TWR vs IRR vs your broker's number: why they all disagree

Last reviewed 11 March 2026.

You linked two brokers, opened your first consolidated view, and one of them says you're up 17%, the other says 11%, and the spreadsheet you keep says 9.7%. Nothing was mistyped. You didn't miss a transaction. The holdings are identical. This is the moment nearly every multi-broker investor hits in their first month, and it feels like a bug in someone's software.

It isn't a bug. It's three different methods measuring three different things, and once you see the formulas side by side the confusion evaporates for good. The gap between them is not noise — it's information about how much your timing of deposits helped or hurt you. Below we'll build both formulas by hand, run one portfolio through all three, and give you a one-line rule for picking the right number every time.

The two formulas, in plain arithmetic

Start with time-weighted return (TWR). The idea is to answer “how good were my investment choices?” without letting the size or timing of my deposits flatter or distort the answer. To do that, you chop the measurement period into sub-periods — a new one begins every time cash moves in or out — calculate the plain percentage return of each sub-period, then chain them together by multiplication:

TWR = (1 + r₁) × (1 + r₂) × … × (1 + rₙ) − 1

Each r is the return of one sub-period, calculated on the balance you actually held during it. Because every deposit starts a fresh sub-period at its own price, the amount you put in never touches the percentages you multiply together. That is exactly why fund managers and indices are quoted in TWR: it isolates skill (or luck) in the holdings from the client's cash-flow behaviour. It is the only fair way to say “did I beat the S&P?”

Now internal rate of return (IRR), also called the money-weighted return. This one asks the opposite question: “across all the pounds I actually committed, and for the length of time each pound was working, what single annual rate did I earn?” IRR is the discount rate that makes the present value of everything you paid in equal to the value of everything you have (plus anything you took out). You list every cash flow with its date and sign — deposits negative, your ending balance positive — and solve for the rate r that satisfies:

∑ CFₜ ÷ (1 + r)^t = 0

There's no clean algebraic rearrangement; you solve it by trial and error (or let a spreadsheet's XIRR do it). The key property is that a large deposit made right before a bad stretch drags the whole result down, because those pounds were exposed to the loss — whereas TWR would shrug that off. IRR is what tells you how your money did, not how the average pound in the strategy did.

One portfolio, three numbers: the worked example

Here is a deliberately simple year, chosen so the two methods pull apart clearly. You will see how the same set of facts produces 17.50%, 11.66% and 9.67%.

You begin the year with £10,000 invested. Over the first six months your holdings rise 25%, so the balance reaches £12,500. At the midpoint you get a bonus and deposit a further £5,000, taking the balance to £17,500. In the second half the market softens and your holdings fall 6%, leaving you with £16,450 at year end. Across the year you deposited £15,000 in total and finished with £16,450 — a gain of £1,450.

Time-weighted return. Two sub-periods, split by the deposit: the first returned +25% (£10,000 → £12,500), the second returned −6% (on the £17,500 you then held). Chain them: (1.25 × 0.94) − 1 = 1.175 − 1 = 17.50%. Notice the £5,000 deposit never appears in the percentages — TWR only sees a 25% period followed by a −6% period.

Internal rate of return. Now the cash flows and their dates matter: −£10,000 at the start, −£5,000 at six months, and +£16,450 at year end. Solving for the annual rate that balances those flows gives 11.66%. It is lower than TWR because the extra £5,000 — a big slug of money — was deposited just before the 6% fall, so a large share of your capital only ever experienced the losing half of the year. That's a real, felt outcome, and IRR is the number that captures it. The gap of roughly six percentage points (17.50% versus 11.66%) is entirely the cost of that timing.

Your broker's number. Most retail platforms don't show either. They show a simple return: total gain divided by total deposited, £1,450 ÷ £15,000 = 9.67%. It's the easiest figure to compute and the easiest to read, but it ignores how long each pound was invested and ignores when the market moved, so it lands in a third place again. Three numbers, one portfolio, none of them wrong.

Which number is the right question?

The rule is short: use TWR to grade the investments, use IRR to grade your money.

If the question is “did my strategy beat a simple index fund?” you want TWR, because indices and funds are quoted time-weighted and comparing IRR to a TWR benchmark is comparing two different measurements. TWR answers “were my picks any good?” cleanly, whatever your saving pattern looked like.

If the question is “did I, personally, actually make money at a decent rate given how and when I invested?” you want IRR. It's the honest scorecard for a real human who dollar-cost-averages, lump-sums a bonus, or sells to fund a house deposit. When IRR sits well below TWR, that gap is a behavioural signal — it means your timing of deposits worked against you, and it is worth understanding rather than hiding.

And your broker's simple return? It's fine for a rough “am I up or down?” glance, but don't benchmark with it and don't read too much into small differences between platforms — they're often just methodology, not performance. The healthiest habit is to know which of the three you're looking at before you draw any conclusion, especially when you're consolidating two, three or four accounts that each compute returns their own way.

What this doesn't tell you

A few honest caveats. First, StackViewer is built for investors thinking in months and years, not for day traders — our figures are not real-time. We sync connected brokers three times a day and refresh values roughly every 30 minutes, so an intraday TWR or IRR reading will lag the live market. Second, IRR can misbehave mathematically: a portfolio with an unusual pattern of large deposits and withdrawals can technically produce more than one valid rate, or none, and any tool (ours included) has to pick a sensible one. Third, dividends, fees, foreign-exchange conversions and taxes all shift these numbers, and different brokers handle them differently — part of why platform returns diverge in the first place. Finally, assets we can't see through a broker connection (physical property, private holdings, some custom or unlisted instruments) sit outside the calculation entirely, and a handful of EU broker connections expose less transaction history than US ones, which can blunt IRR precision. The maths is only ever as complete as the cash-flow data feeding it.

Frequently asked

Is TWR or IRR the better measure of my returns?

Neither is universally better; they answer different questions. Use TWR to judge how well your investments performed, independent of when you added money — it is the fair way to compare yourself to an index or a fund manager. Use IRR (money-weighted return) to judge how well your actual pounds performed, because it accounts for the size and timing of every deposit and withdrawal. If you rarely add cash, the two numbers stay close; if you make large, badly-timed deposits, IRR can fall several points below TWR.

Why is my broker's return different from both TWR and IRR?

Most retail brokers show a simple return: the gain in pounds divided by the total amount you deposited, with no adjustment for how long each pound was invested or when the market moved. It is easy to compute and easy to read, but it is neither time-weighted nor money-weighted, so it will usually sit somewhere between TWR and IRR — or outside both. Brokers also differ on whether they annualise, whether they include dividends, and how they treat pending transfers, which is why two brokers can report two different “returns” for the same holdings.

This article is for information only. It is not financial, investment or tax advice, and nothing here is a recommendation to buy or sell any security. StackViewer analyses decisions you have already made; it does not predict returns. Past performance does not indicate future results.