Blog
Technical, no-hype writing on algorithmic trading, backtesting methodology and Polymarket market mechanics.
Every article here is written to be checked, not just believed — real methodology, real numbers, no promises about returns. Filter by rubric or search to find what you need.
Articles: 36
LatestCopy tradingSeptember 2, 202612 min readDrew Shelem
The popular wallet-picking checklist measures the leader. Your question is a different one — is this a good trader after you've paid to enter? Why profitable and copyable are not the same thing.
September 2026 · 3
September 2, 202611 min readDrew Shelem
Waiting for their profit to sag means being too late. Early behavioural warnings, your own late-but-honest figures, and the two reactions that must not be confused: reduce trust, or exit.
September 2, 202618 min readDrew Shelem
Cold starts, exits, market orders, the crowd of copiers and free cash — the dozen places a beginner quietly loses the result even with a well-chosen leader and a correctly configured bot.
September 2, 202631 min readDrew Shelem
You copy the leader's choice, not their bet size. How to size positions yourself, and the four circuit breakers that keep one of someone else's all-ins from carrying off your deposit.
August 2026 · 1
July 2026 · 31
July 31, 202612 min readDrew Shelem
Your signal watches Binance spot, but the outcome is decided by the Chainlink aggregate at a preset instant — and near the strike these are different prices that determine win and loss.
July 31, 202611 min readDrew Shelem
The two hidden costs of scaling: correlated risk that's invisible in calm times, and trading against yourself, which under the venue's rules is a violation.
July 30, 202615 min readDrew Shelem
Your bot was profitable for months, then the drawdowns stopped being made up — and the fee, the spread, and the BTC chart are all the same. Did the edge die, or are you just unlucky? Here's how to separate them.
July 30, 202611 min readDrew Shelem
Why a confirmed edge can turn out to be a disguised bet on volatility — and why ordinary validation doesn't catch it.
July 29, 202612 min readDrew Shelem
The ruler that tells a modest real result from an impressive fluke — in both directions.
July 29, 202618 min readDrew Shelem
A map of strategies grouped by source of income, not by tactic — and an honest look at what's hard in each.
July 28, 202627 min readDrew Shelem
A map of three places where an edge has a reason — and why a signal from a search is almost always false.
July 28, 202612 min readDrew Shelem
Why a green "robust" verdict is not proof of an edge, and three things this test doesn't see.
July 27, 202618 min readDrew Shelem
The Kelly criterion, fractional Kelly, and why every uncertainty the whole cycle was about forces you to shrink the bet.
July 27, 202616 min readDrew Shelem
Why consecutive trades aren't independent — and how that makes your drawdown twice as deep at the same Sharpe.
July 26, 202615 min readDrew Shelem
Why a good backtest on held-out data can still fool you — and how to check.
July 26, 202621 min readDrew Shelem
The anatomy of the gap between a backtest and live trading on Polymarket's 5-15-minute crypto markets.
July 25, 202631 min readDrew Shelem
An alert threshold set at a round number like "minus twenty percent" carries no information about the state of your strategy. The only informative threshold is one derived from the drawdown distribution of that specific strategy — and it can be computed in advance, before you launch.
July 25, 20268 min readDrew Shelem
Averaging across a portfolio removes the spread and does not remove the bias. A hundred strategies, each slightly inflated, produce a portfolio inflated by exactly the same amount — while looking far more convincing than any one of them alone.
July 24, 202620 min readDrew Shelem
If you tried several variants of a strategy and kept the best one, its result is inflated — and the more variants you tried, the more inflated it is. So the number of variants tested has to be counted and published alongside the result. Without it the result cannot be assessed.
July 24, 202631 min readDrew Shelem
On a prediction market, return-to-risk ratios are derivative quantities. They blend forecast quality, entry price, and cost into a single number. Here those three can be separated, because the true outcome of every trade is known and a ready-made market forecast exists to compare against. Neither equities nor futures offer that.
July 23, 202624 min readDrew Shelem
If you read nothing else: the thing worth measuring is not how often you win. It is how far your forecast beats the price, in percentage points of probability, after the exchange takes its cut. Those two numbers can point in opposite directions, and this article is about why.
July 23, 20268 min readDrew Shelem
A trading bot does not make decisions. It repeats yours, quickly and without flinching.
July 11, 20264 min readQuant-Geek
Monte Carlo is one of the strongest tools for judging whether a strategy is robust or merely lucky. Backtesting shows what happened; Monte Carlo shows what could have happened under other plausible scenarios.
July 11, 20264 min readGogoboss
Markets constantly change regimes — trends turn into sideways movement, calm periods explode into high volatility. The purpose of filters is to allow the strategy to trade only when market conditions align with its statistical edge.
July 9, 20264 min readQuant-Geek
Overfitting (also known as curve-fitting or over-optimization) remains one of the biggest challenges in algorithmic trading. A strategy that performs exceptionally well on historical data often fails dramatically when deployed live.
July 8, 20265 min readQuant-Geek
Developing an algorithmic trading strategy is a process that requires not only a strong market idea but also meticulous work with parameters. Even a powerful concept often delivers mediocre or losing results in live trading due to poor parameter selection. This issue remains one of the central challenges in algorithmic trading.
July 7, 20266 min readQuant-Geek
Walk-Forward Optimization tests a strategy the way it would actually be traded — repeatedly re-optimising on one window of history and measuring on the window that follows. One of the strongest defences against overfitting.
July 7, 20265 min readR2D2
Momentum, quality and a multi-factor synthesis borrowed from equity factor models and adapted to event markets — decomposing a contract's return into drivers instead of guessing the outcome.
July 7, 20269 min readGuest Contributor
A no-hype guide to backtesting Polymarket's 5-minute Bitcoin up/down markets — why win rate misleads, how transaction costs quietly kill most edges, and what actually survives.
July 6, 20265 min readQuant-Geek
Two market states, two families of strategy: trend following and mean reversion — the logic of each, their opposite risk profiles, and how to tell which one the market is currently paying for.
July 5, 20265 min readQuant-Geek
In this article, we will break down why strategies that work perfectly in the past die in the future, and why ruthless "stress tests" on unseen data are the absolute only way to survive in the market.
July 4, 20266 min readQuant-Geek
Why Trading is About Calculating Probabilities, Not Predicting the Future
July 3, 20264 min readGogoboss
Polymarket's hybrid CLOB in practice: EIP-712 signed orders, off-chain matching with on-chain settlement on Polygon, the unified YES/NO book, and the edge cases that break a first bot.
July 2, 20263 min readGogoboss
Why Polymarket runs a real central limit order book instead of an AMM, how price-time priority actually fills your order, and what that changes about the way you place trades.
July 1, 20263 min readGogoboss
On a prediction market your breakeven win rate is your entry price, so a 75% win rate at $0.75 is a coin flip. The breakeven table, plus spread, slippage and API latency — the three quiet killers of an edge.
Press / to search, j and k to move between articles.