Betfair Scalping: What It Actually Requires

Our overview of Betfair trading strategies introduces scalping in a few paragraphs, alongside swing trading, laying the draw and value-driven positioning. This piece leaves those aside and stays on scalping only.

Scalping is the technique most often demonstrated and least often sustained. The mechanics are simple enough to explain in two minutes. The conditions under which those mechanics produce a net positive result are narrow, and most markets do not meet them.

What a scalp is, in exchange terms

An exchange order book has two sides. Backers’ money sits on one side, layers’ money on the other, and between the best available back price and the best available lay price there is a gap — the spread. On a liquid market that gap is often a single tick.

A scalp is the act of getting filled on one side and closing on the other, one or two ticks away, and repeating that many times. The profit on each round trip is either the spread itself (posting on both sides and being matched at both) or a small oscillation in price (taking the offered price on entry, exiting after the market ticks in the chosen direction).

Two structural points follow from this.

First, a scalper is not forecasting the outcome of the event. The position is closed in seconds or minutes, and by the off there is normally no exposure at all. What is being forecast is the next one or two ticks of price movement, or nothing at all if the trade is pure spread capture.

Second, the position is closed by taking the opposite side at a different price, which leaves a position that pays the same regardless of the result. On a two-tick winner that flat position is worth a very small amount. That number is where the analysis has to start.

The tick ladder is not linear

The exchange price ladder does not move in constant increments. Between 1.01 and 2.00 the increment is 0.01. Between 2.00 and 3.00 it is 0.02. Between 3.00 and 4.00 it is 0.05. The increments continue to widen as prices lengthen, and at long prices a single tick is a substantial move.

The consequence matters more than the table. One tick at 1.50 is 0.01, which is about 0.67% of the price. One tick at 8.0 is 0.20, which is 2.5% of the price. The same phrase — “a one-tick scalp” — describes two entirely different trades depending on where on the ladder it happens.

This cuts both ways. Long prices offer more value per tick, but they also move by a tick more easily, and the noise that produces a winning tick produces a losing one just as readily. Short prices give tiny per-tick returns and require volume of trades to amount to anything, which in turn requires depth on both sides.

The arithmetic, with numbers

Take a back at 3.00 for 100 EUR, closed with a lay at 2.98 — one tick, in the 0.02 band.

To flatten the position (the green-up arithmetic is set out step by step elsewhere), the closing lay stake is (3.00 × 100) ÷ 2.98 = 100.67. The trade is now worth 0.67 EUR whatever happens. Commission on net market winnings — take 5% for the example — leaves roughly 0.64 EUR.

Two ticks, closing at 2.96: (3.00 × 100) ÷ 2.96 = 101.35, so 1.35 EUR gross, about 1.28 EUR after commission.

Now the other direction. The same entry, exited three ticks against at 3.06: (3.00 × 100) ÷ 3.06 = 98.04, leaving a loss of 1.96 EUR. Commission does not soften it; commission is charged on net winnings, so a loss simply means there are no winnings to charge on in that market.

One three-tick loser therefore erases three one-tick winners, or a little more than one and a half two-tick winners. And a three-tick adverse move is not an unusual event — it is a normal fluctuation in any market worth trading.

That ratio is the whole problem. A scalper taking one and two ticks needs to be right a large majority of the time and needs losing trades cut at one or two ticks, not three or four. The moment an exit is delayed in the hope of the price coming back, the arithmetic inverts.

The preconditions, and how each one fails

Depth on both sides

A scalp needs money waiting at the exit price before the entry is even placed. Without it, the trade is entered and then cannot be closed at the intended price — only at whatever is left further down the ladder, which is where the one-tick target becomes a three-tick loss.

Failure mode: thin book. The entry fills because someone wanted the other side; the exit does not fill because nobody does.

Oscillation without trend

Scalping is profitable in a market that moves up and down around a stable level. A market that trends turns every scalp into a directional position on the wrong side. A trader repeatedly buying dips in a steadily shortening price is simply being run over more slowly.

Failure mode: money arriving in one direction — team news, an injury, an in-play development filtering through. The oscillation stops and the ladder walks.

Queue position

An order at the best available price does not jump the queue. It sits behind everything already resting at that price. If a market has 4,000 EUR queued at the back price and an order for 100 EUR is added, that order is matched only after the 4,000 EUR ahead of it.

This produces a specific and nasty asymmetry: the order tends to fill when the price is about to move against it (because the queue is being cleared by aggressive money coming through) and tends not to fill when the price is about to move in its favour (because the queue never empties). Automated systems generally post faster and sit ahead in the queue.

Failure mode: adverse selection. The fills that arrive are the ones not wanted.

Reaction speed

Manual scalping means seeing a change and clicking. Latency between the exchange and the screen, plus human reaction time, plus the time to place and confirm, is enough for a one-tick edge to disappear.

Failure mode: the price on screen is history by the time the order lands.

Where scalping is even possible: our liquidity data

Across 39 days of collection between 31 July and 26 September 2026 — not consecutive, as the collection stopped between 7 and 25 September — we logged 6,688 distinct pre-match football and tennis markets (2,980 football, 3,708 tennis), captured four times a day.

The median volume traded per market was 41 EUR.

Not 41 EUR at a moment. Forty-one euros over the entire pre-match life of the market, as far as our last snapshot before the start. Above that median: 13.8% of markets exceeded 1,000 EUR, 6.2% exceeded 3,000 EUR, 2.6% exceeded 10,000 EUR and 0.8% exceeded 50,000 EUR. The largest single market observed traded 369,000 EUR.

The reading is blunt. The overwhelming majority of football and tennis markets cannot be scalped at all — there is no book to enter on and no book to exit into. The useful question is not how to scalp but which small minority of markets contains enough money for the technique to be mechanically possible.

Two caveats on our own figures. These are pre-match markets, and scalping is often done in the minutes around the off or in-play, where volumes are higher. And cumulative traded volume is not the same as depth at a given instant: a market that traded 10,000 EUR over three days may still show only a few hundred euros on each side of the book at any one moment. Cumulative volume sets an upper bound on what is plausible, nothing more.

Horse racing is the traditional venue for scalping, and the reason is usually given as concentrated liquidity in the last minutes before the off. We do not collect racing data. Nothing in this article about racing comes from us, and anyone relying on that claim should source it elsewhere.

One finding from our data that a scalper should read as a warning

On the first month of data, restricted to 247 markets that traded more than 3,000 EUR, we compared how selections performed against the probability implied by their own final traded price.

Selections that shortened by more than 3% won 85 times against 99 expected (z = −2.03; p = 0.04). Selections that drifted by more than 3% won 75 times against 61 expected (z = +2.20; p = 0.03). The control group — prices that moved less than 3% — showed nothing at all (z = +0.10).

The reading is that the market appears to overreact to pre-match money flow. The limits have to be stated plainly: one month of data, and that month was August, the start of the season, when teams are unsettled; several groups were tested, which erodes the value of a p of 0.04; and the calculation excludes exchange commission. It is a hint, not a proof.

For a scalper the relevance is not the betting implication. It is that liquid pre-match markets do trend: prices moved by more than 3% between our snapshots often enough to form groups worth testing. A move of that kind is exactly what converts a scalp into a directional loss.

Demonstration versus sustained result

Scalping is the strategy where the gap between a demonstration and a long-run outcome is widest. A recorded session of twenty winning trades proves little. On 100 EUR stakes around 3.00, twenty one-tick winners come to about 13 EUR after commission, and the recording says nothing about the losing trades that did not make it in. Even a trader who genuinely wins 80% of one-tick trades strings twenty together only about once in ninety attempts, so a clean recording is more likely to be selected than typical.

Published win rates are self-selected. Sessions that go badly are rarely recorded, and the ones that are recorded tend to be short. A win rate quoted without the average size of the losing trades is not information — the loss size is the variable that decides everything.

What separates outcomes over hundreds of trades is not market reading. It is the discipline to close a losing trade at one or two ticks, every time, including the times when the price would have come back. That is a behavioural requirement rather than an analytical one, and it is the part that most people cannot maintain when the position is real.

---

This content is provided for informational purposes only. It does not constitute betting advice or a recommendation to gamble. Gambling can be addictive; participation is restricted to those aged 18 and over.