Reading the Bot Portfolio Tracker: The Three Numbers That Tell You a Strategy Is Breaking Down

Thirteen bots is a lot of green-and-red boxes to glance at every morning. Most traders running a full Push Button Trading lineup check one thing: did the account make money today. That habit works fine right up until a bot quietly stops being the bot you backtested — and by the time the daily P&L makes that obvious, you've already given back weeks of edge.

The Bot Portfolio Tracker exists to catch that shift earlier. It's not just a scoreboard — it's a per-bot and portfolio-wide diagnostic panel, and three specific numbers on it do almost all the early-warning work. Here's what they are, why they matter more than your equity curve, and how to read them without overreacting to normal variance.

What the Bot Portfolio Tracker Actually Shows You

Every bot in your lineup — ORB, Snapback, Scalper, 10 EMA, NR7, Volume Spike, and the rest — logs every trade against its own 4.5-year backtest baseline inside the Tracker. That gives you two views stacked on top of each other: a rolling "live" window (the last 20, 50, or 100 trades, your choice) and the historical sample that bot was built and tested on. The Tracker isn't useful because it shows you trades. It's useful because it puts recent performance next to the number that recent performance is supposed to resemble. A bot that's "working" is one where those two lines stay close. A bot that's breaking down is one where they've started to drift apart — and drift shows up in three places before it shows up in your account balance.

Number One: Rolling Win Rate vs. the Backtest Baseline

Every PBT bot ships with a documented backtest win rate — the Volume Spike bot, for example, wins on a different rhythm than a mean-reversion bot like Snapback. The Tracker plots your rolling win rate against that baseline, and the number to watch isn't the win rate itself — it's the gap.

A bot sitting 3-5 points under its historical win rate over a 20-30 trade window is normal noise; every strategy has cold stretches. What's not noise is a gap that holds for 40+ trades, or one that widens instead of mean-reverting. That pattern usually means the setup the bot was built to catch — an opening-range breakout, a volume anomaly, a mean-reversion snap — is showing up less cleanly in the current market than it did across the backtest sample. The fix isn't to panic and kill the bot on trade 25. It's to let the Tracker's window run long enough to separate a real regime shift from a bad two weeks.

Number Two: Average R-Multiple, Not Total P&L

This is the number traders skip because total P&L feels more concrete. But two bots can post the same monthly return with completely different health: one because it's winning the way it always has, the other because one oversized winner is masking a shrinking edge on every other trade.

The Tracker's average R-multiple (or expectancy-per-trade) line strips that out. If your average winner is getting smaller relative to your average loser — even while win rate holds steady — expectancy decays quietly until a single losing week erases a month of gains. This is exactly the metric that separates "the bot had a losing week" from "the bot is losing its edge," a distinction covered in more depth in how bots handle stops and position sizing while you're away from the screen. Watch the R-multiple trend over 30+ trades, not the dollar total over 30 days.

Number Three: Drawdown Depth and Duration Against the Historical Sample

Every bot's backtest has a worst-case drawdown — a specific depth and a specific length of time it took to recover. The Tracker plots your current drawdown against that historical worst case, and this is the number most traders misread in both directions.

Some traders panic the moment a bot is underwater at all, even when the current drawdown is well inside the backtest's documented range — that's the bot behaving exactly as designed. The real warning sign is a drawdown that exceeds the historical worst case in depth, or one that's taking meaningfully longer to recover than the backtest's longest recovery window. Sample size matters here too: five losing trades in a row on a bot with a small historical sample tells you less than the same streak on a bot with thousands of backtested trades behind it, a nuance covered in what 4.5 years of backtest data actually tells you before you risk a dollar and in how to tell if a strategy is overfit before you trade it live.

Reading the Three Together

Any one of these numbers moving on its own is usually just market noise — win rate dips, expectancy wobbles, and drawdowns happen inside a healthy backtest range all the time. What actually means something is two or three of them drifting in the same direction at once: win rate down, R-multiple shrinking, and drawdown pushing past the historical worst case. That combination, held over a long enough sample, is the Tracker telling you a bot's edge has genuinely changed — not just gone quiet.

When that happens, the next move isn't necessarily to shut the bot off for good. It might mean re-checking correlation with the rest of your lineup inside the Bot Portfolio Builder, cutting size while you watch for confirmation, or benchmarking it again against the question of whether one bot alone was ever a strategy in the first place. The point of the Tracker isn't to make that decision for you — it's to make sure you're making it off three real numbers instead of a gut feeling after a rough week.

Put the Tracker to Work on Your Own Lineup

The Bot Portfolio Tracker ships as part of the full Push Button Trading platform alongside all 13 automated NinjaTrader 8 bots, the Bot Portfolio Builder and Analyzer, and a Trade Copier that scales one bot lineup across up to 20 funded accounts. If you're currently trading on gut feel about which bot is "having a bad week," the 14-day free trial is the fastest way to see your own win-rate, R-multiple, and drawdown numbers next to the backtest baseline. Traders who want the full walkthrough — including how to set thresholds instead of guessing — can start with the 30-Day Bot Workshop for $199 before stepping into full membership.

Start at pushbuttontrading.co and let the Tracker tell you which of your bots still is what it was backtested to be.