Dollar-Cost Averaging
DCA · Split buying · Scaling in · Recurring buy · Regular investing · Cost averaging
Spreading a purchase over several buys instead of one; the classic form, DCA, invests the same amount at fixed intervals.
Simple explanation
Background: Koreans call splitting a purchase 분할 매수 ("split buying"), a term that covers both buying at preset price levels and buying the same amount on a schedule, the "savings-plan" style (적립식) that English calls dollar-cost averaging. Several Korean exchanges now automate it.
Split buying means fixing the total you will spend and putting it in over several purchases, the opposite of a lump-sum buy. There are two main ways to split. By price: buy a slice at ₩1,000, another at ₩900 and another at ₩800, or leave limit orders at several prices. By time: buy the same amount at regular intervals, such as ₩100,000 (about $71) every Monday. The second is dollar-cost averaging (DCA). Selling in slices the same way is called scaling out.
Because DCA spends the same amount every time, it buys more coins when the price is low and fewer when it is high. That makes your average purchase price lower than the plain average of the prices over the period, or equal to it only if the price never changed. That is the "averaging" in the name. But it only spreads the luck of timing across many dates; it does not make the price go up. If the price falls throughout, DCA still loses money, just less than a lump sum bought at the start.
Compared with a lump sum, buying everything at once tends to do better when prices rise steadily, because the money you hold back keeps going in at higher prices. When prices drop right after you start, splitting loses less. You only know which was better afterwards, so split buying gives up the chance of buying everything at the bottom in exchange for avoiding buying everything at the top. Which approach to take is each person's call, and this post does not recommend either.
Korean exchanges charge trading fees as a percentage of the amount filled, so splitting the same total into several buys leaves total fees almost unchanged. Each slice must still clear the minimum order (₩5,000, about $3.60, on Upbit's KRW market). A large market order runs through the order book and suffers slippage, so spreading a big order over time is also a way to reduce slippage. Meanwhile the won not yet invested sits outside the market, so it misses part of any rise and part of any fall.
As of early November 2026 several Korean exchanges offer automatic split buying. Upbit launched 코인 모으기 ("coin collecting") in August 2024, buying a set amount automatically every day, week or month at a time you choose, and Coinone (since December 2025) and Digital X (formerly Korbit) run similar services. According to a report on data the exchanges submitted to the National Assembly, cumulative investment through the three services reached about ₩972.9 billion (about $695 million) by the end of August 2026, around 92% of it on Upbit. Bithumb does not run such a service; it offers an automatic split order (July 2025) that places up to 10 orders at different prices for one coin at once, and TWAP, a time-split order that divides the amount into equal slices and submits them as market orders over the period and interval you set. Bithumb widened TWAP to its app, PC site and mobile web in March 2026; reportedly one TWAP order can be up to ₩1 billion (about $714,000) with at least ₩5,000 per slice, and once running it can only be cancelled as a whole, not changed or partly cancelled. Supported coins, limits and fees differ by exchange and change, so check each exchange's own guide.
An automatic buy ignores price and premium: it buys at the market price at the set time, whether that is just after a spike or right after a listing when the kimchi premium is huge, and if it buys a thinly traded coin at market it takes the slippage too. If your won balance runs short, the remaining orders can stop or end. And the point of split buying is deciding in advance how much, in how many slices and how often; buying more every time the price falls, beyond the amount you set, is just averaging down by another name. Splitting does not reduce the risk of the coin itself either: if it is put under an investment warning or delisted, everything bought so far is affected.
Each slice bought in Korea also pays the kimchi premium of that moment. Keep buying while the premium is high and that premium builds up in your average; when it later shrinks you show a paper loss even if the overseas price has not moved. Slices bought at a reverse premium, on the other hand, were cheaper than abroad. Buying over time averages the premium across many dates as well, but when the premium stays high for months, that helps little.
On CoinGaja, the Kimchi premium by coin table shows the current prices and premiums on the four Korean exchanges, and the CoinGaja Kimchi Premium Index chart (24 hours, 7 days, 30 days, 1 year, all time) shows whether the market-wide premium is running high or low. The Price and premium heatmap shows each coin's 24-hour price change as coloured tiles.
Example
₩100,000 (about $71) a month for 4 months, at prices of ₩1,000 → ₩800 → ₩500 → ₩1,000
Coins bought = 100 + 125 + 200 + 100 = 525 → average = 400,000 ÷ 525 = about ₩761.90 (below the plain average of the prices, ₩825)
At the final ₩1,000: split buying 525 × 1,000 = ₩525,000 (about $375, +31.25%) / ₩400,000 all at once at the start: 400 × 1,000 = ₩400,000 (0.00%)
If prices rise steadily ₩1,000 → ₩1,200 → ₩1,500 → ₩2,000: 100 + 83.33 + 66.67 + 50 = 300 coins → 300 × 2,000 = ₩600,000 (about $429, +50.00%) / all at once: 400 × 2,000 = ₩800,000 (about $571, +100.00%)
If prices fall steadily ₩1,000 → ₩800 → ₩600 → ₩400: 100 + 125 + 166.67 + 250 = 641.67 coins → × 400 = about ₩256,667 (about $183, -35.83%) / all at once: 400 × 400 = ₩160,000 (about $114, -60.00%)
Assume a 0.05% buying fee: four buys of ₩100,000 or one of ₩400,000, the fee is 400,000 × 0.0005 = ₩200 either way
※ Example numbers, for illustration only
See it on CoinGaja
Related terms
Comments · ask anything about this term
Comments
No comments yet.