Trend Flip Scanner
Counts before filters
The market on 4h
Priced inA typical 4h up leg has lasted about 3 days, and the 4h line flips about 8 times a month.
6 more timeframes · not yet measured
- 15mnot yet measured
- 30mnot yet measured
- 1hnot yet measured
- 2hnot yet measured
- 6hnot yet measured
- 8hnot yet measured
Coins
loading the live boardWhat changed
Loading flips.
How it works
The line is SENTINEL Trend's trailing price boundary. Cells use the public Pine presets below on confirmed closes.
HA means Heikin Ashi, a smoothed candle; 12h direction and distance use the HA close, while the chart plots real closes.
At the history edge, the flip predates the loaded window and age is a lower bound; there is no flip date or bitcoin stamp, and age filters exclude it.
| Timeframe | Preset | Stretched past | On the web |
|---|---|---|---|
| 15m | Classic 10 by 3.0 Wilder hl2, Heikin Ashi on, limit 12 percent | not yet measured | not yet measured |
| 30m | Classic 10 by 3.0 Wilder hl2, Heikin Ashi on, limit 12 percent | not yet measured | not yet measured |
| 1h | Classic 10 by 3.0 Wilder hl2, Heikin Ashi on, limit 12 percent | not yet measured | not yet measured |
| 2h | Classic 10 by 3.0 Wilder hl2, Heikin Ashi on, limit 12 percent | not yet measured | not yet measured |
| 4h | 4H Swing 28 by 2.0 SMA hl2, limit 10 percent | loading | live |
| 6h | 4H Swing 28 by 2.0 SMA hl2, limit 10 percent | not yet measured | not yet measured |
| 8h | 4H Swing 28 by 2.0 SMA hl2, limit 10 percent | not yet measured | not yet measured |
| 12h | 12H Watch 28 by 4.0 SMA close, Heikin Ashi on, limit 25 percent | loading | live |
| 1D | Daily Confirm 28 by 3.5 SMA close, limit 25 percent | loading | live |
| 1W | Weekly Hold 14 by 1.0 SMA close, limit 15 percent | loading | live |
Priced in BTC. Priced in BTC reads the same presets on each coin divided by bitcoin, the series TradingView shows as BYBIT:COINUSDT.P/BYBIT:BTCUSDT.P. Every close is exact. The high and the low of a bar depend on how it is built: TradingView divides the two markets' one hour bars for its 4h and 12h charts and their daily bars for its daily and weekly charts, and a bar built from finer or coarser bars comes out taller or shorter, which moves the line and the flips. The board builds its bars the same way, and it compares prices the way TradingView's script does, where two prices closer than 0.0000000001 count as equal. So each timeframe says below how its bars are built and what was checked against TradingView's chart, and a reading marked estimate has not earned that match yet. The stretched limits were measured on dollar pairs, so they are never applied here.
Press B to switch between USDT and BTC, and the number keys 5, 8, 9 and 0 for 4h, 12h, 1D and 1W.
Grading. A leg runs from one confirmed flip close to the next, using real closes. Net is exit over entry minus one, times direction, less the 0.22 percent round trip once; zero or below is losing, and open legs stay out of the record.
The studies. The August 2026 study ran 24,960 backtests across 26 perpetuals and found 58 to 68 percent losing legs out of sample. The September 2026 study covered 148 perpetuals and 8.4 million legs; a coin's calmest past timeframe did not predict its calmest next one. The advisory exit tested on 2h to 8h long legs begun while bitcoin daily was down arms after a 5 percent gain and exits on the first close 2 percent below the highest close. The October 2026 study ran 576 settings on fifteen timeframes for ten coins and then checked 112 more coins over four years: no line whose legs typically lasted under a day paid after cost, and long legs paid more on average while the coin's own daily line was also up, a gap that a wider check of 546 perpetuals could not show to be reliable.