Bitcoin Cycle Indicators Compared: MVRV, Pi Cycle, NUPL, and the BTC/Gold Ratio
An honest side-by-side of the four most-cited Bitcoin cycle indicators — what each one actually measures, when it works, and where it breaks. No indicator wins; each one tells you something the others can't.
If you spend any time around Bitcoin macro analysis, four indicators come up repeatedly: MVRV Z-Score, Pi Cycle Top, NUPL, and the BTC/Gold ratio. They all claim to answer some version of the same question — where are we in the cycle? — but they measure very different things, and each has a specific failure mode that's easy to miss until you've seen it happen live.
This is an honest side-by-side. No indicator wins; each one tells you something the others can't. The point is to understand what you're actually looking at when you read a chart, and where each tool stops being reliable.
MVRV Z-Score
MVRV is short for Market Value to Realized Value. Market value is the familiar number — current price times circulating supply. Realized value sums every coin at the price it last moved on-chain, treating each UTXO as having a "cost basis." The ratio of the two, normalized by its standard deviation, is the Z-Score.
The intuition: when market value is far above realized value, most coins are in heavy paper profit and the market is statistically overheated. Historical tops cluster above a Z-Score of ~7. Historical bottoms cluster below 0.
What it gets right. MVRV is grounded in something economically real — aggregate cost basis. Holders sitting on 10× gains behave differently from holders underwater, and MVRV measures that gap directly.
Where it breaks. The thresholds drift. The 2017 top fired at Z > 7; the 2021 top fired at Z > 7 in April but the higher November top fired at only ~3. As Bitcoin matures and HODLer base widens, realized value catches up faster, compressing the Z-Score range. A 2026 top may not produce a Z > 7 print at all. The indicator needs recalibration each cycle, which means the published threshold is always backward-looking.
Pi Cycle Top Indicator
Pi Cycle is the simplest of the four: a 111-day SMA crossing above 2× the 350-day SMA. The name comes from the rough ratio (350 ÷ 111 ≈ π) but the math is arbitrary — what made it famous is that it fired within ~3 days of the cycle peaks in 2013, 2017, and 2021.
What it gets right. It's mechanical. You can verify any claim about Pi Cycle in five minutes with a price feed and Excel. It doesn't require trust in any analyst or platform.
Where it breaks. No theoretical underpinning, no out-of-sample validation, marks only tops not bottoms, and one failed cycle ends its credibility entirely. Useful as confirmation alongside other indicators; dangerous as a standalone signal.
NUPL (Net Unrealized Profit/Loss)
NUPL takes the same UTXO-level cost-basis data MVRV uses and expresses it as a psychological gradient. Each coin's unrealized P/L is summed across the network and divided by market cap. The result lives between roughly −0.25 (capitulation) and +0.85 (euphoria).
Glassnode broke the range into named zones: capitulation, hope/fear, optimism/anxiety, belief/denial, euphoria/greed. The zones are mnemonic, not statistical — they correspond to historical price-action regions, not to any underlying behavioral model.
What it gets right. NUPL is the cleanest expression of the same idea MVRV measures: how much paper profit is in the system. The zone labels make it intuitive for a non-quant reader.
Where it breaks. Same drift problem as MVRV — the boundaries between zones have crept lower with each cycle as the HODLer base matures. Long-term holders quietly accumulating during a bear market push realized prices higher, which compresses NUPL's upper range. A 2026 cycle top may print in "belief," not "euphoria."
BTC/Gold Ratio Regime
The BTC/Gold ratio is Bitcoin price divided by gold price per ounce. The regime indicator places today's ratio against its own trailing 4-year distribution and classifies the result: when the percentile is low, the ratio is in a Bitcoin-favored regime; when high, gold-favored.
The 4-year window is deliberate — it's roughly one Bitcoin halving cycle, so each comparison covers a full reward-schedule generation. Fixed thresholds become obsolete as Bitcoin matures and the ratio drifts upward; a rolling-window percentile sidesteps the calibration problem entirely.
What it gets right. No on-chain data dependency, no model of fair value, no fragile threshold to maintain. It measures cycle position rather than absolute valuation, which is a more answerable question. It also incorporates information from outside Bitcoin — when gold rallies on macro fear, the ratio drops, which can shift the regime even if Bitcoin's price is unchanged.
Where it breaks. It tells you nothing about absolute valuation. If Bitcoin and gold both double, the ratio is unchanged and the regime doesn't move — which may or may not be what you wanted to know. It's also regime-blind: it cannot see macro context (rates, regulation, institutional adoption) that affects how a given ratio level should be interpreted.
Honest Comparison
| Indicator | Data source | Recalibration needed? | Marks tops? | Marks bottoms? | Out-of-sample track record |
|---|---|---|---|---|---|
| MVRV Z-Score | On-chain (UTXO) | Yes, each cycle | Yes | Yes | 3 cycles, drift confirmed |
| Pi Cycle Top | Price only | Effectively none | Yes | No | 3 cycles, 1 near-miss |
| NUPL | On-chain (UTXO) | Yes, each cycle | Yes | Yes | 3 cycles, drift confirmed |
| BTC/Gold Regime | Price only (BTC + XAU) | None (rolling window) | Yes (probabilistically) | Yes (probabilistically) | 1 cycle out-of-sample, in progress |
Should You Combine Them?
A reasonable instinct is: if four indicators all flash red, that's a stronger signal than any one alone. There's some truth to this, but it's contaminated by a problem most cycle-indicator users underweight: they're not independent. MVRV and NUPL are both derived from the same on-chain cost-basis data. Pi Cycle is just price. The BTC/Gold ratio is price-only too. Three of the four are mathematical transforms of the same underlying source (BTC price), with different smoothing.
Combining them via simple agreement ("3 of 4 say top") is closer to triple-counting price than to genuine multi-signal confirmation. The cleanest separable signals are actually MVRV-or-NUPL (pick one) versus a price-based indicator (Pi Cycle or BTC/Gold) versus something completely external (the equity-market regime, the dollar index, a liquidity proxy).
What This Site Builds On
We focus on the BTC/Gold ratio because it's the cleanest expression of the question we think actually matters: is Bitcoin in the cheap or expensive end of its cycle relative to a comparable store of value? Gold isn't a perfect comparator — its returns are bounded by physical scarcity, while Bitcoin's are bounded by demand for digital scarcity — but it's the closest single asset in economic role, and unlike on-chain measures, it incorporates information from outside Bitcoin itself.
The full methodology, including the walk-forward validation we use to keep the indicator out-of-sample honest, is on the methodology page. If you want to see where the ratio sits today, the current signal page has it.
Get notified when the regime flips
Free email alerts when the BTC/Gold ratio crosses a regime threshold or hits a level you set. No password, takes 10 seconds.