A basket presents itself as a way to hold a market rather than a bet. But somebody decided what goes in, how much of each, and when that changes, and those three decisions do more to determine the outcome than anything happening in the market itself.
The word index suggests measurement, and measurement suggests neutrality. But there is no neutral way to summarise a market in one number: somebody has to decide which assets qualify, how much weight each gets, how often that is revisited, and what happens when something fails. Each of those is a judgement.
The judgements are usually documented in a methodology nobody reads, and they are where the differences between two indices covering the same market come from. Two baskets of the same assets, weighted differently, produce genuinely different results, and the difference is attributable entirely to the construction.
That is not a criticism. Rules-based construction is far better than discretion, and publishing the rules is far better than not. The point is that holding a basket is holding somebody's construction decisions, and those deserve to be read once with the same attention you would give any other position.
There is no neutral way to summarise a market in one number. Somebody decided what qualifies, how much of each, and when it changes.
The most common approach weights each asset by its market value, so larger assets occupy more of the basket. Its main virtue is mechanical: as prices move, the weights adjust themselves, so the basket does not need to trade to stay correctly weighted, which keeps costs low.
Its main property is concentration, and in digital assets that property is extreme. A market where the largest asset represents a very large share of total value produces a size-weighted basket that is mostly one thing, and the diversification such a basket appears to offer is substantially smaller than the number of constituents suggests.
The second property is that it buys more of what has risen. A size-weighted basket automatically increases its allocation to assets whose price has gone up, which is a momentum exposure nobody chose deliberately. Whether that is desirable depends on the market, and the point is that it is a property rather than an absence of one.
Giving every constituent the same weight solves the concentration problem directly and creates two others. The basket now holds as much of the smallest constituent as the largest, which means its behaviour is dominated by small assets that may be far less liquid and far more volatile than the market it claims to represent.
It also requires constant trading. Because prices move at different rates, the weights drift away from equality continuously, and restoring them means selling what rose and buying what fell at every rebalancing. That produces trading costs a size-weighted basket does not have, and those costs are a permanent drag.
The implicit exposure is the mirror image of the other: equal weighting systematically sells strength and buys weakness, which is a contrarian tilt nobody selected. Neither approach is neutral, and choosing between them is choosing which non-neutral exposure you prefer rather than choosing to avoid one.
How often a basket is brought back to its target weights is a genuine trade-off with no correct answer. Frequent rebalancing keeps the basket close to its stated composition and costs more in trading. Infrequent rebalancing costs less and permits the basket to drift substantially from what it claims to be.
There is a subtler effect that matters more than the cost. Rebalancing is itself a strategy: selling what has risen and buying what has fallen, executed mechanically at fixed intervals. In markets that trend, that hurts; in markets that oscillate, it helps. The rebalancing rule is therefore a bet on market character, embedded in something described as passive.
The frequency also creates a predictable event. If a widely-held basket rebalances on a known schedule, the trades it must make are known in advance by anybody who reads the methodology, and predictable flow of any size attracts participants positioned ahead of it. That cost falls on the basket's holders.
The rules deciding what enters and leaves a basket matter more in digital assets than in conventional markets, because the turnover is higher and the exits are more abrupt. An asset that fails does not decline gracefully to a small weight; it can become untradeable quickly, and how the methodology handles that determines what holders actually experience.
The relevant questions are specific. At what point is a failing asset removed, at what price is that removal executed, and does the basket bear the loss or is the constituent quietly dropped from the record. The last one matters for anybody evaluating historical performance, because a basket whose history excludes its failures is describing a different basket.
Inclusion has its own version of the problem. Assets typically enter a basket after they have grown large enough to qualify, which means the basket buys them after their growth rather than before, and the historical record of the index includes that growth only if the index was computed retrospectively. Whether it was is worth checking.
Holding many assets instead of one reduces the risk specific to any single asset: a failure, a protocol problem, a team dispute. That is a real reduction and it is the honest case for a basket, particularly in a market where individual assets fail at a meaningful rate.
It does not reduce the risk common to all of them, and in digital assets that common component is unusually large. When the whole market moves together, a basket moves with it, and the diversification that felt reassuring during ordinary conditions provides very little during exactly the conditions people bought it for.
The honest framing is that a basket converts a specific risk into a general one. That is frequently a good trade, because the specific risks are the ones that produce total losses, but it should be understood as a change in the character of the risk rather than a reduction in its amount.
A basket converts a specific risk into a general one. Frequently a good trade, but a change in the character of the risk rather than a reduction in its amount.
The construction is one question and the implementation is another. A basket can be held by owning each constituent directly, by holding a token whose contract holds the constituents, by holding a claim on a manager who holds them, or through a derivative that references the index without holding anything.
Those are four genuinely different risk positions with the same stated exposure. Direct ownership carries the custody problem for every constituent. A contract-based basket carries the contract's risk. A managed claim carries the manager's solvency. A derivative carries the counterparty and may track the index imperfectly.
The question worth asking about any specific implementation is what you would hold if the thing furthest from the underlying assets failed. That question has a clear answer for each of the four, the answers differ substantially, and it is not the question people ask when they compare the fee.
Six things, and they are all in the document. The universe: what is eligible and what disqualifies. The weighting rule and any cap on individual weights, because a cap changes a size-weighted basket into something quite different. The rebalancing frequency and whether the dates are public.
Then: how a failing constituent is handled and at what price. Whether the historical series was computed as the index ran or reconstructed afterwards, which determines how much of the record is a real result. And what it costs, expressed as an annual percentage, against the drag of the rebalancing it performs.
Ten minutes of reading, and it explains most of the difference between two products that describe themselves identically. The methodology is generally the only document about a basket that contains information rather than description, and it is the one almost nobody opens.
Constructing a basket yourself is entirely possible and it removes the fee, which is the usual motivation. What it does not remove is the work: the rebalancing has to happen, the trading costs are yours, and every decision the methodology would have made becomes a decision you make repeatedly.
The failure mode is specific and common. A self-built basket gets constructed carefully and then rebalanced when the holder remembers, which is neither a rule nor a discretionary strategy but a drift punctuated by irregular intervention. That is usually worse than either alternative done consistently.
If you do build one, the thing that makes it work is writing the rules down before starting and following them without exception, including the parts that feel wrong at the time. The value of a rules-based basket is entirely in the rules being followed, and a rule followed when convenient is not a rule.
No. Somebody decided which assets qualify, how much weight each gets, how often that is revisited, and what happens when something fails. Those judgements are where the differences between two indices covering the same market come from, and they determine most of what the number does.
It keeps costs low because weights adjust themselves as prices move, and it produces extreme concentration in a market where the largest asset is a very large share of total value. It also automatically buys more of what has risen, which is a momentum exposure nobody chose.
It solves concentration and creates two other problems: the basket is dominated by small constituents that may be far less liquid, and it requires constant trading to maintain, which is a permanent drag. It also systematically sells strength and buys weakness, a contrarian tilt nobody selected.
Beyond cost, rebalancing is itself a strategy: selling what rose and buying what fell at fixed intervals. That hurts in trending markets and helps in oscillating ones, so the rule is a bet on market character embedded in something described as passive.
It depends on the methodology, and the questions are specific: at what point is it removed, at what price, and does the basket bear the loss or is the constituent quietly dropped from the record. The last matters because a history excluding its failures describes a different basket.
It reduces the risk specific to any single asset, which is real and is the honest case for a basket. It does not reduce the risk common to all of them, which in digital assets is unusually large. A basket converts a specific risk into a general one rather than reducing the amount. The practical consequence is that the diversification feels most reassuring during ordinary conditions and provides the least during exactly the conditions people bought it for. Knowing that in advance is what separates a basket held for the right reason from one held for a reassurance it cannot deliver.
Yes, and it is rarely asked. Owning each constituent, holding a token whose contract holds them, holding a claim on a manager, and holding a derivative are four different risk positions with the same stated exposure. Ask what you would hold if the layer furthest from the assets failed.
Possible, and it removes the fee but not the work. The common failure is a basket constructed carefully and then rebalanced whenever the holder remembers, which is neither a rule nor a strategy. Write the rules first and follow them without exception, including when it feels wrong.