Kimchi Premium Mean Reversion
Sell the KRW premium when it stretches past 3%, buy it back as it reverts toward zero.
- SignalPremium ≥ 3%
- EntrySell Upbit · buy Binance
- ExitPremium ≤ 0.5%
How it works
Two legs, one spread · holds hours – days- premium60/hKRW premium for every coin, ranked
- forex60/hUSD/KRW conversion rate
- wallet_status12/hGate check before any entry
- SignalPremium ≥ 3%Upbit KRW vs Binance USDT at the live USD/KRW rate, every minute
- EntrySell Upbit · buy BinanceHeld 5 min, all gates open, ≥ $500K volume each side
- ExitPremium ≤ 0.5%Or 72 hours in the trade
- RiskStop past 6% or gate shutMax 10% of book per coin; KRW and USDT pre-positioned
The signal
XRP · Upbit vs Binance · premium % · sample| Hour (KST) | Premium |
|---|---|
| Nov 15, 00:00 | 1.45% |
| Nov 15, 01:00 | 1.55% |
| Nov 15, 02:00 | 1.43% |
| Nov 15, 03:00 | 1.49% |
| Nov 15, 04:00 | 1.59% |
| Nov 15, 05:00 | 1.36% |
| Nov 15, 06:00 | 1.08% |
| Nov 15, 07:00 | 1.09% |
| Nov 15, 08:00 | 1.1% |
| Nov 15, 09:00 | 0.89% |
| Nov 15, 10:00 | 0.84% |
| Nov 15, 11:00 | 1.07% |
| Nov 15, 12:00 | 1.16% |
| Nov 15, 13:00 | 1.03% |
| Nov 15, 14:00 | 4.09% |
| Nov 15, 15:00 | 3.88% |
| Nov 15, 16:00 | 3.38% |
| Nov 15, 17:00 | 2.8% |
| Nov 15, 18:00 | 2.49% |
| Nov 15, 19:00 | 2.23% |
| Nov 15, 20:00 | 1.74% |
| Nov 15, 21:00 | 1.39% |
| Nov 15, 22:00 | 1.39% |
| Nov 15, 23:00 | 1.33% |
| Nov 16, 00:00 | 1.1% |
| Nov 16, 01:00 | 1.1% |
| Nov 16, 02:00 | 1.28% |
| Nov 16, 03:00 | 1.18% |
| Nov 16, 04:00 | 0.93% |
| Nov 16, 05:00 | 0.9% |
| Nov 16, 06:00 | 0.89% |
| Nov 16, 07:00 | 0.59% |
| Nov 16, 08:00 | 0.35% |
| Nov 16, 09:00 | 0.42% |
| Nov 16, 10:00 | 0.47% |
| Nov 16, 11:00 | 0.34% |
| Nov 16, 12:00 | 0.43% |
| Nov 16, 13:00 | 0.77% |
| Nov 16, 14:00 | 0.89% |
| Nov 16, 15:00 | 0.83% |
| Nov 16, 16:00 | 0.98% |
| Nov 16, 17:00 | 1.16% |
| Nov 16, 18:00 | 1.01% |
| Nov 16, 19:00 | 0.81% |
| Nov 16, 20:00 | 0.89% |
| Nov 16, 21:00 | 0.93% |
| Nov 16, 22:00 | 3.56% |
| Nov 16, 23:00 | 3.18% |
| Nov 17, 00:00 | 3.14% |
| Nov 17, 01:00 | 3% |
| Nov 17, 02:00 | 2.72% |
| Nov 17, 03:00 | 2.69% |
| Nov 17, 04:00 | 2.77% |
| Nov 17, 05:00 | 2.54% |
| Nov 17, 06:00 | 2.22% |
| Nov 17, 07:00 | 2.13% |
| Nov 17, 08:00 | 2.02% |
| Nov 17, 09:00 | 1.65% |
| Nov 17, 10:00 | 1.41% |
| Nov 17, 11:00 | 1.48% |
| Nov 17, 12:00 | 1.46% |
| Nov 17, 13:00 | 1.28% |
| Nov 17, 14:00 | 1.33% |
| Nov 17, 15:00 | 1.56% |
| Nov 17, 16:00 | 1.5% |
| Nov 17, 17:00 | 1.3% |
| Nov 17, 18:00 | 1.31% |
| Nov 17, 19:00 | 1.29% |
| Nov 17, 20:00 | 0.96% |
| Nov 17, 21:00 | 0.66% |
| Nov 17, 22:00 | 0.65% |
| Nov 17, 23:00 | 0.58% |
Sample backtest
Sample · illustrative 90D · not real performance| Date (UTC) | Equity |
|---|---|
| Jul 21 | 0.32% |
| Jul 22 | 0.66% |
| Jul 23 | 0.91% |
| Jul 24 | 1.21% |
| Jul 25 | 1.27% |
| Jul 26 | 1.32% |
| Jul 27 | 1.51% |
| Jul 28 | 1.64% |
| Jul 29 | 1.93% |
| Jul 30 | 2.18% |
| Jul 31 | 2.2% |
| Aug 1 | 2.24% |
| Aug 2 | 2.16% |
| Aug 3 | 2.08% |
| Aug 4 | 2.24% |
| Aug 5 | 2.32% |
| Aug 6 | 2.42% |
| Aug 7 | 2.49% |
| Aug 8 | 2.32% |
| Aug 9 | 2.26% |
| Aug 10 | 2.26% |
| Aug 11 | 2.28% |
| Aug 12 | 2.54% |
| Aug 13 | 2.7% |
| Aug 14 | 2.76% |
| Aug 15 | 2.87% |
| Aug 16 | 2.83% |
| Aug 17 | 2.95% |
| Aug 18 | 3.25% |
| Aug 19 | 3.49% |
| Aug 20 | 3.85% |
| Aug 21 | 4.07% |
| Aug 22 | 4.12% |
| Aug 23 | 4.29% |
| Aug 24 | 4.4% |
| Aug 25 | 4.62% |
| Aug 26 | 5% |
| Aug 27 | 5.17% |
| Aug 28 | 5.32% |
| Aug 29 | 5.38% |
| Aug 30 | 5.27% |
| Aug 31 | 5.35% |
| Sep 1 | 5.46% |
| Sep 2 | 5.56% |
| Sep 3 | 5.76% |
| Sep 4 | 5.71% |
| Sep 5 | 5.6% |
| Sep 6 | 5.56% |
| Sep 7 | 5.44% |
| Sep 8 | 5.55% |
| Sep 9 | 5.74% |
| Sep 10 | 5.8% |
| Sep 11 | 5.93% |
| Sep 12 | 5.89% |
| Sep 13 | 5.83% |
| Sep 14 | 6% |
| Sep 15 | 6.15% |
| Sep 16 | 6.44% |
| Sep 17 | 6.78% |
| Sep 18 | 6.89% |
| Sep 19 | 7.04% |
| Sep 20 | 7.15% |
| Sep 21 | 7.25% |
| Sep 22 | 7.61% |
| Sep 23 | 7.9% |
| Sep 24 | 8.14% |
| Sep 25 | 8.38% |
| Sep 26 | 8.36% |
| Sep 27 | 8.39% |
| Sep 28 | 8.51% |
| Sep 29 | 8.59% |
| Sep 30 | 8.84% |
| Oct 1 | 8.99% |
| Oct 2 | 8.93% |
| Oct 3 | 8.93% |
| Oct 4 | 8.78% |
| Oct 5 | 8.72% |
| Oct 6 | 8.87% |
| Oct 7 | 8.93% |
| Oct 8 | 9.06% |
| Oct 9 | 9.12% |
| Oct 10 | 8.97% |
| Oct 11 | 8.99% |
| Oct 12 | 9.03% |
| Oct 13 | 9.15% |
| Oct 14 | 9.49% |
| Oct 15 | 9.68% |
| Oct 16 | 9.82% |
| Oct 17 | 9.96% |
| Oct 18 | 9.97% |
| Oct 19 | 10.18% |
| Oct 20 | 10.5% |
| Oct 21 | 10.76% |
| Oct 22 | 11.13% |
| Oct 23 | 11.29% |
| Oct 24 | 11.32% |
| Oct 25 | 11.46% |
| Oct 26 | 11.51% |
| Oct 27 | 11.72% |
| Oct 28 | 12.02% |
| Oct 29 | 12.1% |
| Oct 30 | 12.2% |
| Oct 31 | 12.15% |
| Nov 1 | 11.99% |
| Nov 2 | 12.06% |
| Nov 3 | 12.12% |
| Nov 4 | 12.21% |
| Nov 5 | 12.38% |
| Nov 6 | 12.29% |
| Nov 7 | 12.21% |
| Nov 8 | 12.18% |
| Nov 9 | 12.12% |
| Nov 10 | 12.32% |
| Nov 11 | 12.54% |
| Nov 12 | 12.67% |
| Nov 13 | 12.85% |
| Nov 14 | 12.85% |
| Nov 15 | 12.88% |
| Nov 16 | 13.12% |
| Nov 17 | 13.33% |
Generated to show the shape of the strategy, not measured on real fills. Ask your agent to backtest on DataMaxi+ history before trading.
Backtest this strategy
One prompt: your agent pulls the history below, tests these rules with fees and slippage, and charts the equity curve and drawdown. Run it before you trust the sample above.
- Period
- The last 180 days, hourly
- History calls
- 50
- Credits
- 250
- Cost per run
- ≈ $0.025
Backtest the "Kimchi Premium Mean Reversion" strategy on DataMaxi+ history. Thesis: Sell the KRW premium when it stretches past 3%, buy it back as it reverts toward zero. Use the DataMaxi+ MCP server at https://mcp.datamaxiplus.com/mcp. Pull the history once, cache it, and run the backtest in your own code sandbox (Python is fine). ## History to pull (split the period with from/to; never refetch a range you have) 1. cex_candle(exchange="upbit", market="spot", symbol="XRP-KRW", interval="1h") — OHLCV, up to 1,000 candles per call, history from Jan 2024; about 25 calls 2. cex_candle(exchange="binance", market="spot", symbol="XRP-USDT", currency="KRW", interval="1h") — OHLCV, up to 1,000 candles per call, history from Jan 2024; about 25 calls Universe: BTC, ETH, XRP, SOL and DOGE; Upbit KRW spot against Binance USDT spot Period: the last 180 days, hourly ## Rules to test Signal: Upbit KRW price vs Binance USDT price at the live USD/KRW rate (premium %), per coin, every minute. Entry: Premium ≥ 3.0% for 5 minutes, both venues' deposit and withdrawal gates open, 24h volume ≥ $500K on each side. Exit: Premium ≤ 0.5%, or 72 hours in the trade. Risk: Stop if premium widens past 6% or any gate closes. Max 10% of book per coin; pre-position KRW and USDT so neither leg waits on a transfer. ## Assumptions Costs: 0.05% Upbit and 0.1% Binance taker per leg, plus 0.1% slippage per leg Data: There is no premium history endpoint: rebuild the premium as Upbit KRW close ÷ Binance close in KRW − 1 (cex_candle with currency=KRW). Wallet gates keep no history, so assume they were open and say so. Use only data stamped before each decision (no look-ahead). Say plainly where the data cannot test a rule. ## Report Net return, annualised return, Sharpe, max drawdown, number of trades, win rate and average hold, next to buy-and-hold over the same period. Chart the equity curve and the drawdown, and list the 5 worst trades. ## Budget About 50 history calls ≈ 250 credits ≈ $0.025 from my prepaid DataMaxi+ balance. Tell me how many calls you made. If DataMaxi+ is not connected, ask me to [connect DataMaxi+](https://claude.ai/customize/connectors?modal=add-custom-connector) using https://mcp.datamaxiplus.com/mcp before continuing.
Estimate from the call counts in the prompt. Your agent may pull less or more; it reports the calls it made.
Rules
| Signal | Upbit KRW price vs Binance USDT price at the live USD/KRW rate (premium %), per coin, every minute. |
|---|---|
| Entry | Premium ≥ 3.0% for 5 minutes, both venues' deposit and withdrawal gates open, 24h volume ≥ $500K on each side. |
| Exit | Premium ≤ 0.5%, or 72 hours in the trade. |
| Risk | Stop if premium widens past 6% or any gate closes. Max 10% of book per coin; pre-position KRW and USDT so neither leg waits on a transfer. |
Data used
≈ $0.3168 / day| MCP tool · REST route | Dataset | Tier | Calls / h | Credits / day | USD / day |
|---|---|---|---|---|---|
premium GET /premium | Premium | Latest · 1 cr | 60 | 1,440 | $0.144 |
forex GET /forex | Index price · forex | Latest · 1 cr | 60 | 1,440 | $0.144 |
wallet_status GET /wallet-status | Wallet status | Latest · 1 cr | 12 | 288 | $0.0288 |
| Total | 132 | 3,168 | $0.3168 |
1 credit = $0.0001. Only successful calls are billed, from one balance shared by the API and MCP. Calls per hour are the suggested cadence; scale them to the symbols you watch.
Build it with your agent
Connect the MCP server →Open it in Claude with one click (it asks you to add the DataMaxi+ connector first if you haven't), or paste it into Cursor or any MCP client. It backtests first, then runs in paper mode and never places orders on its own.
Build the "Kimchi Premium Mean Reversion" strategy for me with DataMaxi+ data. Thesis: Sell the KRW premium when it stretches past 3%, buy it back as it reverts toward zero. Connect to the DataMaxi+ MCP server at https://mcp.datamaxiplus.com/mcp (header X-DTMX-APIKEY, read from env DATAMAXI_API_KEY). ## Data (call these tools; arguments are examples — loop over the symbols and venues you need) 1. premium(source_exchange="binance", target_exchange="upbit", target_quote="KRW", sort="desc", key="pdp", limit="50") — KRW premium for every coin, ranked; about 60×/hour 2. forex(symbol="USD-KRW") — USD/KRW conversion rate; about 60×/hour 3. wallet_status(exchange="upbit", asset="XRP") — Gate check before any entry; about 12×/hour ## Rules Signal: Upbit KRW price vs Binance USDT price at the live USD/KRW rate (premium %), per coin, every minute. Entry: Premium ≥ 3.0% for 5 minutes, both venues' deposit and withdrawal gates open, 24h volume ≥ $500K on each side. Exit: Premium ≤ 0.5%, or 72 hours in the trade. Risk: Stop if premium widens past 6% or any gate closes. Max 10% of book per coin; pre-position KRW and USDT so neither leg waits on a transfer. ## What to do 1. Build the signal from the tools above and backtest the rules on the history you can pull; report trades, win rate, average trade and max drawdown. 2. Then run it live in paper mode and show me a table of coins with premium ≥ 3% and open gates, with suggested size, refreshed on the cadence above. 3. Never place real orders. Propose trades and wait for my confirmation. ## Budget Every call is billed from my prepaid DataMaxi+ balance. Expected load: ~132 calls/hour ≈ 3,168 credits/day (≈ $0.32/day). Cache results, never repeat a call inside its refresh window, and print the number of calls each refresh made. If DataMaxi+ is not connected, ask me to [connect DataMaxi+](https://claude.ai/customize/connectors?modal=add-custom-connector) using https://mcp.datamaxiplus.com/mcp before continuing.
Risk notes
- A closed deposit gate pins the premium — never enter without fresh wallet_status.
- KRW off-ramp and FX moves sit on top of the spread; the USD/KRW rate must be refreshed with the price.
- Regulatory events in Korea can reprice the whole market at once.
Educational example, not investment advice. Sample charts and figures are illustrative. See the live screen it builds on: Kimchi Premium Tracker.
Keep exploring
Live screens, dashboards and the data behind this strategy.