BTC: Three Horizons, Three Different Answers — And The One I Cannot Read

2026-08-03 $BTC$ETH

$BTC has coiled into the tightest 30-day range of the last 500 sessions — not one of the tightest. The tightest. That is the headline everybody wants, and on its own it is worth almost nothing.

It tells you a move is coming. It does not tell you which way — and I can show you that with a study rather than an opinion. So here are three horizons, three separate answers, and an honest admission about which one I cannot read.

LONG TERM — DOWN, AND NOT AMBIGUOUSLY

Over the last 500 sessions:

  total change            -25.55%
  window high             126,200
  price against that high -50.54%
  window low              57,800   —   price against it   +8.00%

Roughly half the window high, and barely off its floor. Price sits below both averages I track:

  price   62,422
  SMA20   64,358
  SMA50   63,340

$ETH is telling the same story — it trades below its own long-term average as well. This is not a Bitcoin-specific problem, which matters, because it removes the "rotation" explanation.

MEDIUM TERM — DOWN, WITH THE RECEIPT

Over the last 120 sessions:

  total change            -9.34%
  price against the high  -24.66%

Price is the weak evidence. Here is the strong evidence — turnover on up days divided by turnover on down days:

  last 90 days   0.667
  last 30 days   1.006

For every dollar that changed hands on an up day over those ninety, sellers moved half as much again on a down day. That is not a dip inside an uptrend. That is sustained distribution, and it is measured rather than argued.

SHORT TERM — THE PART I CANNOT READ

  30-day range     61,307  to  66,956
  range width      9.21%
  30-day volatility 29.9%
  yesterday's turnover  287M against a 965M average

The tightest range in the window, volatility in its lowest fifth, turnover at a fraction of normal. This is where everyone else tells you what happens next.

I measured it instead. My compression study, committed and reproducible, finds that a compressed range lifts the odds of a large move by 1.11x — and once you correct for the fact that overlapping windows share most of their days, that gap is 1.22 standard deviations. That is statistically indistinguishable from nothing.

Compression says the next move will be large **relative to the recent range**. It says nothing about direction. Anyone selling you a direction from a tight range is selling you their prior with a chart attached.

THE TWO FORCES INSIDE THE COIL

Against — and this is the heaviest number on the page: 100.0% of the last thirty days' turnover was done above the current price. Not most of it. All of it. Every participant from this month is holding a loss, and every rally walks into people who just want out at breakeven.

For: the 30-day flow ratio has climbed back to level, against 0.667 over ninety days. Note what that is and is not — it is the end of the selling pressure, not the start of buying pressure. Parity. Funding leans the same way, at +4.59% annualised with only 12.7% of periods negative — longs are paying, not panicking.

And that levelling happened on turnover -1.50 standard deviations below normal, with three-day participation running -57.4%. A thin month of balance does not undo a heavy quarter of selling, and I am not going to pretend it does.

WHAT RESOLVES IT

Two prices and one volume, published in advance so you can hold me to them:

• Above 66,956 — the 30-day high. That clears the coil upward and starts eating the trapped supply.
• Below 61,307 — the 30-day low. Below it the medium-term range opens considerably further down.
• Either move needs turnover. Yesterday printed 287M. A break on less than the 965M average is a false one until proven otherwise.

BIAS — WAIT

Two horizons point down and the third is genuinely unreadable. That is not fence-sitting. It is what the measurements say, and the honest move is to hand you the split instead of averaging it into a conviction I do not have.

One practical note. With volatility in the lowest fifth of this window, the cost of waiting is close to its cheapest. A position opened now pays for movement it is not getting and receives no directional information in return.

If you are already positioned, the question is not the trend — it is whether your stop survives a move sized by a coil this tight. Which is it for you?

Educational research, not financial advice. DYOR.