Death by a Thousand Cuts

Retail traders evaluating AI quant tools obsess over the win rate. They look at the 70% accuracy claim and start planning their retirement. They ignore friction.

Friction is the difference between the theoretical return of a model in a vacuum and the actual return of a human executing it in a live brokerage account. It consists of three main components: Subscription Fees, Slippage, and Taxes.

1. Subscription Fee Drag

As we explored in our Fee vs. Alpha Calculator, fixed monthly costs are devastating to small accounts. If you pay $100 a month on a $5,000 account, you start every year down 24%.

The math is brutal. You don't just need the AI to beat the market; you need it to beat the market by 24% just to match a free index fund.

2. Slippage: The Illusion of the Perfect Fill

Backtests assume you can buy a stock at exactly the price the signal was generated.

If Tickeron issues a buy signal on a micro-cap stock breaking out, thousands of other users (and high-frequency algorithms) see the same signal. By the time your market order reaches the exchange, the price has jumped 5 cents.

Entry Slippage: Buying higher than the signal price.
Exit Slippage: Selling lower than the signal price (especially during a panic sell).

The Slippage Math

Assume you trade a $50 stock, buying 100 shares ($5,000). You suffer a realistic $0.05 slippage on entry and $0.05 on exit.

That is $10 lost per trade to friction. If your AI model is a high-frequency system making 10 trades a week, that is $100 a week, or $5,200 a year lost purely to bad fills.

3. The Tax Man Cometh (Short-Term Capital Gains)

AI models are usually hyper-active. They hold stocks for hours, days, or weeks. This means every profit is taxed as ordinary income (short-term capital gains), which can be up to 37% in the US, plus state taxes.

Holding the SPY ETF for over a year incurs long-term capital gains (usually 15-20%).

To justify using an active AI system over a passive ETF, the AI must out-earn the ETF after paying an extra 15-20% penalty on all its wins. Check out our Tax Guide for more details.

The Conclusion

Before buying a subscription to Trade Ideas or Danelfin, you must run your expected friction math. A model boasting a 30% annual return often translates to a 5% net loss for a retail trader once fees, $0.05 slippage, and short-term taxes are applied.

Monitor our Model Portfolio to see how we apply rigid friction penalties to our live tracking.