Every bull market opens with the same sentence: "This time is different." Every bear market closes with the same one: "Crypto is dead." Seven years in, I have lived through two complete cycles and heard each of those sentences repeated dozens of times. Neither one has been accurate yet. This piece covers five signals I actually use to read the cycle, a handful of contrarian indicators that have outperformed those signals, and the dumb, boring method I rely on most of the time: no chart-reading, no top-and-bottom calls, no directional bets.

The honest opening: nobody calls the exact top or bottom

Before any signals, the uncomfortable truth: nobody catches the exact top or exact bottom. The people on X who claim to have done it almost always have curated tweet history. The June 2018 catchphrase among the crypto people I knew was the question "you still around?" — which is itself a sentiment indicator that beat every chart-based signal I had at the time. Tony Greer, Raoul Pal, Arthur Hayes, the people who do this for a living and have access to data desks you don't — they have all called tops and bottoms wrong publicly. So have I. Stop trying to catch the precise candle. You almost certainly won't, and even if you do, you won't have the conviction to size the trade.

The thing I actually do is different: I try to identify roughly which phase of the cycle we are in — early, mid, late, transition, deep bear, or accumulation. Not precise to a week. Coarse to a quarter or two. Coarse is enough. As long as you avoid leverage, being directionally right at the quarter-level produces results that most active traders never reach.

November 2021. New York. BTC had just touched $69,000. I caught a cab from a meeting in Midtown. My driver, mid-forties, started telling me his nephew said "this Bored Ape thing is going to make us rich" and asked whether he should buy. I felt my stomach turn but didn't sell that night. I held three more months and trimmed half my BTC stack in the $40K range. In hindsight, the day that driver asked me the question was the top. I never needed to know whether the precise top was $69K or $73K — when the cab driver asks, it's time to leave.

Signal 1 · On-chain (MVRV / Realized Cap / SOPR)

On-chain data is the signal class I trust most, because it comes from actual transactions — not opinions, not sentiment, not survey responses. The three metrics I check live on Glassnode, CryptoQuant and DeFiLlama, all of which have free tiers.

MVRV (Market Value to Realized Value)

MVRV = market cap divided by realized cap — roughly, "how much profit is the holder base sitting on right now?" It is the single best single-number indicator of cycle phase that I know.

My personal rule of thumb: when MVRV hits 3.0 I trim 10% of the BTC stack. At 3.5 I trim another 20%. When MVRV drops under 1 I resume DCA. The thresholds are not magical; they are a discipline. Pre-committing to the thresholds prevents the "but what if it goes higher" emotional override that turns trim plans into hold-and-pray plans.

Realized Cap

Realized Cap values every BTC at the price it last moved on-chain. It is, effectively, the aggregate cost basis of the holder base. It's a slow indicator — boring to watch in real time — but the velocity of Realized Cap growth tells you whether new money is entering. Accelerating realized cap = new capital flowing in = bull-market dynamics. Flat realized cap = stagnation = range-bound or bear.

SOPR (Spent Output Profit Ratio)

SOPR is the aggregate profit/loss ratio on coins moving on-chain on a given day. SOPR > 1 means people are selling at a profit in aggregate (bull-market behavior). SOPR < 1 means people are realizing losses (capitulation behavior). The cleanest signal SOPR gives is when it crosses back above 1 after sustained sub-1 readings — that crossover has marked early bull transitions in April 2019 and again in January 2023. Worth keeping an eye on.

Signal 2 · Funding rate

Funding rate is the thermometer of the perpetual-futures market. Perp contracts pay an 8-hour settlement between longs and shorts: when longs dominate, longs pay shorts (positive funding); when shorts dominate, the opposite (negative funding).

March 2024, when BTC hit $73K for the first time, funding annualized at 80%. I don't trade, so I didn't short — but I paused my DCA for a month and waited for the pullback to about $62K before resuming.

The tool I use is Coinglass, once a week. If funding annualized stays above 50% for a full week, I know we're in a short-term overheated zone. Is it precise enough to time a top? No. But it's accurate enough to keep you from buying into the most overheated weeks — which is most of the value.

Signal 3 · Exchange BTC balances

A crude but useful indicator: the total BTC sitting on all exchange wallets. The logic is mechanical — coins on exchanges are coins ready to sell; coins withdrawn to self-custody are coins ready to hold.

October 2020 through April 2021 saw exchange balances fall by roughly 30% — a clear bull-market signal. After the January 2024 spot ETF approvals, balances fell hard again (ETFs absorbed everything coming out of exchange wallets).

⚠ This metric has lost some power since 2024

With spot ETFs now in the picture, large amounts of BTC live in BlackRock and Fidelity custody accounts rather than "exchanges" — Glassnode has started splitting ETF balances into a separate series. The way to read this metric today is to look at exchanges + ETF wallets combined, not exchanges alone. The signal is still useful; the data definition just needs to be updated.

Signal 4 · Sentiment (Fear & Greed)

The Crypto Fear & Greed Index compiles a 0-100 composite from volatility, momentum, social-media chatter, Google Trends, market dominance, and survey data. It's well known and easy to find.

By itself the index is not enough — the 2021 April top printed a 95, but the March 2024 high only got to 80, and a few unprecedented all-time-high prints in 2024-2025 happened with sub-70 readings. It's a coarse screen, not a precise timer. My usage: a full week above 85 = lean defensive; a full week below 20 = consider scaling buys. Anything in between, ignore.

Signal 5 · The Fed rate cycle

This is the highest-level macro signal — and after seven years, I think it's the single most important one. Bitcoin is highly correlated with global liquidity. Loose policy (cuts + QE) lifts risk assets; tight policy (hikes + QT) drains them.

Fed actionBTC outcome
March 2020 · emergency cut + unlimited QELaunched the $5K → $69K bull run
March 2022 · hike cycle begins + QTBTC fell from $47K to $15.5K
November 2023 · hike cycle hinted as endingBTC ran from $35K to $73K by March 2024
September 2024 · first rate cutTriggered the next leg of the bull

Operationally: watch the Fed's tone, not the rate decision itself. Markets price the expected path 3-6 months ahead of the actual move. The FOMC press conference on the last Wednesday of each meeting cycle matters more than the dot plot — pay attention to Powell's wording, the questions he ducks, and the changes in language between consecutive meetings.

A practical aside: none of the five signals above tells you when altcoins should be in play. BTC dominance (BTC.D) is the cleanest indicator for that question, with four threshold zones that have repeated across cycles. I built a live BTC dominance tool for exactly this — pair it with the five signals above and the cycle picture sharpens up.

Contrarian indicators (often more useful than the five above)

All five signals above require looking at data. There's a parallel class of indicators that you cannot avoid because they live in everyday life. They have outperformed the data signals at the most important inflection points of the last two cycles.

June 2022. Two months after LUNA died, Celsius froze withdrawals, and 3AC went into administration. I scrolled X and counted: six of the ten crypto influencers I followed had changed their profile pictures — to AI projects, commodities, "local services." Search interest for "bitcoin scam" hit a four-year high. The mood that month was that crypto was finished. I did not sell. I kept DCAing. The reason: I had seen this exact pattern in December 2018. Everybody had said "zero" then too. A year later BTC had tripled.

What contrarian indicators measure is whether the marginal-seller cohort has finished leaving. When the last people who shouldn't have been here are gone, the selling pressure that defined the bear is gone with them. What follows is the floor.

Cycle accounting · four tops and bottoms

Running the signals back across the four cycles I've watched (the first one I learned from after the fact):

2013 · the original rollercoaster

Before my time, but the post-mortem is well-documented. April peak at $260 → $50 crash, November second peak at $1,200 → $200. On-chain data was barely available; the cleanest signal was the first major retail wave (Chinese Baidu search for "bitcoin" hit a record in November 2013).

2017 · the first retail mania

BTC from $1K to $20K in a year — a clean 20x. Top signals:

2021 · the double-top cycle

Different from both prior cycles — this one printed two tops: $65K in April, $69K in November. The first top was MVRV at 3.8 and funding rate at extreme highs. The second was driven by NFT mania, Bored Apes, and "Web3" institutional pitching. Both tops shared the same signatures: market unconditionally optimistic, retail mania penetrating to taxi drivers, institutional voices declaring "this time is different."

2024-2025 · ETF cycle + structural shift

This cycle is the strangest of the four. With US spot ETFs sucking up coins and BlackRock as a marginal buyer, the drawdown profile is compressed (no 77% bear-market collapse like 2022 yet). My current read is that the top of this cycle has not yet printed, on these grounds:

But I won't go all-in on that read either. The most important lesson of seven years is do not trust your own analysis too much. Lean toward what the data favors, but never bet the farm.

§ US/EU reader perspective · the indicators specific to your market

The five signals above are jurisdiction-neutral — they apply equally well in Seoul or São Paulo. But for US and European readers there are additional indicators that have become important since 2024, and a few cultural markers worth tracking. This section is the part that doesn't appear in the Chinese version.

The Fed cycle as the master signal

In retrospect, the cleanest cycle calls of the last decade have all come from the Fed. The 2018 hike cycle and 2022 hike cycle both produced bear markets. The 2020 emergency cuts and 2023-2024 hike-pause both produced bull markets. If you only watched one number, the federal funds rate would have outperformed any on-chain indicator. Mark the FOMC dates on your calendar; read the press conference transcripts; ignore the noise from rates traders trying to predict each 25-bp move and focus on the meta-question of "is the Fed easing or tightening this year."

ETF flows · the new institutional sentiment proxy

BlackRock's IBIT publishes daily inflow / outflow numbers. So do Fidelity's FBTC, Bitwise's BITB, and Ark/21Shares' ARKB. The aggregate daily net flow across the eleven US spot Bitcoin ETFs is the single best proxy for US institutional sentiment that has ever existed for this asset class. Three or four consecutive days of $300M+ inflows on IBIT signals real bid; two weeks of net outflows is genuine distribution. Farside Investors publishes the daily table for free. I check it weekly.

MicroStrategy (MSTR) premium to NAV

Michael Saylor's MicroStrategy (now formally "Strategy") trades at a premium or discount to the actual value of the BTC it holds on its balance sheet. That premium is a real-time retail sentiment indicator: MSTR at a 100% premium to NAV means retail is paying double for leveraged BTC exposure, which historically marks euphoric phases. When MSTR trades near or below NAV, retail has lost interest — which has historically been close to capitulation. Check the premium on strategytracker or by dividing market cap by reported BTC holdings × spot price.

CME futures basis · the institutional position indicator

The premium of the CME Bitcoin futures contract over spot — the "basis" — tells you whether institutional money is positioned long. A basis annualized above 15% indicates real positioning by hedge funds and trading desks. A flat or negative basis (backwardation) indicates institutional caution or active shorting. The basis has been one of the most reliable signals around cycle inflections — backwardation appeared right before the 2022 bottom and during the May 2021 sell-off. CME publishes the data; aggregator dashboards like Coinglass display it in a more readable form.

Coinbase App Store ranking

A folk indicator that has worked surprisingly well: when Coinbase climbs into the top 10 of the US iPhone App Store free chart, US retail FOMO has peaked. This signal flashed clearly in May 2021 (Coinbase reached #1 the day BTC hit $64K) and again in March 2024. AppFigures publishes the daily ranking. A free check that takes thirty seconds.

Google Trends US "bitcoin"

Google Trends data for the United States — searching the keyword "bitcoin" — is one of the cleanest leading indicators for US retail attention. A spike to a multi-year high in US-only Trends precedes top by 4-8 weeks reliably. Conversely, multi-year Trends lows often coincide with cycle bottoms. The query is free and takes seconds to run. Compare US-only Trends to global Trends and the divergence itself becomes a signal — when US Trends spikes but global Trends doesn't, US retail is running ahead of the rest of the world (often late-cycle).

Political rhetoric · a new factor in 2024-2025

The 2024 US election campaign saw both major presidential candidates make crypto-positive statements — a complete reversal from the regulatory hostility of 2021-2023. Trump's Nashville speech in July 2024 floated a "strategic Bitcoin reserve" idea; the Harris campaign issued a more measured crypto-positive policy paper in August. This is a new factor that didn't exist in any prior cycle. Watch the regulatory tone change carefully: SEC chair turnover, CFTC commissioner statements, House Financial Services Committee hearings. Political-rhetoric shifts in 2024-2025 have moved markets the way Fed rhetoric did in 2018-2022.

Stablecoin market cap (USDT + USDC)

Aggregate stablecoin market cap, viewable on DefiLlama or CoinMarketCap, is a near-real-time proxy for capital sitting on the sidelines waiting to be deployed. Rising stablecoin cap during a sideways market often precedes a leg up. Falling stablecoin cap after a strong rally suggests retail has cashed out into fiat — late cycle. The relationship is not perfectly clean but the trend is interpretable at the weekly chart level.

US-specific calendar markers

Two recurring patterns that show up clearly in the US data:

Putting the US-specific stack together

None of the seven US-specific indicators above is a standalone timing tool. They work as a cross-check on the global signals. The rough operating model I use: when three or more US-specific indicators align in the same direction as the on-chain signal stack, the read becomes higher conviction. A March 2024 example: MVRV was nearly 3.5 (top-zone), funding rate had spiked to annualized 80% (overcrowded longs), MSTR was trading at 100%+ premium to NAV (retail euphoria), Coinbase had climbed to #2 on the US App Store, and IBIT was seeing $300M+ daily inflows day after day. Five aligned signals. I cut DCA and waited. The subsequent retracement to $62K vindicated the pause. Single signals are noisy; converging signals are how you stay disciplined.

Why this cycle's structure is different · and what it implies

Three structural shifts from prior cycles, all of which I think bear on how the 2024-2026 phase ends:

The combination of these three shifts is why I don't believe the simple "2x retail mania = peak" model from 2017/2021 applies cleanly to 2024-2026. The retail mania component may simply be lower this cycle because the institutional component absorbed the demand. If that's right, the absence of taxi-driver FOMO at a $108K BTC is not evidence that this isn't the top — it might just be evidence that the cycle composition changed. I am still leaning toward "top not yet in" on the broader data, but with much less conviction than I would have had in a 2017/2021 setup. Reasonable people can disagree on this read.

The old-trader method · DCA + don't watch + no leverage

All of the above is useful. Most of the time I don't use it. The method I default to is duller:

Method 1 · Calendar-driven DCA

On the first of every month, I buy a fixed dollar amount of BTC and ETH. Same number every month, no exceptions. This has run unbroken since 2018 — seven years. Run the math on public monthly prices and the blended cost basis lands well below today's $90K+ spot — this isn't my private ledger, it's a long-run pattern anyone can reproduce from public data: DCA compounds ahead of most active traders. The real value of DCA is not the arithmetic. It is psychological: no timing decisions, no FOMO, no panic. During bears, DCA is actively pleasant — the same money buys more coins. During bulls you accumulate less but your stack is already large by then.

Method 2 · Cut screen time by 90%

I check the price once a day for five seconds. Once a week I look at on-chain and funding rate data for ten minutes. Once a month I do the DCA, which takes five. Total: under an hour per month. More watching produces more action; more action produces more mistakes. I have verified this empirically over seven years. The moments where I "felt the urge" to do something — sell, buy more, hedge — were wrong 80% of the time. The 20% where they were right did not produce enough alpha to offset the 80%.

Method 3 · Zero leverage, ever

Read my leverage piece for the full version. Leverage is the dominant way this market kills people. I have never taken a futures or perp position in seven years. If I'm right about the cycle, no-leverage still captures the bulk of the move. If I'm wrong, no-leverage means I lose opportunity cost — not the account. Asymmetric risk is the killer.

Method 4 · Trim, never sell out

I never fully exit. The most I'll trim at a perceived top is 30-40% of the stack. The remaining 60-70% stays on. The reasoning: if I'm wrong (which is often), the core position is still riding. If I'm right, the trim cash is available to deploy in the next bear. People who fully sell have two endings. One: they buy back lower. Two: they watch the recovery from the sidelines and never re-enter. Trimming is the path between those failure modes.

Closing thoughts (less inspirational than they should be)

After all the signals and contrarian indicators and methods: the most useful single sentence is that you don't need to predict bull and bear precisely. You need to keep buying assets you believe in, sized to what you can afford to lose, over a long enough horizon for the cycle to play out. The signals will help you avoid the worst mistakes, but they will not generate alpha by themselves. Whether you make real money depends on whether the underlying asset class actually grows over the next decade — whether BTC becomes a more broadly held reserve asset, whether ETH becomes a meaningful settlement layer, whether crypto plugs into the real economy in a way that justifies the multi-trillion-dollar market cap. I can't answer those questions for you. But I am betting it does — which is why I have stayed seven years. You'll have to make your own bet, and then once you've made it, the remaining job is simple: watch less, act less, stay alive.

The thing that has kept me in this market is not any signal, any indicator, or any contrarian read. It's a single sentence I keep coming back to: "Don't worry about the things that don't matter; do very little about the things that do."