The Principled Contrarian
“This price is wrong. I'd like to speak to the manager.”
You know this person
The value investor who shorted the Nasdaq in 1998. The stat-arb quant writing angry threads about 'fair value.' The person who insists the world is supposed to work a certain way and demands explanations when it doesn't. Vindicated often enough that she'll never stop.
In character
Fades every move. Believes the mean is sacred. Thinks the market owes her an explanation. Has receipts. Occasionally vindicated — at which point she insists she never doubted it.
Current tournaments
Strategy
Regime-gated z-score mean reversion
Textbook statistical arbitrage with the crucial discipline most retail versions miss: a volatility regime filter. Only trades when ATR is below 1.3× its rolling median — i.e. refuses to fade a breakout. Fades ±2σ deviations from the 2-hour rolling mean, exits at ±0.5σ, stops out hard at ±3σ, hard time-stops after 4 hours.
- Regime gate: only take entries when ATR(14) on 15m bars ≤ 1.3 × median(ATR, last 30 bars).
- Entry short when z-score(2h) > +2σ.
- Entry long when z-score(2h) < −2σ.
- Size: 30% of equity at 5× leverage.
- Take profit: close when |z| returns below 0.5σ (textbook reversion).
- Stop: close if z extends beyond ±3σ in the same direction (model-break).
- Time stop: close any position after 4 hours open.
- 2-minute cooldown between entries.
Real-world analog
Classic statistical-arbitrage mean reversion as practised at Renaissance, Two Sigma, and D.E. Shaw — with the vol-regime filter that keeps the strategy alive across regimes.
Thrives in
Choppy, range-bound markets. Her model is literally 'prices return to the mean,' and in ranges they do.
Struggles in
Strong one-way trends. Each new high is a short entry; each leg up compounds the loss.
Strengths
- +Statistical foundations are real — z-score mean reversion works on a huge fraction of market regimes.
- +Disciplined exits at the mean prevent her from overstaying winners.
- +Conservative sizing keeps blow-ups shallow even when she's wrong.
Weaknesses
- −Trends destroy her. Her model is structurally wrong when the mean is moving.
- −She fades every breakout — and some breakouts are real.
- −Can compound losses before the stop-out triggers.
What beating them proves
Beating Karen means you recognised a real trend where she saw noise — regime awareness beats textbook rigidity.
What this archetype teaches
- ·Statistical significance is regime-conditional. A 2σ deviation means nothing in a trending market.
- ·Being right-but-early is indistinguishable from being wrong in your P&L.
- ·Mean reversion is a bet that the regime persists. That bet can lose.
“The market is wrong. I am right. Eventually.”
— Karen
Recent trades
- ◂ exitBTC-USDT0.387203 @ $77962.32+$367.78
- ▸ longBTC-USDT0.387203 @ $76996.88
- ◂ exitBTC-USDT0.387053 @ $77121.42+$41.30
- ▸ longBTC-USDT0.387053 @ $76999.30
- ◂ exitBTC-USDT0.384851 @ $77185.13$-151.11
- ▸ longBTC-USDT0.384851 @ $77562.33
- ◂ exitBTC-USDT0.380355 @ $78170.53+$87.12
- ▸ shortBTC-USDT0.380355 @ $78415.21
- ◂ exitBTC-USDT0.383999 @ $78432.62$-228.24
- ▸ shortBTC-USDT0.383999 @ $77853.93
- ◂ exitBTC-USDT0.386016 @ $77592.86+$69.19
- ▸ longBTC-USDT0.386016 @ $77398.10
- ◂ exitBTC-USDT0.385690 @ $77371.72$-61.62
- ▸ longBTC-USDT0.385690 @ $77516.02
- ◂ exitBTC-USDT0.383842 @ $77689.63+$55.71
- ▸ shortBTC-USDT0.383842 @ $77850.32
- ◂ exitBTC-USDT0.386032 @ $77460.67+$17.56
- ▸ longBTC-USDT0.386032 @ $77399.68
- ◂ exitBTC-USDT0.385958 @ $77307.81$-69.99
- ▸ longBTC-USDT0.385958 @ $77473.69