The most expensive word in crypto is 'probably.' This week a trader operating under the handle Killa did something thousands of accounts do every day and one thing almost none of them do: he stated a directional bias โ Bitcoin long, no new low โ and simultaneously refused to defend it with a stop, a size, or a strike. The second part is the story. His plan, as relayed by the outlet that carried it, is conditional and deferred: hold the long, wait for the US midterm election to approach, read the tape, and only then decide whether to layer on a hedge. No entry. No invalidation level. No disclosure of whether the 'long' is spot, a perpetual, or a levered position. A view without a stop is not a thesis; it is a mood with a ticker attached. That is the entire hard dataset. Everything else โ the confidence, the 'known trader' framing, the implication of foresight โ is narrative. And narrative, as any forensic auditor of this market eventually learns, is the most heavily front-run instrument in the book.
I have spent enough time inside transaction pools to know the difference between a signal and a sentiment. In August 2017, when gas hit triple digits and the mempool turned into a parking lot, the signal was never the headline 'Ethereum is congested.' The signal was which bots were paying up and which were withdrawing. The crowd read the fee; I read the order flow. Killa's statement is the fee. The order flow is missing. That gap โ between the claim and the evidence that would validate it โ is where this entire analysis lives.
Bitcoin does not trade like a technology. It trades like a macro instrument wearing a technology costume. By the time spot ETFs cleared in January 2024, that costume had mostly come off. IBIT and FBTC converted the asset into a line item inside model portfolios, and the marginal buyer shifted from a self-custodying believer to an allocation committee operating on a quarterly rebalancing calendar. That shift is the reason a midterm election now matters to a Bitcoin trader at all.
Congress does not set the Federal Reserve's policy rate. But it controls the fiscal impulse, the debt-ceiling theater, and โ critically โ the legislative calendar that determines whether market-structure bills and stablecoin frameworks advance or die in committee. When the composition of one chamber is uncertain, the discount rate applied to every regulated crypto cash flow becomes uncertain with it. The election is not a Bitcoin event; it is a discount-rate event that Bitcoin imports. That distinction matters, because it tells you what kind of volatility to expect. It is not protocol volatility. It is policy volatility, and policy volatility trades through the risk-appetite channel, not the developer channel.
The historical tape on midterms is thin, but it is not empty, and it is not flattering to the bulls. The 2018 midterm arrived in the teeth of the post-ICO winter; Bitcoin was already grinding lower into year-end, and the election did nothing to interrupt the downtrend. The 2022 midterm landed days after the FTX collapse โ a self-inflicted structural shock that had nothing to do with the ballot box but everything to do with the tape. In both cases, the pattern that stuck in trader memory is simple: post-midterm, Bitcoin was weak. That memory is now a folk model, and folk models in markets are dangerous precisely because they are traded. When enough participants believe a seasonal pattern, the pattern migrates from the calendar to the order book, and it can arrive early, arrive inverted, or arrive as a liquidity vacuum rather than a directional move.
The macro backdrop amplifies everything above. Bitcoin's correlation to the dollar index and to real yields is not constant, but it is directional: when real yields fall and the dollar softens, risk appetite returns and Bitcoin catches a bid; when real yields rise and the dollar firms, the bid retreats. A midterm election is a macro event precisely because it can shift the expected path of fiscal policy, and fiscal policy feeds back into the yield curve. A divided Congress means gridlock, which means less new spending, which is disinflationary at the margin and, paradoxically, sometimes supportive of long-duration assets. A unified Congress means a larger fiscal impulse, higher issuance, and a steeper curve, which pressures the discount rate. The trader who reduces the election to 'risk on or risk off' misses the transmission channel entirely. The election does not move Bitcoin directly; it moves the discount rate, and the discount rate moves everything that trades as a long-duration claim on the future. Bitcoin, whatever its detractors say, trades like a long-duration asset.
So when Killa says he does not believe Bitcoin will make a new low, he is making a claim against that folk model. He is saying the seasonal bearishness is either already discounted or is not the binding constraint this cycle. That is a legitimate and interesting position. What is not legitimate is presenting it without the one thing that would make it testable: the level at which he would admit he is wrong. 'No new low' is a statement about the entire future. It has no expiry, no strike, and no falsification condition. An unfalsifiable forecast is not analysis; it is branding. And in a market where the truth is hidden in the block height โ where every claim either settles on-chain or evaporates โ branding is the cheapest commodity available.
Let me be precise about what 'de-risking' means, because the word gets used loosely and it hides a mechanical process. De-risking is not selling. De-risking is the reduction of gross exposure across a book, and it happens in a specific order. First, leverage comes off the derivatives layer โ funding rates normalize toward zero, open interest contracts, and the basis between perpetuals and spot compresses. Second, spot rotation occurs โ capital moves from high-beta assets toward cash-equivalents and, in crypto, toward stablecoins. Third, the marginal buyer disappears โ not because anyone sold, but because nobody is bidding. That third stage is where the damage actually lives, because a market with no bid does not need sellers to fall; it only needs an absence of buyers.
If Killa's de-risking thesis is correct, the sequence above is the map of what happens after the election. And here is the part that his framing glosses over: the same sequence is what happens if the market merely believes the election will produce de-risking. Reflexivity does not wait for the vote. The market front-runs its own assumptions, which is why the trade you are waiting to place has usually already been placed by someone faster. If enough desks decide to 'wait and see' before the election, the waiting itself becomes the selling, and the event they were hedging against arrives early, at a lower price, without a catalyst. This is not speculation; it is the standard mechanics of anticipation in a leveraged market.
Now layer the ETF structure on top. This is where my own work becomes relevant, because the ETF flow data is the single most misread dataset in the current cycle. In early 2024, after the approval, I pulled the on-chain flow data from IBIT and FBTC and found a persistent discrepancy between exchange inflows and ETF creation-unit activity. The gap suggested that a meaningful share of institutional accumulation was happening off-exchange, through custodians, rather than through visible spot purchases. The conclusion I published โ that the ETF was draining liquid supply rather than generating immediate sell pressure โ ran against the loudest narrative of the moment. It also illustrated a general principle: the visible tape and the settled tape are two different documents, and only one of them is auditable. When you read 'ETF inflows' headlines, you are reading the visible tape. When you read custodian wallet movements and creation-unit redemptions, you are reading the settled tape. The election trade, like the ETF trade, lives in the gap between them.
Here is the heart of it. Killa's position contains a structural contradiction that he either has not noticed or has chosen not to surface. On one hand, he holds a long and asserts no new low โ a directional, risk-on stance. On the other hand, he flags post-election de-risking and reserves the right to build a hedge โ a defensive, risk-off posture. Those two stances can coexist in a portfolio, but they cannot coexist in a thesis without a stated trigger, because the trigger is the entire trade. The value of a conditional strategy is not in its legs; it is in the condition that switches between them. And the condition โ 'wait for the election to approach, then watch the reaction' โ is not a condition at all. It is a promise to decide later. Deciding later is not a strategy. It is the absence of one, dressed in the language of patience.
Let me decompose what is actually missing, because the gaps are the story.
First, the time frame. Midterm elections happen on a fixed calendar, but the 'approach' to an election is not a date; it is a liquidity regime. Positioning builds over weeks and unwinds over days. A trader who says 'I will act when the election approaches' is implicitly saying they will act during the period of maximum positioning congestion, which is precisely when slippage is worst and signals are noisiest. Speed is the only moat in a borderless war, but speed applied at the wrong moment is just faster slippage.
Second, the instrument. A long can be spot, a dated future, a perpetual, or a synthetic. Each carries a completely different de-risking exposure. A spot long has no funding cost and no liquidation level; it can survive a de-risking event indefinitely. A levered perpetual long is exactly the position that gets liquidated during the third stage of de-risking, when the bid disappears. Killa does not tell us which he holds. Without that, 'I am long' is not information. A position is defined by its financing, not by its direction. Two traders can both be 'long Bitcoin' and be on opposite sides of the same liquidation cascade.
Third, the hedge. He reserves the right to 'consider' a short. But a hedge placed after a de-risking event has already begun is not a hedge; it is a momentum short. A hedge is only a hedge if it is in place before the risk it is meant to offset. The word 'consider' is doing enormous work here, and it is the tell. A hedge you have not placed is not protection; it is an intention, and intentions do not settle. If the de-risking arrives as a gap โ and it usually does โ there is no 'considering' window. There is only the fill you get, or the fill you don't.
Now let me do the thing I actually do, which is to trace the mechanism rather than the sentiment. Because the question is not whether Killa is right. The question is whether his stated framework can be right in the way he implies.
Start with the derivatives layer, because that is where de-risking is legible first. Bitcoin's perpetual futures market runs on funding rates. When longs are crowded, funding goes positive and longs pay shorts to hold the position. When the crowd gets nervous, the first observable change is not price โ it is funding. Positive funding decays toward zero, then flips negative, and the flip is the moment leverage is unwinding. Open interest is the second tell: it contracts as positions close, and the contraction is the physical footprint of de-risking. The basis โ the spread between the perpetual and the spot index โ compresses toward parity. None of these require a price move to occur. They are the pre-move, and they are observable in real time by anyone watching the tape.
This matters for Killa's thesis because his de-risking call is, mechanically, a prediction about funding, open interest, and basis. If he is right that post-election de-risking is coming, then the tell will appear in the derivatives layer before it appears in price. And if the tell appears before the election โ which reflexivity all but guarantees โ then his 'wait for the election' plan is a plan to arrive after the signal he is waiting for. The market does not schedule its de-risking around a ballot; it schedules it around the perception of risk, and perception moves first. That is the front-run that catches even experienced desks: not being wrong about the outcome, but being late to the mechanism.
Now the on-chain layer, which is where the settlement happens. If it is not on-chain, it did not happen โ that is not a slogan, it is an accounting rule. De-risking at the institutional level shows up as exchange reserve changes and custodian flows, not as headlines. When capital rotates from Bitcoin to stablecoins, the exchange reserves of BTC rise as coins move in for sale, and the stablecoin supply composition shifts. When capital rotates out of crypto entirely, stablecoin supply contracts as redemptions flow back to fiat rails. These are the fingerprints. They are slower than the derivatives tape but more honest, because a custodian wallet cannot spin a narrative. It only updates. The ledger never sleeps, only updates, and it does not care what Killa thinks.
Understand the ETF arbitrage mechanism, because it is the machine that connects the two tapes. Authorized participants create shares by delivering Bitcoin to the custodian and receive ETF units, which they sell into the market when the ETF trades at a premium, or redeem units for Bitcoin when it trades at a discount. Creations drain liquid supply from the spot market; redemptions return supply to it. This means the ETF is not a passive holder of Bitcoin; it is a bidirectional pump that responds to the spread between the fund's price and its net asset value. When you see 'ETF inflows' headlines, you are seeing the net result of this arbitrage, not the intent of any investor. A day of inflows can reflect a single authorized participant closing a premium, not a wave of institutional conviction. The ETF flow number is a settlement artifact, not a sentiment survey, and treating it as a survey is how retail gets the cycle wrong. In a de-risking event, the tell is not the headline flow; it is whether creations stall and whether the premium collapses toward zero, which is the machine running out of demand.
Stablecoin supply is the de-risking thermometer, and almost nobody watches it correctly. When capital rotates out of Bitcoin but stays in crypto, stablecoin supply holds steady while BTC exchange reserves rise. When capital leaves crypto entirely, stablecoin supply contracts as redemptions flow to fiat. The distinction is the difference between a rotation and an exit, and the two have opposite implications for Bitcoin's floor. A rotation is a pause; an exit is a repricing. A trader positioning for a post-election de-risking needs to know which one is coming, and stablecoin supply composition is the instrument that tells him. Watch the stablecoin float: if it holds while BTC reserves rise, you are in a rotation, and rotations mean-revert; if it shrinks, you are in an exit, and exits do not.
Here is the genuinely new insight, and it is the one the discourse is missing. The de-risking trade and the 'no new low' call are not in tension because of market direction. They are in tension because they operate on different time horizons, and Killa has collapsed them into one sentence. 'No new low' is a structural claim โ it is a statement about the floor of the cycle. 'Post-election de-risking' is a tactical claim โ it is a statement about a weeks-long window. A market can make a higher low on a structural basis while still suffering a violent tactical drawdown. Both can be true. And when both are true, the trader who holds an unhedged levered long through the tactical window gets liquidated at the exact moment the structural thesis is most correct. Being right about the destination and wrong about the route is still a liquidation. The two claims do not reconcile; they coexist as a trap for anyone who reads the structural call as a license to ignore the tactical risk.
Let me push further, because this is where most readers stop and where the real edge begins. The reason the midterm folk model exists is that elections resolve uncertainty, and markets tend to sell the uncertainty and buy the resolution. That is the 'sell the rumor, buy the news' structure. But it only works when the uncertainty is genuinely unresolved and genuinely priced. In the current cycle, the question is whether the market has already paid for the midterm risk or whether it is carrying it for free. If positioning is already defensive โ if funding is already neutral, if open interest is already light, if exchange reserves already reflect rotation โ then the midterm 'de-risking' is pre-paid, and the actual event may be a relief rally, the 'sell the rumor, buy the news' outcome Killa's folk model would miss. Conversely, if positioning is extended โ if funding is rich, if open interest is heavy, if the crowd is levered long into the vote โ then the de-risking has room to run and Killa's hedge instinct is correct but his timing is late.
Which scenario we are in is knowable. It is not a matter of opinion. Funding, open interest, basis, exchange reserves, and ETF creation-unit activity are all public or semi-public data. A trader with a real framework would reference them. Killa does not. That absence is not a small omission. When a market call omits the data that would confirm or deny it, the omission is the signal. It tells you the call is built on conviction, chart intuition, or a position that benefits from the narrative, rather than on the observable state of the market.
Now I want to be fair to the man, because the skeptical frame can become lazy. There is a version of Killa's position that is entirely defensible, and it goes like this. The midterm seasonal weakness is a well-known pattern. Well-known patterns get arbitraged. Therefore the weakness is likely front-run and therefore already in the price. Therefore the risk-reward of holding a long into the event is better than the folk model implies, and the correct move is to stay long, keep powder dry, and only hedge if the tape confirms the folk model is reasserting itself. Under that reading, 'no new low' is not arrogance; it is a bet that the crowd's memory is the crowd's mistake. That is a coherent and sometimes profitable view. I have traded that view myself, in a different form, when I audited the Uniswap V2 factory before launch and bet that the market's understanding of the liquidity-bootstrapping mechanism was wrong. The difference is that I had the code. I could point to the constant-product formula and show why the prevailing narrative was incomplete. Conviction without a verifiable source is a coin flip; conviction with one is a position. Killa has the conviction. The question is whether he has the source, and the relayed material gives us no evidence that he does.
There is a second defensible reading, and it is more interesting. It is possible that Killa is not making a price call at all, but a risk-management call. Under this reading, 'no new low' is shorthand for 'I am not willing to short into a possible bottom,' and 'wait for the election' is shorthand for 'I will not add leverage until the policy uncertainty clears.' That is a statement about what he will not do, not about what price will do. And statements about what one will not do are, in a leveraged market, often more valuable than statements about direction, because most blowups come from doing the wrong thing at the wrong time rather than from being directionally wrong. If that is the real content of the message, then it is sound advice wearing a misleading costume. But even then, it is not actionable for anyone else, because it is calibrated to his book, his size, and his pain tolerance โ none of which are disclosed. Advice calibrated to an undisclosed book is not advice; it is autobiography.
Which brings me to the deepest structural point, and the one I want the reader to carry out of this piece. In the current market structure, the dominant flow is passive. ETFs, index products, and allocation mandates buy on schedules, not on signals. That means the marginal buyer is increasingly price-insensitive in the short run and price-sensitive in the long run. De-risking, in that structure, does not look like selling. It looks like a pause in buying. The bid thins, the order book hollows out, and price drifts or gaps lower on relatively light volume. This is why de-risking events are so treacherous for leveraged longs: there is no visible seller to fight, only an absence of buyers to survive. A trader who is mentally prepared for 'the market will sell' is not prepared for 'the market will stop bidding,' and the two require completely different positioning. You cannot short an absence, and you cannot hedge a pause โ you can only reduce your need to be bailed out by a bid that may not arrive.
If that is the true shape of the post-election risk, then Killa's plan โ hold long, wait, maybe hedge โ is structurally mismatched to the threat. A hedge short does nothing against a slow bid-less grind except cost carry. The correct defense against a bid-less market is lower gross exposure and more cash, not a short overlay. His framework appears to be built for a directional de-risking โ a sell-off โ when the more likely mechanism is a liquidity de-risking โ a fade. That is the kind of mismatch that does not show up in a backtest of price, because price is the last thing to move. It shows up in realized slippage, in funding costs, and in the quiet erosion of a position that never got the exit it expected.
Let me close the core analysis with the verifiability problem, because it is the thread running through everything. Killa's claim has four components: a direction (long), a prediction (no new low), a risk assessment (de-risking likely), and a plan (wait, then maybe hedge). Of these, only the direction is stated as fact. The prediction is unfalsifiable as stated. The risk assessment is plausible but unsourced. The plan is conditional on an undefined trigger. A framework in which only the comfortable part is verifiable and every risk is hedged with a 'maybe' is not a framework; it is a comfort object. And comfort objects are precisely what a sideways market produces, because sideways markets starve traders of signal and feed them of narrative. When price does not move, stories move in its place.
Here is the angle almost nobody is running. The interesting question is not whether Killa is right about Bitcoin. It is why this particular statement โ a directional call with no stop, no size, and no evidence โ is circulating as news at all. The answer is that the crypto media complex has a structural appetite for exactly this kind of content, and understanding that appetite is more valuable than any price forecast.
Consider what a piece of 'trader says X' content actually is. It is a sentiment reading dressed as analysis. It has no hard event behind it, no protocol change, no on-chain data, no regulatory filing. Its informational content is a single data point: one participant's current lean. And a single participant's lean, absent position size and disclosure, is close to informationally empty. Yet it circulates widely, because it is cheap to produce, easy to headline, and emotionally satisfying to an audience that wants permission to hold. A sentiment quote is not a signal; it is a mirror, and readers mistake their own reflection for a forecast. The mirror is the product. The forecast is the packaging.
Now the disclosure problem, which is the sharpest edge here. Killa is described as holding a long and believing no new low. He is not described as disclosing the size of that long, when he entered, at what price, or with what leverage. In any other market โ equities, rates, commodities โ a public directional call from a named participant without position disclosure would be treated as suspect, because the incentive to talk one's book is obvious and old. In crypto, the disclosure norm is essentially absent, and the result is a permanent conflict-of-interest blind spot. A trader who is long and publicly says 'no new low' has a direct financial interest in other people believing him, because their buying supports his position. That does not make him a fraud. It makes him a participant with a book, which is exactly why the claim needs to be discounted, not amplified. The louder the conviction and the quieter the disclosure, the more the conviction should be read as positioning rather than prophecy. I learned this the hard way in the NFT cycle, when I audited the Bored Ape minting contract and found that the 'full ownership' story the community was repeating did not match the actual IP transfer terms. The market had priced a narrative that the legal text did not support. The gap between narrative and text is where retail gets hurt, and it is always widest when a confident voice is doing the talking.
The amplification loop is the second contrarian point, and it is mechanical. A KOL statement is picked up by outlets, which syndicates it, which pushes it into feeds, which reaches traders who treat the coverage as corroboration. Now a single opinion appears to be a consensus, because it has been multiplied across channels without any new underlying fact being added. This is not journalism; it is reverb. And reverb has a market effect: it can create a short-term self-fulfilling bid, as readers act on what they believe is broad agreement. The result is a temporary, fragile support that has no fundamental basis and that evaporates the moment the next headline arrives. In a market with no new information, the loudest voice becomes the marginal buyer โ until it becomes the marginal seller. The trader who understands this does not trade the opinion; he trades the reflexive bump the opinion produces, and he exits before the reverb fades.
Now the hardest question, and the one that flips the whole frame. What if the midterm de-risking is already priced? What if the crowd has read the same historical tape, drawn the same conclusion, and positioned defensively in advance? If so, then the de-risking has already happened in the form of light positioning and neutral funding, and the actual event may deliver the opposite of the folk model โ a relief move as the uncertainty clears and sidelined capital re-enters. This is the 'buy the news' scenario, and it is exactly the scenario that a defensively positioned market produces. The folk model says midterms are weak. But folk models that everyone shares do not stay profitable; they get arbitraged into the price and then inverted at the event. If that is the current setup, then Killa's instinct to hold long is correct, but for the wrong reason, and his reserve to hedge is the more dangerous part of his plan, because a hedge placed into a relief rally is a hedge placed against the actual move. The most crowded trade in a sideways market is the one everyone agrees will happen after the catalyst.
There is one more blind spot, and it is the one Killa's framework treats as background noise: regulation. The midterm election matters to crypto not because of the vote itself but because of what it does to the legislative pipeline. Market-structure legislation, stablecoin frameworks, and the jurisdictional turf war between the SEC and the CFTC all advance or stall based on committee composition. A trader who says 'I will wait for the election and then decide' is implicitly saying the regulatory outcome is unknowable until the votes are counted โ which is true โ but he is not pricing the asymmetry. If the outcome is friendly, it is a structural catalyst that lifts the entire asset class and makes the de-risking thesis wrong. If it is hostile, it is a structural headwind that makes the de-risking thesis insufficiently bearish. Either way, the regulatory tail is fat on both sides, and a stopless long into a fat-tailed binary is a position defined by the tail, not by the thesis. A directional call into an unhedged binary event is not conviction; it is leverage on a coin flip. And the disclosure problem returns here too: the projects that talk loudest about decentralization are the ones whose foundation wallets are most traceable, and the same gap between narrative and structure applies to the election trade. The story is always cleaner than the cap table, and the election story is no exception.
There is a fourth angle, and it is the one that keeps me skeptical of the whole genre: the technical-fundamental void. This entire discussion contains zero protocol-level information. No discussion of Bitcoin's hashrate, no miner economics, no halving-supply dynamics, no Lightning capacity, no Taproot adoption, no L2 development, no ETF creation mechanics. It is a pure price-and-policy conversation. That is not necessarily wrong โ Bitcoin at this point in its life genuinely does trade more on macro than on code โ but it means the 'no new low' claim rests entirely on chart intuition and seasonal memory, with no structural anchor. A claim with no structural anchor is a claim that cannot survive a structural shock. If a Mt. Gox distribution, a sovereign sale, or a custody failure hits during the tactical window, 'no new low' evaporates instantly, because nothing in the thesis was load-bearing against that class of event. A forecast that survives only if nothing unexpected happens is not a forecast; it is a hope with a chart attached. The absence of any technical or on-chain evidence is, in itself, the most important technical finding in the entire dataset.
And here is the meta-point that ties it together, the one I would put in bold if I could put the whole article in bold. The reason this content exists โ the reason a stopless directional call is circulating as news โ is that the market is sideways. Sideways markets are information deserts. Price does not trend, so trend-following fails. Volatility compresses, so options sellers win and buyers bleed. The only thing that moves is narrative, and narrative is manufactured by voices. So the market fills its information vacuum with KOL content, and the content becomes a substitute for signal. This is the trap of the current regime, and it is the trap Killa's message walks into. When price stops talking, everyone else starts, and the loudest substitute for signal is usually the most expensive one. The trader who recognizes this does not consume more commentary; he consumes less, and he watches the data that does not lie โ funding, open interest, reserves, and the settled tape.
Let me also correct a subtle but important error in the folk model itself, because precision here is edge. The historical claim is 'Bitcoin was weak after midterms.' But 2018's weakness was the tail of the ICO bust, and 2022's weakness was FTX. In neither case can the election be cleanly identified as the cause. What the data actually shows is that midterms coincided with weakness that had other, larger causes. Correlation with a catalyst is not causation by a catalyst, and a pattern built on two contaminated observations is a superstition, not a statistic. This matters because it means the folk model is even weaker than it appears, which cuts in Killa's favor โ but it also means his 'no new low' claim, which is built on the same thin tape read the other way, is equally unsupported. Both the bull and the bear case here are running on folklore. Neither is running on evidence.
So watch the right things, because the debate about Killa is a distraction from the debate that matters. In the weeks ahead, the tape will tell you which regime you are in, and it will tell you through data, not through voices. Watch funding: if it stays neutral or flips negative into the election, the de-risking is pre-paid and the folk model is likely to invert. Watch open interest: if it is light, there is no fuel for a cascade and no need for a hedge. Watch exchange reserves and ETF creation-unit activity: if coins are not moving in for sale and creations are not stalling, the bid is intact and the 'no new low' call has a floor under it. And watch the settled tape above all, because the settled tape is the only document that cannot be spun. If it is not on-chain, it did not happen โ and that rule applies equally to a trader's conviction and to a market's fear. The next move will not announce itself in a quote. It will announce itself in the block height, in the funding print, and in the silence of a bid that either holds or does not. The question is not whether you believe Killa. It is whether you are watching the data that would make belief unnecessary.
