What Is Paper Trading? (And Why Most of It Lies)

Paper trading is simulated trading. You enter orders the same way you would on a live broker, the orders fill against simulated market conditions, and you track P&L without real money on the line. Done right, it's a learning environment where execution mistakes are cheap and behavioral habits get built. Done wrong, it's a video game that teaches you to expect fills you'll never get live.

Most retail paper-trading platforms do it wrong. This is not an opinion. It's a statement about what their fill engines actually model and what they don't.

What "wrong" looks like

The most common shortcut retail paper platforms take is filling at mid-price. A market order to buy 1,000 shares of a stock with a bid of $42.10 and an ask of $42.14 fills the paper trader at $42.12 — the midpoint. The next click goes to mark-to-market and the trade looks instantly profitable by half the spread.

In live markets, that's not what happens. Your market buy hits the ask at $42.14. If the ask only had 500 shares available, the next 500 fill at $42.15 or higher as the order walks the book. You're starting the trade at the bottom of a small but real hole — the bid-ask spread plus any size impact — and the live broker rightfully shows it. The paper-trading platform that fills at mid hid this from you, and after 200 paper trades you've trained your brain to expect entry quality that doesn't exist.

The second common shortcut is no enforcement of regulatory halts. SSR (Reg SHO Rule 201, the uptick rule that triggers when a stock drops 10% from prior close) and LULD (Limit Up / Limit Down volatility halts) are two of the most common regulatory events a US-equity day trader runs into live. Paper platforms that don't enforce them let you short freely on a -15% day, let you trade through a 5-minute halt, and let your alerts fire on prices that wouldn't have been touchable in reality. You learn that the trade was easy. The next time it happens live, you discover that it isn't.

The third common shortcut is no borrow cost on shorts. Real shorts pay a daily carry equal to position value times the annualized borrow rate, divided by 365. Hard-to-borrow names — exactly the names retail short-squeeze traders are drawn to — can carry annualized fees of 50%, 200%, sometimes more. A 1,000-share short of a $4 stock at 150% borrow accrues over $16/day in carry, independent of whether the trade is profitable. Paper platforms that don't model borrow let you sit on a short for two weeks for free. You learn that you can wait shorts out. Live, you can't.

What "right" looks like

Realistic paper trading models the three shortcuts above directly.

Fills walk the order book. A market buy hits the best ask. If the best ask doesn't have the size you wanted, the rest of the order consumes the next level, and the next, until the share count is satisfied. The reported fill price is the volume-weighted average of the levels consumed. Partial fills happen when book depth is genuinely thinner than the requested size — and the paper engine returns the partial rather than topping you up at a price that doesn't exist. The methodology behind Tapeboard's L2 walk-book documents this end to end.

SSR and LULD are enforced server-side. When a stock drops 10% from prior close, SSR flips on and the paper engine refuses short fills below the last consolidated print. When a stock's 5-min VWAP band breaks, the paper engine halts trading for 5 minutes and resting limit orders sit until resume. The rejection codes match what a live broker returns, so the paper trader learns the same defensive habits they'll need on the real account. Full enforcement logic at Tapeboard's SSR + LULD methodology.

Borrow rate carry charges daily. Every paper short accrues carry at the published IBKR annualized rate divided by 365, applied each trading day. Hard-to-borrow names cost what they cost. Easy-to-borrow names pay the median. Either way, the paper trader feels the cost of holding a short and adjusts their hold-time expectations accordingly. Full math at Tapeboard's short borrow methodology.

The combined effect is that a paper trader who hits 50 profitable trades on a realistic engine has a much better chance of those skills surviving the transition to live trading than a paper trader who hit 200 profitable trades on a mid-price engine. The realistic engine is harder. That's the point.

Why retail platforms take the shortcuts

The shortcuts are computationally cheaper and emotionally more flattering. Mid-price fills don't require a live L2 book in the simulator — the platform doesn't need to subscribe to depth-of-book data for every paper account. SSR and LULD enforcement require server-side state tracking per symbol. Borrow rate carry requires nightly ingestion from a stock-loan feed and a daily mark-to-market routine on every open short position.

All three are real engineering work. A paper-trading product that's competing on user growth has the incentive to skip them, because users who lose less to slippage feel like they're learning faster, and a feature like "we charge you to short" is a hard sell in a free-tier marketing page.

The shortcuts also flatter the platform's marketing. "Paper traders made $X million in simulated profits this month" makes a better headline if the fills were at mid-price. The same paper traders running through an L2-walking engine would have made meaningfully less and the marketing number would be smaller. Most platforms pick the bigger number.

What you should actually look for

Three questions to ask any paper-trading platform before you commit to learning execution on it.

Question 1: At what price do market orders fill? If the answer is "at the mid" or "at the last print," the engine is not modeling real spread cost. If the answer is "by walking the order book until the size is satisfied," the engine is doing real work.

Question 2: Does it enforce SSR and LULD? If the platform can't answer this, or answers with marketing language, the engine doesn't enforce them. Real enforcement is a yes/no question with a server-side implementation behind it.

Question 3: Does it charge borrow rate carry on shorts? Same yes/no. If the answer is "we model fees in a generic way," that's a no with extra steps. Real carry is a daily charge against open short positions sourced from a stock-loan feed.

A platform that answers yes to all three is doing the work. A platform that hedges on any of them is not. The Tapeboard paper-trading simulator is one of the platforms that answers yes to all three; the methodology pages linked above document the implementation. Pick whichever platform fits your workflow, but ask the questions first.

What paper trading still can't teach you

Even a perfect simulator has limits. Two big ones.

No real-money fear. Paper trading has no skin in the game. You will not feel the same emotional pressure on a paper drawdown that you'll feel on a live drawdown of the same size. This is not fixable — it's structural. The fix is to size your first live trades small enough that the dollar amounts don't trigger the fear response, then scale up as confidence builds.

No fill-time anxiety. A paper trader who hits buy at the open never has to wait the 200ms while the order routes to the exchange and gets confirmed. Live fills come back with latency, the latency varies, and the trader sometimes second-guesses the entry while the fill is in flight. Paper simulators that snap to instant fills hide this.

Both of these gaps are real. Neither is solvable by improving the simulator. The fix is the third leg of the paper-to-live transition: micro-sizing live early so the gap between paper and live closes incrementally, not all at once.

The simulator is a learning environment. Done right, it teaches you the mechanical pieces of trading — fills, halts, borrow costs, position tracking, risk management — in a place where the cost of being wrong is zero. Done wrong, it teaches you bad habits and gives you false confidence. The difference between the two is whether the engine takes shortcuts.