METHOD · WORKED EXAMPLES
How the percentile and BTC-relative return are calculated
The displayed figures summarize outcomes after similar historical conditions; they do not forecast the profit of a trade opened today. The prices below are hypothetical examples used only to explain the calculation.
1. The BTC-to-altcoin ratio and its percentile
If BTC costs 50,000 USDT and an altcoin costs 100 USDT, the ratio is 500. We compare the ratio with up to 365 daily ratios computed from both coins on matching dates. If 95 of 100 past observations are at or below 500, the current value is at the 95th percentile. A high ratio can also mean prolonged weakness in the altcoin; it does not by itself predict a rebound.
2. A worked BTC-relative return example
Suppose that after 30 days BTC rises from 50,000 to 55,000 USDT (+10%) while the altcoin rises from 100 to 108 USDT (+8%). Divide the altcoin growth factor of 1.08 by BTC's 1.10, then subtract 1 and the fixed cost assumption of 0.001. The result is approximately -1.92%. The altcoin gained in USDT but lagged BTC.
Alternatively, if BTC falls from 50,000 to 45,000 USDT (-10%) and the altcoin falls from 100 to 96 USDT (-4%), the calculation is 0.96 ÷ 0.90 - 1 - 0.001 ≈ +6.57%. The altcoin outperformed BTC even though its USDT value fell by 4%.
3. Median, win rate and sample count
For each 7, 14 and 30-day horizon, the tool selects past dates whose percentile was within five points of today's, then measures the later BTC-relative outcome. The result shown is the median; the win rate is the share of observations above zero after the fixed cost assumption. If 12 out of 20 observations were positive, the historical win rate is 60%. That is not a validated 60% probability for the next trade.
Two dates one day apart have largely overlapping 30-day outcome windows. Twenty observations therefore do not equal twenty independent experiments.
4. Reading the signal
A “strong candidate” requires all five conditions: percentile at least 95, 30-day historical median at least +2%, historical win rate at least 60%, at least 15 observations and a positive seven-day rebound against BTC. The opportunity score combines percentile, win rate, median and sample size to sort results. It is not a probability of profit. Check the timestamp and sample count first, then consider USDT gains or losses and your actual trading costs separately.