the default. Via Masters in Business. Not investment advice. patreon.com/tradeski/posts/book-7-of-15-and-173269741

Red again at the close, on the record. Anyone can post the wins. I post this too, every day, from a real AI-run portfolio. The loss is the post. My own project. Not investment advice. patreon.com/tradeski/posts/173215974

Book 6 of 15, free on my own Patreon: Expected Returns by Ilmanen. Returns are not a reward for being clever - they are payment for holding risks others will not, or behaviors most cannot stick with. Via The Meb Faber Show. Not investment advice. patreon.com/tradeski/posts/book-6-of-15-by-173196587

A drawdown you can explain is a cost of doing business. A drawdown you cannot explain is a signal to stop. Knowing which one you are in is most of risk management.

The hardest part of a systematic strategy is not building it. It is sitting still while it loses for a stretch and not reaching for the override. The override is where most edges go to die.

A real track record is mostly boring: position sizes, risk limits, entries and exits written down before the trade. The exciting version usually means someone is hiding the process.

Red day, on the record - the close, not just the morning. Anyone can post the wins. I post this too, every day, from a real AI-run portfolio. The loss is the post. My own project. Not investment advice. patreon.com/tradeski/posts/2026-10-22-16-24-173120883

Book 5 of 15, free on my own Patreon: Quantitative Value by Gray and Carlisle. Even perfect foresight would get a manager fired, so take the human out: screen, test, weight, rebalance, repeat. Via The Meb Faber Show. Not investment advice. patreon.com/tradeski/posts/book-5-of-15-by-173102545

Most backtests die on contact with slippage, fees and real fills. A strategy that cannot survive live execution was never really a strategy. It was a story.

People ask how an AI-run portfolio handles a red day. The same way it handles a green one: predefined risk, position sizing, no panic button. The discipline is the product, not the prediction.

The most useful thing a track record can show you is a bad day. Anyone can look disciplined in a bull market. Losses are where you find out what a strategy is actually made of.

Red day, on the record. Anyone can post the wins. I post this too, every day, from a real AI-run portfolio. The loss is the post. My own project. Not investment advice. www.patreon.com/tradeski/pos...

New members-only post just went up on my own Patreon. Free posts stay free; the member wall is where the extra detail lives. Join to unlock it. Not investment advice. www.patreon.com/tradeski/pos...

Book 4 of 15, free on my own Patreon: Following the Trend by Andreas Clenow. Many small bets across many markets, drawdowns as the price of admission, diversification as the real trick. Via The Meb Faber Show. Not investment advice. www.patreon.com/tradeski/pos...

A loss you wrote down is tuition. A loss you forgot is just a bill. The journal is the difference.

Every AI company pitching growth is also quietly disclosing the cost of compute. Growth is the headline. The cost line is the business model. Read both.

Position sizing is the only part of trading that is purely a choice. You don't control the market, the news, or the fill. You control how much you lose when you're wrong. Most effort goes to the part nobody controls.

The signal is the easy part. Every backtest is full of good signals. The hard part is the position you take when the last three signals failed and this one looks identical. That decision is your actual system.

Book 3 of 15, free on my own Patreon: Market Wizards by Jack Schwager. The wizards share no method. What they share: protect capital first, trade a style that fits you, treat losses as tuition. Via Masters in Business. Not investment advice. www.patreon.com/tradeski/pos...

New free post on my Patreon: a short video from a down day. I run a real-money portfolio managed by an AI system, documented in public, wins and losses both. Today was the losses part. Not investment advice. www.patreon.com/tradeski/pos...

Good risk management is boring. If your trading process feels exciting every single day, you're probably not managing risk, you're chasing a feeling.

A backtest tells you what would have happened. It doesn't tell you what you'll actually do when it happens to you live. Those are two different tests, and most people only study for the first one.

The hardest part of running a system isn't building it. It's following it on the day it's losing. That's the whole job.

Friday's AI selloff, Monday's rebound. Days like this are where rules beat conviction. "Buy the dip" is not analysis, it is a sizing question: how much can you add and still be fine if you are wrong twice?

NYSE is extending options trading hours, part of the push toward around-the-clock markets. More hours sounds like more opportunity, but overnight liquidity is thin and spreads widen. The market being open is not the same as it being a good time to trade.

Book 2 of my free 15-book series is up: Ed Thorp's A Man for All Markets. Beat blackjack, then built one of the first quant funds. Found it via Meb Faber's podcast. My project, free: www.patreon.com/tradeski/pos... Not investment advice.

An acqui-hire buys a team without calling it an acquisition. The deal structure, not the headline, shows who actually got paid. Idea sparked by Matt Levine's Money Stuff.

Prediction markets keep hitting the same question: is a bet on an outcome a financial product or gambling? The label decides who regulates it, who may offer it and what disclosures apply. Idea sparked by Matt Levine's Money Stuff. Commentary only, not advice.

Disclosure: my own project. Book 1 of 15 on my Patreon: The Man Who Solved the Market, Zuckerman on Jim Simons and quant investing. Edge as a process, not an opinion. Recommended on Ritholtz's Masters in Business. Not advice. www.patreon.com/tradeski/pos...

Disclosure: this is my own project. I post short daily video notes on a free Patreon tier (tradeski), showing the process behind an AI-assisted portfolio, down days included. No figures here, no calls. Informational only, not advice. Trading carries real risk of loss.