Every flip we have recorded for Gate, and what the price did next.
A candle is one bar of price over a fixed time: 4 hours, a day, a week or a month. Only closed candles count. The flip level is a price line that follows the trend. When a candle closes on the other side of it, the trend flips from bullish to bearish, or back. Each flip is written when its candle closes and is never edited afterwards.
What happened after Gate's weekly flipsCounted: "4 of 5 flips" means four of the five flips are old enough for that many days to have passed. Each median uses only those. With an even number of flips we take the lower of the two middle results.Range of the median: the flips we have are only a sample, so their median could have come out a little different. This range shows how far it could plausibly have moved. The percentage in brackets is how often a range built this way would hold the median of many more flips like these. With fewer than 6 flips the range simply runs from the lowest result to the highest.Ordinary days: we took every ordinary day in the same period and measured the price change over the same number of days. This is the range where the median of that many ordinary days would land 95 times out of 100. "Against ordinary days" says whether the median after the flips is inside that range (it looks like an ordinary day) or outside it. About 1 in 20 falls outside by chance alone.Same dates: for each flip, the median result of all tracked coins over the same days, and the share of them that went up. These are medians and shares; by averages the picture can differ. Coins tend to rise and fall together, so flips that fall on the same days are not independent evidence: they count as fewer separate periods than flips.The range for these rows is a 95% interval: if whole calendar months of flips were drawn again at random, many times over, the median would land inside it 95 times out of 100. A row with fewer than 30 separate periods or 30 months behind it says "too few independent periods to compare" instead.
After bearish flips the median price was still higher 90 days later. A bearish flip did not by itself mean the price went down. Both medians are within the range of ordinary days.
Green = the price rose, red = it fell, whatever the flip said. Each change is measured from the close of the flip candle.
| Days later | Counted | Median change | Range of the median | Ordinary days | Against ordinary days |
|---|---|---|---|---|---|
| After bullish flips | |||||
| 7 days | 8 of 8 flips | +6.9% | +0.1% to +23.4% (93%) | −5.8% to +5.0% | Outside range |
| 30 days | 8 of 8 flips | +12.9% | −6.6% to +122.8% (93%) | −12.9% to +12.5% | Outside range |
| 90 days | 7 of 8 flips | +12.7% | −32.4% to +737.5% (98%) | −21.1% to +51.7% | Within range |
| After bearish flips | |||||
| 7 days | 7 of 7 flips | −4.8% | −18.5% to +18.3% (98%) | −5.0% to +7.4% | Within range |
| 30 days | 7 of 7 flips | 0.0% | −28.6% to +57.9% (98%) | −10.6% to +18.3% | Within range |
| 90 days | 7 of 7 flips | +17.0% | −11.0% to +241.4% (98%) | −21.1% to +51.7% | Within range |
| Any day since 9 Feb 2020, for comparison | |||||
| 7 days | 2,424 days | +0.1% | – | – | – |
| 30 days | 2,401 days | +0.8% | – | – | – |
| 90 days | 2,341 days | +2.0% | – | – | – |
| All tracked coins on the same dates as the bullish flips | |||||
| 7 days | 8 flips on 8 days | too few independent periods to compare | too few independent periods to compare | 56% up | – |
| 30 days | 8 flips on 8 days | too few independent periods to compare | too few independent periods to compare | 56% up | – |
| 90 days | 7 flips on 7 days | too few independent periods to compare | too few independent periods to compare | 37% up | – |
| All tracked coins on the same dates as the bearish flips | |||||
| 7 days | 7 flips on 7 days | too few independent periods to compare | too few independent periods to compare | 48% up | – |
| 30 days | 7 flips on 7 days | too few independent periods to compare | too few independent periods to compare | 63% up | – |
| 90 days | 7 flips on 7 days | too few independent periods to compare | too few independent periods to compare | 73% up | – |
About 1 in 20 medians falls outside the ordinary range by chance alone.
Coins tend to rise and fall together, so flips on the same days count as fewer separate periods than flips.
The latest flip is too recent to count in the 90 day column.
Prices are weekly closes on Gate. Returns are measured from the close of the flip candle.
What we tested
We tested whether other settings for the trend line would have separated what followed bullish and bearish flips better than the published one. We found no setting that was clearly better.
We tried 56 different settings and filters on daily and weekly candles (100 combinations of rule and candle size), on 129 coins with at least three years of history (128 without Bitcoin for the settings grid). 4-hour and monthly lines and the Bitcoin-priced lines were not tested. The main test period is January 2022 to October 2026: one mostly falling stretch, using only coins that are still listed.
For altcoins on daily candles, after 2022, line factors from 2 to 3 gave similar results: the 30-day gap between what followed bullish and bearish flips was 2.4 to 5.0 points, against 3.4 for the published setting, and none of the 90-day differences was clear. Before 2022 the published setting was at the low end of that range, and a factor of 1.5 was clearly worse in both periods. For Bitcoin, one series with 45 daily flips since 2022, the evidence is too thin to defend or replace its setting. That the published setting sat in a flat area does not show it was chosen without hindsight.
Rules that wait for extra closes, demand a margin beyond the line, or use a slower factor changed the label less often or later. They did not clearly improve how well it separated what followed after 2022, and the margin rules fell further from the peak before turning bearish. Filters that go neutral when the trend looks weak made the label change more often, not less (daily: 15 to 86 changes per coin per year, against 14), and the flips they let through were followed by much the same results as the rest.
For altcoins on weekly candles the published setting sits next to a cliff: a wider band turned later and separated what followed less (the 90-day gap fell by 12 to 26 points; the 95% intervals are wide, because they rest on only 19 calendar quarters). Bitcoin's weekly line has only 4 flips since 2022 and shows no such pattern.
Coins tend to turn together: on 69% of the days we looked at (each coin on each day), a coin's daily trend matched Bitcoin's. About 50% would be expected by chance; it was 63% to 65% in 2025 and 2026, and 80% in 2022. After 2022 the 30-day median gap between bullish and bearish flips was +3.3 points, but “about zero” (−0.2 points, 95% interval −0.7 to +0.2) once each flip is compared with all coins on the same dates. That holds for medians; by averages the same-date gap is +2.4 points (95% interval 0.2 to 5.2), driven by a few large rises. The 95% interval for the raw gap includes zero.
Flips undone within 3 candles (3 weeks)
A flip is undone when the next flip, back the other way, comes within 3 candles (3 weeks). A flip is only counted once 3 candles have closed after it.
| Which flips | Flips | Undone within 3 candles (3 weeks) |
|---|---|---|
| After bullish flips | 8 | 0 of 8 |
| After bearish flips | 7 | 0 of 7 |
Gate's weekly line has flipped 2.3 times a year since 9 Feb 2020.
Gate, weekly flips: 15 flips
Each figure is the price change from the flip price: 7, 30 and 90 days later, and at the next flip. Green means the price rose and red that it fell, whatever the flip said. Closed past the line: how far the flip candle closed beyond the flip level it crossed. Score agreed: when the trend score, a blend of six slower measures of the trend, first reached a bull score after a bullish flip or a bear score after a bearish one. The flip line reacts faster, so the score often agrees a few candles later.
| Candle closed | Flipped to | Closed past the line | Score agreed | Price at the flip | 7 days later | 30 days later | 90 days later | Until the next flip |
|---|---|---|---|---|---|---|---|---|
| 23 Aug 2026Lasted 6 weeks so far | Bull | +2.1% | after 2 weeks | $7.83 | +2.7% | +41.4% | not yet | +42.4% so far |
| 15 Jun 2025Lasted 62 weeks | Bear | −5.3% | after 3 weeks | $16.71 | −4.8% | −6.4% | +3.7% | −53.1% |
| 12 Nov 2023Lasted 83 weeks | Bull | +4.0% | after 3 weeks | $4.22 | +0.1% | +12.9% | +12.7% | +296.3% |
| 4 Jun 2023Lasted 23 weeks | Bear | −6.7% | after 1 week | $4.26 | −5.6% | +3.1% | −8.6% | −0.9% |
| 22 Jan 2023Lasted 19 weeks | Bull | +3.4% | after 4 weeks | $4.05 | +6.9% | +22.6% | +29.2% | +5.1% |
| 12 Jun 2022Lasted 32 weeks | Bear | −7.9% | at the flip | $4.86 | −18.5% | −22.1% | −11.0% | −16.6% |
| 31 Oct 2021Lasted 32 weeks | Bull | +19.3% | at the flip | $6.30 | +18.2% | +9.2% | +7.5% | −22.8% |
| 23 May 2021Lasted 23 weeks | Bear | −45.5% | after 5 weeks | $3.89 | +11.2% | −28.6% | +17.0% | +61.9% |
| 4 Apr 2021Lasted 7 weeks | Bull | +7.5% | at the flip | $2.34 | +82.3% | +122.8% | +42.1% | +66.5% |
| 28 Feb 2021Lasted 5 weeks | Bear | −1.0% | not before the next flip | $1.25 | +18.3% | +57.9% | +241.4% | +86.4% |
| 31 Jan 2021Lasted 4 weeks | Bull | +9.1% | at the flip | $0.5863 | +23.4% | +135.1% | +737.5% | +113.7% |
| 1 Nov 2020Lasted 13 weeks | Bear | −5.2% | at the flip | $0.4540 | −7.9% | 0.0% | +23.3% | +29.1% |
| 16 Aug 2020Lasted 11 weeks | Bull | +4.3% | at the flip | $0.6243 | −4.3% | −6.6% | −32.4% | −27.3% |
| 15 Mar 2020Lasted 22 weeks | Bear | −18.7% | at the flip | $0.3601 | +2.2% | +17.7% | +44.2% | +73.4% |
| 9 Feb 2020Lasted 5 weeks | Bull | +6.6% | after 1 week | $0.5705 | +12.2% | −11.1% | −21.1% | −36.9% |
How to read this
Closed candles only
Flips are computed on closed candles only, so a flip never appears and then vanishes while a candle is still open. The ledger is append-only. When the rules change, the rule set gets a new version number and the old entries stay exactly as they were.
Compare with any day
The comparison rows show what the price did after any day since 9 Feb 2020, from the first flip on these candles on, so they cover the same period as the flips. A row that rests on fewer starting days than four times its number of days says too little history.
Survivorship
Only coins in today's list have a record here. Coins that left the list before 1 Oct 2026 are missing, which tends to make past results look better than they were. Each coin's any-day row has the same gap, so compare the two.
A small sample
15 flips are a small sample. The numbers describe the past and are not a forecast, and nothing here is advice.