Betfair trading is profitable only for traders whose edge beats everything a trade costs, and the arithmetic of the exchange works against most people who try it. No official figure counts how many traders finish ahead, and there is no guarantee of profit. In an illustrative month of 400 trades, winning 60% of them keeps $102.40 after an illustrative 6% commission, about $2.56 for each of 40 hours, while winning 57% of them finishes behind.
A trade costs commission on winning markets, the spread on the way in and out, any extra charges the exchange places on some winning customers, and the hours it takes. The Interactive Gambling Act (s 8A(3) and s 10B) limits online sports betting to before the event begins, so both halves of a sports trade go on before the start. Every racing trade on this page is closed before the jump as well.
A month of trading, costed
An exchange trading profit has to survive every cost. With illustrative figures: 400 trades in a month, one per market, averaging +$4.00 on winners and -$5.00 on losers before commission, over 40 hours at the screen.
| Line | Working | Amount |
|---|---|---|
| 240 winning trades (60%) | 240 x 4.00 | +$960.00 |
| 160 losing trades | 160 x 5.00 | -$800.00 |
| Trading result before commission | 960 - 800 | +$160.00 |
| Commission at an illustrative 6% | 960 x 0.06 | -$57.60 |
| Net for the month | 160 - 57.60 | +$102.40 |
| Per hour at the screen | 102.40 / 40 | $2.56 |
Commission comes off each winning market's net winnings, so with one trade per market it is 6% of every winning trade and nothing on a loser. The break-even win rate is loss / (net win + loss) = 5.00 / (3.76 + 5.00) = 57.1%, against 5.00 / 9.00 = 55.6% with no commission at all.
How thin that margin is:
| Win rate | Before commission | Commission at 6% | Net |
|---|---|---|---|
| 60% (240 winners) | 960 - 800 = +$160.00 | $57.60 | +$102.40 |
| 59% (236) | 944 - 820 = +$124.00 | $56.64 | +$67.36 |
| 58% (232) | 928 - 840 = +$88.00 | $55.68 | +$32.32 |
| 57% (228) | 912 - 860 = +$52.00 | $54.72 | -$2.72 |
| 56% (224) | 896 - 880 = +$16.00 | $53.76 | -$37.76 |
Three points of win rate, 12 trades in 400, separate a winning month from a losing one.
Commission is the exchange's to set and to change, so every rate here is an input; re-run the sums whenever yours moves. With the same 400 trades at 60% winners:
| Commission rate (illustrative input) | Net for the month | Break-even win rate |
|---|---|---|
| 4% | 160 - 960 x 0.04 = +$121.60 | 5.00 / (3.84 + 5.00) = 56.6% |
| 6% | 160 - 960 x 0.06 = +$102.40 | 5.00 / (3.76 + 5.00) = 57.1% |
| 8% | 160 - 960 x 0.08 = +$83.20 | 5.00 / (3.68 + 5.00) = 57.6% |
| 10% | 160 - 960 x 0.10 = +$64.00 | 5.00 / (3.60 + 5.00) = 58.1% |
For how the exchange works out commission on a whole market's net winnings, see exchange commission explained.
The spread, extra charges and your time
- The spread. It has no line of its own because it sits inside the $4.00 and $5.00 averages. With illustrative prices, backing $69 at $3.40 and laying straight back at $3.45, one tick higher, means a lay of 69 x 3.40 / 3.45 = $68.00 and a loss of $1.00 either way. If, as an illustration, half of the 400 trades paid one tick of spread like that, the spread took 200 x 1.00 = $200 out of the averages. Without it, the same trades would have made 160 + 200 = $360 before commission.
- Extra charges. The exchange's terms can add charges for some winning customers. The exchange decides what triggers them, what they are called and how much they take, and it can change all three. Read its current terms and model the charge as an input: an illustrative 20% charge on a $500 result leaves $400. Because it falls on winning accounts, it cuts into exactly the result a trader is working towards.
- Software and time. An illustrative $60 a month for software or data leaves 102.40 - 60 = $42.40, or 42.40 / 40 = $1.06 an hour.
Risk: Betting involves risk. Every figure in these tables is illustrative, a small change in win rate or commission turns the month negative, and there is no guarantee of profit. See responsible gambling for limits and support.
Why the average trader loses or only breaks even
Nobody publishes an official count of exchange traders who finish ahead. The reasons to expect a loss, or a result close to zero, come instead from how the exchange works:
- Each matched bet pairs someone who backed with someone who laid. One side's win is the other side's loss, and commission comes off the net winnings of whoever finishes a market ahead, so a trader's profit is paid out of other customers' losses, less the exchange's share.
- A trader with no edge loses on average on every trade: the spread on any entry or exit that takes a price, plus commission on the trades that win.
- Small wins and bigger losses. Closing winners after a tick or two while letting losers run makes the average loss bigger than the average win, which pushes the win rate needed well past 50%, as the $4.00 and $5.00 averages do.
- Faster competition. Some of the money on the ladder is placed by software, as the next section explains.
- Short records mislead. A trader with no edge at all still has winning months, and a winning month feels like proof.
Take a trader exactly at break-even in the costed month: 57.1% winners, netting $3.76 on each winner and losing $5.00 on each loser, a swing of $8.76 between the two. Over 400 trades, chance alone moves the month's result by about 8.76 x square root of (400 x 0.571 x 0.429) = 8.76 x 9.90 = $86.72 either way for one standard deviation.
So roughly 95% of such months land within 1.96 x 86.72 = about $170 of zero. A month of +$102.40 sits well inside that range and says nothing about skill on its own. Sample size in betting shows how many results it takes to tell.
Competition from automated traders
Exchanges let customers connect software through an API, on conditions each exchange sets in its terms, so some of the money on the ladder is placed by programs rather than people. A program can place, move and cancel offers faster than anyone can click, across many markets at once.
That bites hardest on the small, quick trades a manual trader is most likely to try:
- Queue position. Offers at one price are matched in the order they arrived, and software that re-places an offer as soon as a price changes tends to sit nearer the front.
- Which offers get taken. A waiting offer is most likely to be taken when someone with better or faster information wants that side, which is often just before the price moves against it.
- News. A scratching or a sharp bookmaker move is priced in by the fastest participants first.
None of that makes trading by hand impossible, but it means a manual trader's edge has to come from somewhere speed does not decide, such as a reason for a move that you can name and test. Betting exchange bots explains how exchange software connects and what can go wrong with it. Before connecting software of your own, check what the exchange's terms allow; many bookmakers restrict automated betting in their terms.
What Betfair trading success takes: discipline and records
No official source sets out what profitable exchange traders do, but those costs set the requirements any trader has to meet:
- An edge you can state in one sentence and test, such as a pattern in how bookmaker moves reach the exchange on the races you trade.
- Rules written before each trade: entry, target, stop, stake and a deadline before the jump or the start.
- Stakes sized from what the stop costs, and a daily loss limit you keep.
- A record of every trade: date, market, minutes before the start, entry and exit prices, ticks, stake, result, commission and time spent.
- A monthly check of that record against your activity statement. The National Consumer Protection Framework requires Australian online wagering providers to email active accounts a monthly statement of wins, losses and the net result, and to make your transaction records from at least the last 7 years available.
- Enough trades before judging. Given the $170 range chance alone can produce, one month says little either way.
- A stopping rule. If the record after every cost is flat or negative over a fair sample, stop rather than raise stakes.
Risk: Betting involves risk. Good records show whether a method is working; they do not make it work, and past results are no guarantee of future results. See responsible gambling for limits and support.
Tax on trading profits
The ATO's starting point is that most punters' winnings are outside their assessable income, and their losses cannot be deducted, unless betting is carried on as a business. Its ruling IT 2655 says each case depends on its own facts, and the main tests it records from the courts include how systematic, organised and businesslike the betting is. It also notes that betting with a significant element of skill is more likely to have tax consequences than betting on random events. The professional gambler tax guide explains each test.
Trading is usually more systematic than ordinary betting, so those are the tests to read, but whether they apply depends on your own facts. This is general information, not tax advice: ask the ATO or a registered tax agent about your situation. Tax on gambling winnings in Australia covers the ATO's position in full.
When trading is the wrong choice
Stop, or do not start, when:
- your records after every cost sit inside the range chance alone would produce;
- the hours are coming out of work, sleep or family, or are worth more to you than the result;
- you are trading to win back losses, or with money you need for something else;
- someone is selling you a trading course or system with income claims. Schemes sold as sports trading are among the sports investment schemes Scamwatch warns about: the ACCC's scam service calls them just another form of gambling, not an investment, and often outright scams (checked October 2026).
For free and confidential support call 1800 858 858 or visit gamblinghelponline.org.au.
What B337 can and cannot add
B337 does not trade, back or lay on the exchange, and nothing in it makes a trade profitable. Its Terminal shows Betfair's back and lay prices beside prices from 40+ bookmakers across racing and sports, with flucs, so you can set bookmaker moves against the exchange price while you test an edge such as the bookmaker-move pattern above. Bookmaker names are trade marks of their owners. B337 is not affiliated with them.
On racing, it also shows closing lines, the last prices before the jump. Logging the close beside each entry price tests a method sooner than profit and loss alone can, which is the idea behind closing line value. The Terminal reads prices on a repeating cycle, so a closing line on screen can trail the last price the exchange or bookmaker actually showed; treat it as a yardstick for your log.
How laying and commission work is set out in the lay betting hub, and the mechanics of a single trade in Betfair trading explained.
Risk: Betting involves risk. A run of entries better than the closing price can still end in a loss, Terminal prices can lag the exchange's, and there is no guarantee of profit. See responsible gambling for limits and support.