Two funds of the same vintage: Mandala Fund III shows TVPI 1.9x, DPI 0.6x and a 21% IRR; Kshitij Fund II shows TVPI 1.7x, DPI 1.4x and a 17% IRR. Add a public market comparison and say which performance you trust more.

Trust Kshitij's numbers more; Mandala may be the better fund, but most of its lead is unrealised and part of its IRR is timing. Kshitij has returned 1.4x of its 1.7x, 82% in cash; Mandala has returned 0.6x of 1.9x. Mandala's 21% IRR falls to 16% if its calls had not been delayed a year. Both beat a 12% index on a PME basis, Mandala more so, but only if its marks hold: a 15% markdown erases its TVPI lead.

A friend who says his flat has doubled in value and one who has sold his flat for double are telling you different things. DPIDistributions to paid-in capital: cash returned to investors divided by cash they have put in. It cannot be revised. is cash already returned and cannot change; RVPIResidual value to paid-in capital: the manager current valuation of investments still held, divided by capital paid in. is the manager's valuation of what is still held; TVPI is the two added together. Mandala's 1.9x is 0.6x of cash and 1.3x of marks, 32% realised. Kshitij's 1.7x is 1.4x of cash and 0.3x of marks, 82% realised. A five year old fund with two thirds of its value still in marks is asking you to trust its valuations.

Rebuild the IRRs from the dated flows, with the reported value treated as if received at the end of year 5. Mandala's IRR comes out at 21.1% and Kshitij's at 17.0%, but if Mandala had called its capital a year earlier, as Kshitij did, its IRR would be 15.9%: about 5 points of its IRR come from the subscription line, not from the deals. The money multiple is unchanged by the timing; only the IRR moves. That is why investors ask managers to report IRR with and without the line, and why you should compare the two funds on both.

A PMEPublic market equivalent: compares a fund with what the same cash flows would have earned in a public index. Above 1.0 the fund beat the index. asks whether the same rupees, put into an index on the same dates, would have done better. Grow every call and every distribution to year 5 at the index return, add the reported value, and divide. At an assumed 12% a year, Mandala's PME is 1.30 and Kshitij's is 1.16: both beat the index, Mandala by more. The PME still uses Mandala's marks, though. Cut them 20% and Mandala's PME falls to 1.13, below Kshitij, its TVPI to 1.64x and its IRR to 16.0%.

Trust Kshitij's record more, because 82% of it is cash in investors' hands; treat Mandala's as promising but unproven until the marks turn into exits. Then ask Mandala four things: how each holding is valued and against which listed peers, how the marks compare with the last round or offer for each company, which exits are in progress and at what prices, and its IRR without the subscription line. If the top three holdings make up most of the Rs 130 crore, diligence those three companies directly. The limit of the comparison is time: five years in, Mandala may simply be earlier in its harvest, and in two years its DPI may settle the question in its favour.