How we grade

An S&P 500 fund is trying to track the index as cheaply as it can. A bond fund is trying to pay steady income. Neither is trying to beat the other, so it would be unfair to rank them against each other. We rank each fund against its own kind instead, and that set of funds is what we call a category.

Because every Grade is measured inside a category, an A means the same thing everywhere. An A on a semiconductor fund and an A on a Treasury fund both say the fund is near the top of its own category. They do not say which of the two funds is better. That question has no answer, because the two funds are trying to do different things.

The math is the easy part, and the rest of this page walks through it. The hard part is deciding which fund belongs in which category. We use AI for that, and the next section explains how.

The name is not enough to go on. Two funds with “emerging markets” in the name can be completely different products: one holds stocks, the other holds bonds paid in US dollars. If we sorted by name, we would end up grading a bond fund against a stock fund.

So the AI reads the paperwork instead: the prospectus, the fact sheet and the issuer’s own documents. From those it pulls out what the fund promises to do, which the industry calls its mandate. Each fact it takes carries a link back to the document it came from, so nothing in the mandate is a guess.

Buffer funds are a good example of why this matters. A fund that absorbs the first 15% of S&P 500 losses has the S&P 500 in its name. It is not a normal stock fund, though. It belongs with the other funds that offer the same index, the same protection and the same kind of time window. Put it in a broad stock category and you would be grading an options payoff as if it were a basket of stocks.

Every number later on this page assumes this step is right. The AI proposes the categories, but it does not get the last word: a person reviews every change before it goes live.

There are two levels. The category is the narrow one: the funds a Grade is measured against. Above it sits the group, which is broader: buffered funds, option income funds, Treasury and government bond funds, active equity funds, and so on. Each group holds several categories, and both levels are versioned, so every past change is on record.

How many funds does a category need? A category needs at least 5 members before it grades anything. Below that the comparison is too thin, so we grade the fund in its parent category: the next broader set of funds that make a similar promise. A copper miners category with four funds, for example, grades inside the broader global themes category. If the parent category is too thin as well, the fund gets no letter.

You will not find “bond”, “equity” or “thematic” on that list. Labels that broad are only where the sorting starts.

Two of the rows show how far apart similar-sounding funds can be. A semiconductor fund and an option-income fund on Nvidia both sound like technology funds. But one holds a whole industry and the other sells options on one company. They are graded in different categories.

Once a fund is in a category with enough members, we score it on six pillars. Each pillar score is a percentile: where the fund ranks inside its category, from 0 at the bottom to 100 at the top. We then combine the six, using the fixed weights below, into one Score from 0 to 100. The Score sets the letter.

The six pillars are Cost 30%, Mission 25%, Risk 15%, Tradability 15%, Holdings 10%, Durability 5%, and the weights sum to 100%. A Score of 70 or higher is A, 55 to under 70 is B, 40 to under 55 is C, 25 to under 40 is D, under 25 is F.

The weights never change. Every fund type and every investing goal gets the same six. Cost carries the most weight because every fund in a category charges a fee, so it is always there to compare. Mission is next. It leans hardest on the category work. The question it asks is whether the fund did the job it wrote down, not whether it had a good year.

The cutoffs are fixed. We do not grade on a curve, and we do not force a set share of funds into each letter. If every fund in a category is cheap and well run, several of them can hold an A at once.

When a pillar cannot be measured for a fund, we drop it and spread its weight across the pillars that remain. That way a fund is not punished for a metric that does not apply to it. The same goes for data we do not have yet. A gap in our data is our problem to fix, never a mark against the fund. There is a floor, though: a letter needs at least 4 of the 6 pillars, and one of them has to be Mission or Risk.

We run the numbers every day, but the letter does not move every day. A new letter has to hold for 3 consecutive published runs before it replaces the one on the page. That keeps day-to-day noise out of the Grade. A day on which we publish nothing pauses the count; it does not reset it. Three things skip that wait: a fund’s first letter, an audited reset of the methodology or the categories, and a change in whether the fund can be graded at all. The Score underneath is different. It updates every day and we do not smooth it.

Some funds also carry a Grade flag, which says what kind of product the fund is. A flag never adds or removes points. A leveraged fund is not marked down for being leveraged, and it can still earn an A inside its category. The flags below use the same names you will see next to a fund’s Grade.

Crypto derivatives

This fund gets its cryptocurrency exposure through futures contracts rather than holding coins directly. Futures-based returns can differ from the spot price of the coin.

Leveraged · daily reset

This fund multiplies its index's move each day and resets that multiple daily. Because daily returns compound, its long-run return can differ from the headline multiple.

Single-stock strategy

This fund concentrates its exposure in one company, using derivatives rather than holding the shares directly. Its results depend on that single stock.

Volatility product

This fund tracks market volatility itself rather than stocks or bonds. It is built as a short-term trading instrument, not a buy-and-hold investment.

Cost30%

One number, the fund’s net expense ratio, ranked against its category. There are no adjustments and no judgment calls.

Mission25%

Whether the fund did what it promised. The test depends on the promise. For an index or bond fund we check that the returns move with the stated mandate. If it has a duration target, we check it delivers that duration. For a buffer or defined-outcome fund we check the protection depth, the cap and the outcome period against the prospectus. For a leveraged fund we check the stated daily multiple.

In categories where every fund is cash-like, such as money-market funds, prices barely move, so the return test would only measure noise. There Mission rests on duration delivery alone.

If the evidence is missing we leave Mission unscored and spread its weight across the other pillars.

Risk15%

How the fund behaved when markets fell, over one-year and three-year windows: downside volatility, maximum drawdown, how long the drawdown lasted, the Sortino ratio and tail loss. The comparison stays inside the category, so a long-bond fund is measured against other long-bond funds rather than against Treasury bills.

Tradability15%

How easy the fund is to trade: daily trading volume, and how far the market price strays from the value of what the fund holds (its net asset value).

Holdings10%

How concentrated the fund is: the weight of its top ten positions, a standard concentration index, and how complete and up to date the disclosed holdings are.

Durability5%

Whether the fund is likely to stay open. We look at its age, its assets and the issuer’s assets. Small, young or neglected funds are the ones that tend to get closed. If we cannot observe a fund’s assets at all, we leave that part unscored. Unknown size is unknown, not small.

A new fund does not have three years of history. The metrics that need it use the one-year window instead. The Grade section on every fund page tells you which category the fund was ranked in, and how many funds were in it.

Sometimes there is no fair comparison to make. If the category is too small we try the parent category, and if that is too small as well we publish no letter. A fund without a letter always shows why, in plain words:

One ceiling can also cap a Score. It applies when something is wrong with how the product is run, and unlike a flag, a cap does remove points.

No fund pays to be graded, and no fund can pay to improve a Grade or to buy a place in a category. Bluerock sponsors this site as its sole backer.

The Grade does not predict returns. It cannot tell you what a fund will do next month or next year. It tells you where the fund stands today against the funds that make the same promise, on what we can measure.

It does not rank funds across categories either. An index fund’s A and an active bond fund’s A are both real, but they answer different questions.

And when the data cannot support a letter, we publish no letter and say why.