Momentum explained for betting — Statometrics Academy

In one sentence

Momentum trading assumes a price that has started moving will continue in the same direction for a while, so you join the move and exit before it ends.

How it works

On the exchange, information arrives unevenly. Someone with early team news, a strong view or a better model starts backing a team. The price shortens, other traders notice, and more money follows. That chain can keep a steamer shortening for several minutes.

A momentum trader tries to get on early in the move and off before it stops. The signal is usually the recent change in price (in ticks or implied probability) over a fixed window, sometimes combined with traded volume to confirm that real money, not just shuffling orders, is behind it.

Momentum and mean reversion are opposite bets on the same price chart. Moves driven by information tend to continue; moves driven by a single large order tend to reverse. A momentum strategy is really a claim that you can tell which is which.

The maths

A simple momentum signal compares the current implied probability with its value a fixed time ago:

M_t = \frac{p_t - p_{t-k}}{p_{t-k}}, \qquad \text{follow if } M_t \gt m^{*}

and the break-even hit rate for a trade with a fixed target and stop is:

h^{*} = \frac{L}{W + L}

pₜ: implied probability now (1 ÷ price).
k: the look-back window, for example 10 minutes.
Mₜ: percentage change in implied probability over that window.
m*: your threshold for acting.
W: profit if the target is hit (after commission).
L: loss if the stop is hit.
h*: the fraction of trades you must win to break even.

In plain English: act when the price has moved enough, and know in advance how often you must be right for the trade to pay.

Worked betting example

Just after team news, the away side shortens from 8.0 to 6.0 in Match Odds on heavy volume. Implied probability has gone from 12.5% to 16.7%, a momentum reading of +33%.

Entry. Back £50 at 6.0.
Target. You aim to exit at 5.0. Lay £60 (50 × 6.0 ÷ 5.0). Profit £10 whatever the result, £9.80 after 2% commission. That is 10 ticks away (ticks are 0.1 between 5 and 6).
Stop. If the move stalls and the price drifts back to 7.0, lay £42.86 (50 × 6.0 ÷ 7.0). Loss £7.14 whatever the result. That is only 5 ticks away, because ticks are 0.2 above 6.
Break-even. h* = 7.14 ÷ (9.80 + 7.14) ≈ 42%.
Reality check. The trade is sensible only if, in your records, steamers of this size with this volume go on to reach the target before the stop more than about 42% of the time. That number must come from your own data, not intuition.

Where it's good

Pre-match steamers, often around team news, backed by rising traded volume and weight of money.
Drifters in the same conditions, traded from the lay side.
In-play football after a sustained spell of pressure, where the market adjusts in steps rather than all at once.
As a filter: avoid fading a move that has strong momentum and volume behind it.

Limitations and pitfalls

Late entry is the main killer. By the time a steamer is obvious, much of the move is done and you are buying from earlier traders.
Spoofing and bluffing exist: large orders can be placed to create an apparent move and then pulled.
Moves often end abruptly, especially just before kick-off.
The Betfair ladder is uneven; the same number of ticks means different £ amounts on either side of 6.0, which distorts risk and reward.
There is no reliable published evidence that simple momentum rules beat commission on the exchange; expect any edge to be small and fragile.
Short-term momentum and short-term mean reversion both appear in price data at different time scales; mixing them up reverses your results.

How to build it

pandas for rolling price changes and volume windows; lightgbm or logistic regression to predict "hits target before stop" from features.
Data: time-stamped prices, traded volume and order book depth from the Betfair stream API or historical data files.
Practical tip: label each historical move with its outcome (target hit, stop hit, neither by kick-off) and study the hit rate by price band and time to kick-off before risking money.

Mean reversion – the opposite trade on the same chart.
Moving averages and smoothing – crossover signals for trend strength.
Informed trader models – why informed money creates persistent moves.
Weight of money and VWAP – confirming that real money is behind a move.
Change-point detection – spotting when a new move has started.