ROI vs yield in betting comes down to what you divide your profit by. Yield, also called profit on turnover (POT), divides profit by the total amount you staked. Betting records and trackers often use ROI to mean the same thing, while many explainers define ROI as profit over the starting bankroll, which gives a bigger number whenever the bank has been staked more than once over. With illustrative figures, $500 profit on $20,000 of turnover is a 2.5% yield, and on a $2,000 starting bank the same $500 is a 25% return on bankroll.
Use yield to judge how well a method bets, and the bankroll figure only to see how hard a bank worked. Never compare one person's yield with another person's bankroll figure, and never read either one without the number of bets behind it.
The formulas for yield, ROI and return on bankroll
Profit is everything that came back from settled bets, stakes included, minus everything staked on them, after exchange commission. Betting turnover is the total staked on those same bets. Then:
- yield (profit on turnover) = profit / turnover x 100
- ROI on bankroll, or return on bankroll = profit / starting bankroll x 100
- return on turnover = total returned / turnover x 100, which is the yield plus 100
| Measure | Divides | What it answers | Do bigger stakes change it? |
|---|---|---|---|
| Yield, POT, or ROI as most betting records use it | Profit by turnover | What each dollar staked earned on average | No, at level stakes |
| ROI on bankroll | Profit by the starting bank | How much the bank grew | Yes |
| Return on turnover | Everything returned by turnover | The same as yield, plus 100% | No |
The third measure causes the most confusion. Getting $20,500 back from $20,000 staked is a 102.5% return on turnover and a 2.5% yield: nobody doubled their money. When a record quotes a figure above 100%, check whether it is a return or a profit before reading anything into it.
Worked example: a 2.5% yield on $20,000 of turnover
Example: Illustrative figures, not a real record: 400 settled bets of $50 each, from a starting bank of $2,000.
| Line | Working | Result |
|---|---|---|
| Turnover | 400 x $50 | $20,000 |
| Total returned | Stakes back on winners, plus winnings | $20,500 |
| Profit | $20,500 - $20,000 | $500 |
| Yield | $500 / $20,000 x 100 | 2.5% |
| Times the bank was turned over | $20,000 / $2,000 | 10 |
| Return on bankroll | $500 / $2,000 x 100 | 25% |
The two figures are tied together: return on bankroll = yield x (turnover / starting bank) = 2.5% x 10 = 25%. A punter who says 25% and one who says 2.5% can be describing the same 400 bets. The bankroll figure is ten times bigger because the bank was staked ten times over, not because any bet was better.
Why yield compares betting methods fairly
Turnover-based figures compare methods fairly because yield does not move when stakes are scaled. Place the same 400 bets at $100 instead of $50 and turnover doubles to $40,000, profit doubles to $1,000, and the yield stays at 2.5%. Return on the same $2,000 bank doubles to 50%, for the same selections at the same prices. A bankroll figure rewards the risk taken; yield measures the betting.
Two illustrative records, both started with a $1,500 bank, show how far the two measures can point in different directions:
| Method | Bets | Turnover | Profit | Yield | Return on bankroll |
|---|---|---|---|---|---|
| A | 80 at $25 | $2,000 | $240 | 12% | 16% |
| B | 1,500 at $20 | $30,000 | $900 | 3% | 60% |
Method B grew its bank more because it turned it over 20 times. Method A earned more per dollar staked, but 80 bets cannot tell a 12% yield from luck, and how many bets that takes is the job of sample size in betting.
Recompute at level stakes
A yield from varying stakes is weighted toward the big bets. Take an illustrative record of 99 bets at $10 that lost $50 between them, plus one $500 bet at $3.00 that won $1,000. Turnover is $1,490, profit is $950, and the yield is 950 / 1,490 = 63.8%. Stake that last bet at $10 like the others and it wins $20: profit is -$30 on $1,000 of turnover, a -3% yield.
One staking decision made that record look strong. Restate any record at one stake size, or in betting units, before you compare it with another.
What counts as turnover and profit
Most arguments about a yield are really about what went into it. With illustrative amounts:
| Case | How to count it | Why |
|---|---|---|
| Each way bet | Both halves: $10 each way is $20 of turnover | It is two bets |
| Void, scratched or refunded bet | Leave it out of turnover | The stake came back, and counting it dilutes the yield |
| Bet not yet settled | Leave it out until it settles | A stake with no result is not a result |
| Dead heat or deduction | Full stake in turnover, the reduced amount in returns | The whole stake was at risk |
| Exchange commission | Take it off profit | It is a cost of the bet |
| Multi | One stake, counted once | It is one bet, however many legs it has |
Lay bets need a choice. Lay a runner for a $40 lay stake at $4.50 (illustrative) and your lay liability is $40 x (4.50 - 1) = $140. If the runner loses, you win $40 before commission: 100% of the lay stake, but 40 / 140 = 28.6% of the money you had at risk. Liability sits more fairly beside back bets, and whichever base you use, say so beside the figure.
Bonus bets need their own line. A bonus bet costs no cash, so adding it to turnover at face value drags your cash yield down, and leaving it out while counting its winnings lifts the yield on money you never staked. Keep cash bets and bonus bets as two separate records.
Your bookmaker's monthly activity statement, which the National Consumer Protection Framework has required for active online accounts since 31 July 2022, shows the amount spent and the net win or loss for the month. It is a useful cross-check for one account, but check what each line counts before dividing one by the other, because refunds and bonus bets may sit in either.
Betting activity statements explains how to read one, and how to track betting results lists the fields worth recording for every bet.
What a realistic yield looks like
A realistic yield is a sample question, not a number anyone can promise. Two facts frame it.
First, zero is not the baseline: every bookmaker market carries a margin, and a punter betting at random pays it. With illustrative prices of $1.93 on both sides of a two-way market, the book adds up to 1 / 1.93 + 1 / 1.93 = 103.63%. A random bet then returns 1 / 1.0363 = $0.965 per dollar on average, a yield of about -3.5%. Every point above that comes from the prices taken or the selections made.
Second, luck moves a yield a long way. Take an illustrative 3% yield over 500 bets at an average price of $2.20: one standard error of luck is the square root of 1.2 / the square root of 500 = 4.9 percentage points, so two standard errors either side runs from about -7% to +13%, and longer prices widen the band. That record cannot yet tell a winning method from a losing one, and a yield is something a betting strategy measures only after enough bets, at prices you can actually get.
Risk: Betting involves risk. A yield is an average of past bets, a few hundred of them can leave it several points away from the method's real figure, and there is no guarantee of profit. See responsible gambling for limits and support.
How to read a tipster's ROI claim
A tipster's ROI is usually a yield on their own advised prices, at their own staking. Before it means anything to you, run these checks:
- Ask which denominator. Profit over stakes is a yield; profit over a bank is a bankroll figure, which can be ten or more times bigger for the same tips.
- Ask for level stakes. If tips are staked from 1 to 5 units, recompute every tip at 1 unit, the way the level-stakes example does.
- Price the tips at what you can get. With illustrative figures, say a feed shows an 8% yield at advised prices and your prices average 5% lower. Your yield is then about (1 + 0.08) x (1 - 0.05) - 1 = 2.6%, assuming the same winners. The gap is price slippage.
- Take the fee off. An illustrative $60 monthly fee on $1,500 of monthly turnover costs 60 / 1,500 = 4 percentage points of yield, which turns that 2.6% into about -1.4%.
- Count the tips and the months. A yield over 150 tips is a far weaker claim than the same yield over 3,000.
- Check the record is complete: every tip with its price, stake and the time it was posted, losers included, tracked as it happened rather than typed up afterwards. How to check a tipster record sets out what a full record shows.
Past results are no guarantee of future results, so even a record that passes all six checks describes the past, not the next hundred tips.
Mistakes that distort a yield
With illustrative figures, here is what each common mistake does:
| Mistake | What it does to the figure |
|---|---|
| Quoting return on turnover as profit | 102.5% sounds like doubling, when it is a 2.5% yield |
| Dividing by deposits instead of stakes | $500 profit on $3,000 deposited reads as 16.7%, when on $20,000 staked it is 2.5% |
| Dropping losing months or deleted bets | What is left is the survivorship bias version of the record, not the record |
| Counting open bets | An open $100 stake with nothing returned yet counts as a $100 loss until it settles |
| Comparing yields at very different prices | Over the same number of bets, 10% at $1.50 is far more reliable than 10% at $10.00 |
| Calling a hot run the method's yield | The best 100 bets of a 1,000-bet record are a selection, not a forecast |