Confluence Zone does not answer "Should I buy this stock?" Instead it scores the same stock with investing strategies that think differently, and shows where their judgments overlap and where they split. Overlap (confluence) is rare, so it is worth noticing in itself, and a split is information too.
| Service | Confluence Zone |
|---|---|
| Representative | 이승도 |
| Business registration no. | 886-22-02291 |
| Contact | support@conf-zone.com |
| Address | https://conf-zone.com |
Seungdo Lee, software developer and individual investor
Designs, builds and runs all of Confluence Zone himself, including data processing, indicator calculations, strategy models and the backtest system.
| Role | Service design and development, financial data collection and processing, indicator and strategy rules, backtest engine development and testing |
|---|---|
| Experience | Three years of software development, and years of personal investing in Korean and overseas stocks and crypto |
| Focus Areas | Financial data processing, quantitative investment analysis, technical indicators, backtesting systems |
Confluence Zone does not copy a single strategy or source. It turns ideas from published investment theory, financial analysis, technical analysis and market research into its own numeric rules, and checks them with its own backtests.
They were not chosen for good performance. They were picked from different families so that they reach opposite judgments.
| Trend 3 | Trend breakout, new-high growth momentum, trend template. They buy what has risen |
|---|---|
| Contrarian 2 | Contrarian bargain buying, contrarian low P/E. They buy what has fallen |
| Value 3 | Margin-of-safety value, quality value, absolute-return margin of safety |
| Growth 1 | Growth at a reasonable price |
| Quant 1 | Quant quality & value |
Adding only trend-following or only value strategies would tilt the average and defeat the purpose of seeing several. So the design rule is not to grow one family alone. Each strategy looks only at the criteria of its own source; strategies do not share indicators or get averaged together, because the conflict is intended.
The regime is not mixed into the score. Mixing them creates states like "downtrend but 55 points", where each hides the other. Every verdict uses daily bars, whatever timeframe the chart shows.
Sources: Financial Services Commission public data, FSS DART, Binance, Yahoo Finance, Marketstack (EOD), TradingView and others. Confluence Zone does not take price and financial data from a single place. It checks each source's license (commercial use, display and processing) and mixes several sources, which may change by market, purpose and time. Analysis and backtest screens show the data timestamp instead of the source. Prices are adjusted for splits.
| Indicator formulas | Same as the TradingView standard (EMA seeded with an SMA, RSI with Wilder smoothing) |
|---|
Verified: the confluence score (technical indicators) has been backtested. Signals are decided at the close and filled at the next bar's open, and automated tests check that no future data is used.
Not verified: the strategy panel scores have not been backtested, because historical financial data for past dates is not available, so only the present can be scored. The master strategies in the backtest (Minervini, Livermore, O'Neil, Templeton) are separate models of their price and volume rules only, not a test of the panel scores. The regime gate is also for diagnosis and display and is not used as a backtest entry condition.
Keeping this distinction clear is a line this service holds. Backtest results are statistics on past data and do not guarantee future returns.
The assumptions behind the numbers on the backtesting screen and in Questions and answers. Different assumptions give different results.
| Universe | Daily bars of one stock chosen on screen. Several strategies (1 to 3 columns) are compared side by side over the same period, capital and cash flows. It is not a multi-stock scan or a portfolio backtest |
|---|---|
| Strategies | Threshold crossings of the confluence score (nine indicators), models of the price and volume rules only of master strategies (Minervini, Livermore, O'Neil, Templeton), custom rules (moving-average cross, RSI recovery, price breakout), and the wiki basket and leverage models (US TQQQ only). Strategy panel scores and the regime gate are not used |
| Fills | Signals are decided at the daily close and filled at the next trading day's open. The incomplete current bar is not used. Stops and trailing stops are checked against the bar's low during the day, and a gap past the line fills at the open |
| Capital rules | Initial capital, plus an optional deposit on the first trading day of each month (the same for all strategies, off by default). Signal strategies go all in on a buy signal and all to cash on exit by default. With the allocation strategy on, they start at 70 stocks / 30 cash and move only within ±10 points (60 to 80% stocks): to the upper band on a buy signal and the lower band on exit. The basket and leverage models follow their own rules (55 to 85% stocks for leverage) |
| Costs | A one-way cost of 0.15% (fees, taxes and slippage combined) applies to every fill. It can be changed on screen |
| Dividends | Dividends are included by default. Fills and valuations use prices that reinvest each dividend on its ex-dividend date (adjusted prices), and the last price matches the actual one. Trading signals use the same prices as the chart (adjusted for splits only). The basket model fills slots with whole shares at actual prices, so it leaves dividends out. Dividend taxes are not deducted, and crypto is not affected |
| Price check | The 1-, 3-, 5- and 10-year annualized returns of QQQ, SPY, QLD and TQQQ calculated from these dividend-adjusted prices were checked against the official market-price performance published by each issuer as of 2026-08-31. The differences were 0.13 percentage points or less |
| Default rules | No stop, trailing stop or take-profit. Positions exit only on the strategy's own exit signal (all to cash, or to the lower band with the allocation strategy on). Stops can be turned on for each strategy column |
| Currency | Amounts are entered and totaled in dollars. Korean stocks are calculated with won capital at a fixed exchange rate (1,400 by default), converted back to dollars, and only the final value also shows a won amount |
| Output metrics | Return on money invested, time-weighted return (deposit timing removed), annualized return (over 1 year), maximum drawdown (MDD), number of fills and the trade list. The benchmark is buy and hold of the same stock; the wiki models also show QQQ and QLD holding lines |
| Look-ahead prevention | Runs start after the indicator warm-up, and automated tests check that changing future data does not change past orders |
What is not reflected (limits)
① Tick size, fill liquidity and trading halts (fills at the daily open are assumed) ② Exchange-rate changes (fixed rate), interest on cash and separate taxes, including taxes on dividends ③ Risk-adjusted measures such as the Sharpe ratio are not yet provided ④ There is no in-sample/out-of-sample split, so repeatedly tuning settings on the same data risks overfitting; the default ladder portions of the leverage model were also chosen by the operator after looking at 2011 to 2026 data ⑤ Only currently listed stocks can be queried, which is survivorship bias. Financial data is not used in backtests, so reporting-lag issues do not apply.
Investment decisions and their results are your own responsibility. See the investment notice for details.
If you find a calculation or data error, please tell us at the official email. Once an error is confirmed and fixed, the rule is to publish when it was found, when it was fixed, what it affected and what changed on this page. Rather than promising no errors, we publish how errors are handled.
This site currently has no ads, affiliate links, paid stock placements or sponsored content. If any appear, they will be labeled where they appear. Membership only adds custom settings and alert automation; the analysis engine, scores and backtesting are the same for guests, free members and members. Accuracy is not what is sold.
The 25 default symbols work without logging in, and all symbols with a free login. Membership pricing and scope are on the membership screen.
