Every portfolio eventually meets an asset that is not a stock, a bond, or real estate: cryptocurrency, gold, a friend's startup, collectibles, a private fund with a glossy deck. The question is never really whether the asset is good or bad — reasonable people disagree, and the future is unwritten. The question that decides whether the experiment helps or hurts is how much. This page is a sizing framework: what job the asset would have to do, what its loss would cost the plan, and the ceiling that keeps a speculation a speculation.

Start with the job description

An asset earns a portfolio slot by doing a job the core cannot: diversifying a specific risk, hedging a specific fear, or satisfying a genuine interest at a known price. "It might go up a lot" is not a job — it is a hope every asset shares. Gold's traditional job is crisis ballast; it pays nothing, produces nothing, and its return is entirely what the next buyer will pay — which is also, word for word, the description of most collectibles, whose long-term gains the tax code treats specially (collectibles gains face a maximum federal rate of 28%, higher than the 20% top rate for securities). Cryptocurrency's job, for most holders, is a high-volatility growth bet with no cash flows to value — which means its size cannot be derived from fundamentals, only from the loss you can shrug off. Private investments add a second cost to size: illiquidity. Money locked for seven to ten years cannot rebalance, cannot bridge a job loss, and cannot be harvested in a crisis.

The arithmetic of the slice going to zero

Size the position by its worst case, because for speculative assets the worst case is a real case. On a $500,000 portfolio:

Speculative sliceDollars at riskGain the remaining portfolio needs just to get back to $500,000
5%$25,000+5.3%
10%$50,000+11.1%
20%$100,000+25.0%

Read the last column as the plan's recovery burden. A 5% slice lost entirely is a bad year the core portfolio routinely delivers on its own; the plan barely notices. A 20% slice lost is a setback measured in years of savings. This is why practitioners who permit speculative positions at all tend to fence them near 5% of the portfolio — a size at which being wrong is tuition, not damage — and why the slice is measured against the whole portfolio, not against the "play money" account it sits in. The fence is also a rebalancing rule: a winner that grows from 5% to 15% of the portfolio has quietly rewritten your allocation, and trimming it back is Rebalancing doing its least comfortable job.

The rules that keep the slice a slice

Fund it from the portfolio, never from the plan. The emergency fund, the bridge money, and the retirement core are not venture capital; the priority order in Where Your Next Dollar Goes puts speculation last for a reason. Write the exit before the entry. A price, a date, or a portfolio percentage at which you trim or sell — written down while you are calm. Count it in the allocation. A crypto position is not "outside" the portfolio when the statement arrives; it belongs in the risk picture drawn by Asset Allocation by Age and Risk, on the aggressive end of it. Refuse leverage and lockups you cannot afford. Margin turns a slice that can go to zero into a liability that can exceed it; long lockups convert a sizing decision into a decade-long commitment. The core portfolio — Index Funds and ETFs, plus real estate exposure if you want it via Rentals vs. REITs — is where retirements are actually built. Alternatives are the spice rack: real flavor, tiny quantities, and never mistaken for the meal.

Official documentation and references

Keep going

Educational disclaimer: This article is general education, not personalized financial, tax, legal, or benefits advice. Primary sources, plan documents, and — for insurance and estate topics — your state's law and licensed professionals govern your situation. All numbers in examples are illustrative.