How often our price ranges held
Every coin page links to price ranges built from the coin’s own volatility. Here is how those ranges would have done in the past: every week, for every coin with enough history, we built the ranges the method gives and checked where the price really was 30 and 90 days later.
Updated · Replayed on daily closes (Bitstamp for Bitcoin, Binance for other coins) · How it's calculated
Across 71 coins and 17,769 weekly tests, the wide range (aimed at 90%) contained the price 30 days later 91.5% of the time, and the likely range (aimed at 68%) 74.2% of the time. Over 90 days the figures were 90.6% and 71.7%. When the price ended outside the wide range it was above it 1,012 times and below it 500 times, so the ranges miss rallies more often than falls.
By market size
| Coins | Coins tested | Wide range held, 30 days | Likely range held, 30 days | Wide range held, 90 days | Misses above / below (30 days) |
|---|---|---|---|---|---|
| Top 10 | 6 | 89.5% | 74.1% | 88.5% | 193 / 83 |
| Ranks 11–50 | 20 | 91.9% | 74.9% | 91.4% | 320 / 124 |
| Ranks 51–150 | 45 | 91.8% | 73.8% | 90.7% | 499 / 293 |
Coin by coin
Sorted by how often the wide range held over 30 days; select a column heading to re-sort. A coin far below 90% has had more sudden moves than its own past volatility suggested.
| Coin | |||||
|---|---|---|---|---|---|
| Pepe since 2024 | 38 | 128 | 97.7% | 76.6% | 97.5% |
| Floki Inu since 2024 | 120 | 128 | 97.7% | 76.6% | 97.5% |
| Render since 2025 | 63 | 64 | 96.9% | 71.9% | 92.9% |
| Pyth Network since 2025 | 91 | 89 | 96.6% | 71.9% | 98.8% |
| Lido DAO since 2023 | 105 | 180 | 96.1% | 78.3% | 97.7% |
| Shiba Inu since 2022 | 33 | 232 | 95.7% | 83.2% | 93.7% |
| Decred since 2021 | 118 | 272 | 95.6% | 78.7% | 97.7% |
| dogwifhat since 2025 | 140 | 85 | 95.3% | 76.5% | 100.0% |
| Celestia since 2024 | 97 | 103 | 95.1% | 68.9% | 92.6% |
| Bittensor since 2025 | 34 | 79 | 94.9% | 79.7% | 98.6% |
| Aave since 2021 | 39 | 261 | 94.3% | 72.4% | 95.7% |
| LayerZero since 2025 | 90 | 69 | 94.2% | 76.8% | 95.1% |
| Pendle since 2024 | 106 | 120 | 94.2% | 72.5% | 98.2% |
| JasmyCoin since 2022 | 126 | 204 | 94.1% | 77.9% | 92.3% |
| Worldcoin since 2024 | 71 | 117 | 94.0% | 77.8% | 97.2% |
| Tezos since 2020 | 108 | 317 | 94.0% | 77.9% | 93.2% |
| Litecoin since 2018 | 26 | 410 | 93.9% | 72.7% | 94.0% |
| Nexo since 2023 | 79 | 181 | 93.9% | 76.2% | 90.8% |
| Ethereum Name Service since 2022 | 133 | 206 | 93.7% | 74.8% | 93.4% |
| Algorand since 2020 | 64 | 330 | 93.6% | 73.0% | 91.3% |
| Conflux since 2022 | 122 | 238 | 93.3% | 81.1% | 91.7% |
| Cosmos since 2020 | 67 | 338 | 93.2% | 75.7% | 90.6% |
| Compound since 2021 | 123 | 277 | 93.1% | 79.1% | 97.0% |
| Basic Attention Token since 2020 | 136 | 346 | 93.1% | 77.2% | 93.2% |
| TRON since 2019 | 8 | 384 | 93.0% | 79.4% | 93.9% |
| JUST since 2021 | 42 | 271 | 93.0% | 81.9% | 95.8% |
| Sei since 2024 | 107 | 114 | 93.0% | 70.2% | 94.3% |
| Uniswap since 2021 | 27 | 265 | 92.8% | 74.0% | 93.8% |
| Polkadot since 2021 | 50 | 270 | 92.6% | 73.0% | 92.0% |
| Arweave since 2022 | 124 | 231 | 92.6% | 74.5% | 88.8% |
| Quant since 2022 | 35 | 220 | 92.3% | 81.4% | 92.9% |
| Stellar since 2019 | 23 | 385 | 92.2% | 80.3% | 93.6% |
| Starknet since 2025 | 137 | 87 | 92.0% | 74.7% | 93.6% |
| Trust Wallet Token since 2022 | 130 | 247 | 91.9% | 75.7% | 95.0% |
| ChainLink since 2020 | 19 | 353 | 91.8% | 70.0% | 92.7% |
| Internet Computer (IOU) since 2022 | 45 | 232 | 91.8% | 76.7% | 89.7% |
| Dash since 2020 | 77 | 342 | 91.8% | 78.7% | 93.7% |
| Ethereum Classic since 2019 | 54 | 384 | 91.7% | 75.5% | 93.3% |
| Gnosis since 2022 | 103 | 216 | 91.7% | 77.8% | 82.1% |
| XRP since 2019 | 6 | 389 | 91.5% | 77.4% | 93.7% |
| Dogecoin since 2020 | 13 | 328 | 91.5% | 77.4% | 85.0% |
| Bitcoin Cash since 2020 | 24 | 307 | 91.5% | 76.5% | 89.3% |
| Optimism since 2023 | 131 | 177 | 91.5% | 72.3% | 88.7% |
| Decentraland since 2021 | 149 | 271 | 91.5% | 73.1% | 91.3% |
| The Graph since 2021 | 121 | 252 | 91.3% | 67.5% | 90.6% |
| Injective Protocol since 2021 | 66 | 261 | 91.2% | 73.2% | 88.9% |
| Curve DAO Token since 2021 | 99 | 270 | 91.1% | 77.0% | 93.1% |
| PancakeSwap since 2022 | 70 | 243 | 90.9% | 74.1% | 87.7% |
| Hedera Hashgraph since 2020 | 32 | 316 | 90.8% | 78.2% | 91.2% |
| Terra Classic since 2023 | 111 | 162 | 90.7% | 74.7% | 92.9% |
| THORChain since 2021 | 117 | 267 | 90.6% | 71.5% | 88.8% |
| Binance Coin since 2018 | 4 | 415 | 90.4% | 75.2% | 88.7% |
| ether.fi since 2025 | 102 | 83 | 90.4% | 77.1% | 100.0% |
| Fetch.ai since 2020 | 89 | 346 | 89.9% | 69.1% | 80.2% |
| Stacks since 2020 | 78 | 312 | 89.7% | 70.8% | 82.9% |
| Filecoin since 2021 | 80 | 261 | 89.7% | 73.6% | 92.9% |
| Theta Token since 2020 | 141 | 341 | 89.7% | 68.6% | 84.0% |
| Zcash since 2020 | 11 | 343 | 89.5% | 68.8% | 91.0% |
| Avalanche since 2021 | 28 | 265 | 89.4% | 71.7% | 89.1% |
| Cardano since 2019 | 21 | 392 | 89.3% | 70.2% | 86.7% |
| NEAR Protocol since 2021 | 25 | 262 | 89.3% | 71.0% | 86.6% |
| Ethereum since 2018 | 2 | 426 | 89.0% | 69.0% | 90.4% |
| Ethena since 2025 | 56 | 81 | 88.9% | 61.7% | 93.1% |
| VeChain since 2019 | 81 | 378 | 88.9% | 67.2% | 85.4% |
| Immutable since 2023 | 112 | 197 | 88.8% | 68.5% | 85.6% |
| Solana since 2021 | 7 | 271 | 88.6% | 72.3% | 92.4% |
| Aptos since 2023 | 92 | 157 | 87.3% | 68.2% | 89.2% |
| Sui since 2024 | 31 | 129 | 86.8% | 66.7% | 85.8% |
| Bitcoin since 2012 | 1 | 739 | 86.7% | 72.7% | 80.6% |
| Arbitrum since 2024 | 61 | 134 | 85.8% | 62.7% | 83.3% |
| Raydium since 2022 | 84 | 219 | 85.4% | 71.2% | 81.4% |
How this test works
For each coin and each week with at least a year of earlier daily closes, we take the standard deviation of the previous 365 daily log returns (σ) and build the two ranges exactly as the coin’s range page does: today’s price × e±zσ√T, with z = 1 for the likely range and 1.645 for the wide range, for T = 30 and 90 days. We then check whether the close T days later fell inside. A well-calibrated model would land inside the wide range about 90% of the time and inside the likely range about 68% of the time.
Two limits. The test uses history the model could not have known about in advance, which is the point, but overlapping weekly windows are not independent, so the figures are less precise than the large test counts suggest. And past calibration says nothing certain about the next range: crypto regularly produces moves bigger than its own history implies, mostly upward in this sample.
Full method: price ranges. Ranges are not forecasts or advice.