How this is built, and what it means

Over Bitcoin's whole life, log price has tracked a straight line against log age — price grows as a fixed power of days since genesis. We fit that line and draw a corridor from 0.42× to 4× around it — the range price has held since 2011 (the infant 2010 market briefly broke below the floor to ~0.28×, and the early manias blew through the ceiling). The deviation you see is just price ÷ fair − 1.

The catch · It's a descriptive fit to the past, not a law. The corridor is drawn from the same history it's judged against, and nothing forces the next decade to obey it — long-run context, not a target.

Its own log-space regression — a separate, slightly steeper fit than the corridor's power law (price ∝ age^6.1 vs ^5.8, so the two centers disagree) — fanned into seven affordability bands by fixed log-price offsets — deep-value blue at the bottom, bubble red at the top. We tuned the offsets against the real price residuals so band occupancy mirrors history: the deepest ~2% of days land in Fire Sale, the 2011–2014 manias in Bubble red (each later cycle has crested lower — 2017 and 2021 never reached it), and most of the time sits in the middle.

The catch · It's folk-famous and deliberately unscientific — a colour mood-ring laid over its own curve, which is NOT the corridor's fair line (the two models can and do disagree). The band you're standing in is a vibe, not a signal and not advice.

Each halving is reset to day zero and price replotted as a multiple of the halving-day price, so the four epochs line up on one axis. It shows how the current cycle is tracking against the three before it — and that each cycle's gains have so far come in smaller than the last.

The catch · Four cycles is a tiny sample and past cycles are not a forecast. Diminishing returns is a pattern, not a promise, and aligning by halving is one choice among several.

Month-over-month returns (month-end close ÷ prior month-end − 1) laid out as a years × months heatmap, with the MEDIAN per calendar month quoted as the typical read — the median because early moon-shots like November 2013's +451% drag the mean around.

The catch · Sixteen-odd years of one asset is a tiny sample. Seasonality is a rear-view pattern, not a schedule; any single month can and does break it.

The underwater curve — each day's close as a percent below the running all-time high — plus every episode of −40% or worse: its depth, time down, and time from trough back to a new high.

The catch · That Bitcoin has recovered from four falls of 75%+ is survivorship, not a law. Recovery has never been guaranteed and past bears don't bound the next one.

Realized volatility — the annualized standard deviation of daily log returns over trailing 30- and 90-day windows. The long arc bends down as the market deepens, from ~160% in 2011 toward equity-multiple levels today.

The catch · Realized vol looks strictly backward. A calm reading is not a safety guarantee — quiet stretches have preceded both grinds and sharp breaks.

Deterministic protocol arithmetic, not market data: the 50-BTC subsidy halving every 210,000 blocks, anchored to the four known halving heights and interpolated at ~146 blocks/day, giving supply, %-mined, and the inflation rate at any date.

The catch · "Mined" counts coins issued, not coins still spendable — millions are estimated lost forever, so the effective float is smaller. Future halving dates are estimates; heights are exact.

Every entry day in the series, held for 1-8 years: the share that ended in profit, the median outcome, and the single worst entry — all-history and modern-era (2017+) cohorts side by side, because early micro-cap years inflate medians.

The catch · Overlapping entries aren't independent trials — it's a handful of regimes sliced daily. And the whole table exists only because Bitcoin survived; survivorship is the biggest assumption.

Replays a recurring buy (your amount, weekly or monthly, from a chosen era) against weekly closes: sats stacked, cost basis, value now, the journey's worst drawdown — and always the honest rival, the same total as a lump sum on day one.

The catch · No fees, spreads, or taxes are modelled, so real results run slightly worse. And a backtest replays one past — it says nothing about the window that starts today.

Difficulty is the target Bitcoin sets so blocks are found about every ten minutes on average — an exact protocol value in every block header, re-adjusted once every 2,016 blocks (~two weeks) from how fast the last epoch's blocks arrived. We read the latest value from mempool.space.

The catch · It's precise but coarse: the number only moves once per ~2-week adjustment, so between epochs it holds a flat step rather than updating live. And high difficulty is a floor on honest mining effort per block — not the dollar cost to attack the chain, which depends on rentable hashpower, energy, and price.

Total hashrate can't be measured directly — no registry counts every machine. It's inferred from difficulty and how quickly blocks are actually found over a recent window: when blocks arrive faster than the ten-minute target, more hashpower is implied. We use mempool.space's estimate, expressed in EH/s.

The catch · It's a statistical estimate, not a counted quantity — block discovery is random, so short-window figures are noisy and can swing several percent before settling, which is why hashrate never shows as live.

Hashprice gauges miner economics: revenue per unit of hashrate. We compute it subsidy-only and difficulty-anchored — today's per-block subsidy × the blocks per day × the BTC price, divided by the network hashrate implied by difficulty — and quote it in USD per PH/s per day, the standard public unit.

The catch · It's subsidy-only, so it understates real revenue by leaving out transaction fees (which spike during congestion), and it divides by a difficulty snapshot up to a week old — so even when the price leg is current, the reading isn't live and won't exactly match fee-inclusive indices like Luxor's.

The electricity cost to mine one bitcoin, computed from our own hashrate and the protocol issuance: network power (hashrate × an assumed fleet efficiency in J/TH) gives daily energy, spread over the coins issued that day, priced at an assumed $/kWh. We show it as a band across fleet efficiency (~15 J/TH newest ASICs to ~32 older gear) rather than a single number, with the BTC price drawn through it.

The catch · It rests on two assumptions we state up front — fleet efficiency and a $0.05/kWh electricity price — and it counts electricity only, not hardware, cooling, or overhead, so it is a soft floor, not a hard one. It is deliberately NOT the Cambridge CBECI dataset (which is non-commercially licensed); it's our own arithmetic, and a model rather than a measurement.