Dollar-Cost Averaging Strategy Performance Analysis
정기 투자 (적립식 투자) 전략 백테스팅 시뮬레이터
Dollar-cost averaging means putting the same amount of money into the same asset on a fixed schedule, regardless of what the price is doing on the day. It buys more shares when the price is low and fewer when it is high, so the question it raises is not whether the asset went up — it is whether spreading the purchases out did better or worse than committing the money at the start. This page answers that question for one ticker over one historical window, using daily closing prices.
Three schedules are available. A monthly plan buys on the first trading day of each calendar month. A daily plan buys on every trading day in the range. A lump sum buys once, on the first trading day. In all three the full contribution is invested each time: the simulator holds fractional shares rather than rounding down to a whole one, so no part of a contribution is left uninvested.
Total invested capital is the sum of the contributions actually made, and final portfolio value is the shares accumulated valued at the last closing price in the range. The difference between them is the absolute profit, and the total return is that difference as a percentage of what was paid in.
The annualized return is a money-weighted internal rate of return. A contribution plan puts capital in gradually, so a time-weighted CAGR is not defined for it: money paid in last month was never exposed for a year. The IRR is the single yearly rate that, applied to each contribution over the days it was actually invested, reproduces the final portfolio value. It is solved numerically, and when no rate satisfies the cash flows — which happens for some short or fully loss-making ranges — the page shows 0.00%. A displayed 0.00% therefore means either “no gain” or “not solvable”, and the absolute profit above it tells you which.
MDD, the maximum drawdown, is the largest percentage fall from a running peak, measured on daily closes over the period. It is computed on the unit value of the plan — the value of one unit of the portfolio, which contributions do not move — rather than on the portfolio balance. That is the right basis: measured on the balance, every incoming contribution lifts the running peak and erases part of a falling market, so a plan that bought weekly through a crash would report a smaller drawdown than the crash itself. On unit value the figure answers “how far did the asset fall from its high while I held it?” and does not change if you alter how much or how often you pay in.
Buy & Hold is the control. It takes the total capital the schedule would contribute over the whole period and buys it all at the first closing price in the range, then values it at the last. Because both sides spend the same amount of money and both hold fractional shares, the comparison is like for like — the only difference is the timing of the purchases, which is exactly the thing being tested.
Sharpe ratio is a rough volatility-adjusted figure: day-over-day changes in unit value, divided by their standard deviation, annualized by the square root of the number of trading days in a year, with the risk-free rate treated as zero. Because a contribution buys more units rather than raising the unit value, the ratio measures the volatility of the asset over the period held, not the rhythm of the payments. Treat it as indicative only: the zero risk-free rate and the square-root-of-252 scaling are both approximations, and a short range gives too few observations for the standard deviation to mean much.
The two risk figures above — the maximum drawdown and the Sharpe ratio — are built on the unit value series rather than on the raw account balance. Contributions increase the number of units held and never the price of a unit, so neither of those two numbers can be flattered by paying in more often; what they report is the behaviour of the asset over the window, which is the thing a schedule cannot control. The annualized return is deliberately the opposite. It is money-weighted and is solved from the contribution cash flows and the final portfolio value alone, so it does depend on when you paid in — which is exactly what makes it the right yearly figure for a contribution plan, and why it is not comparable with the two risk figures beside it.
The simulation runs on unadjusted daily closing prices from Yahoo Finance. It does not reinvest dividends, and it does not include brokerage commissions, spreads, account fees, taxes on gains or distributions, or currency conversion. A real plan pays those, so a real result will differ from this one — for a dividend-paying asset the omitted distributions push in your favour, while costs and taxes push the other way. If the price feed cannot be reached for the ticker and range you entered, the page falls back to generated sample data so the interface still responds; a run whose numbers look nothing like the asset you asked for is the sign of that, and it is worth retrying.
Above all, this is a backtest. It reports what one schedule would have produced over one window that has already happened, chosen by you after the fact. Past performance is no guarantee of future results, a different start date can reverse the conclusion entirely, and nothing on this page is investment advice or a recommendation to buy or sell any security. Do your own research, and speak to a licensed adviser before committing money.