The Misfiled Bear: Bitcoin, Midterms, and the Four-Year Wall Nobody Blames on Politics

There is a chart making the rounds that stops me cold, not because it frightens me, but because it is lazy. It plots Bitcoin's drawdown after every United States midterm election: minus 72% in 2010, minus 65% in 2014, minus 52% in 2018, minus 27% in 2022. Four elections. Four declines. The caption writes its own conclusion, and the timeline has already signed it. American politics kills the Bitcoin bull. I have traded through two of those candles. I remember the texture of 2018 and 2022 not as political events but as slow, grinding repricings of leverage that had been allowed to grow fat during the expansion phase. Nobody in my order book was checking the composition of the House of Representatives. They were checking margin. And yet here we are again, in the sideways chop of a market that cannot decide whether it is a reserve asset or a risk asset, being handed the same tidy story with a fresh election date stapled to it. The material I am responding to is not a protocol. It is a piece of content, an AI opinion compilation that asked three chatbots, ChatGPT, Perplexity, and Gemini, to forecast Bitcoin's reaction to the coming midterms. All three agreed: short-term pullback, not a trend reversal. A five to ten percent dip, with fifteen percent as the stress case. It reads like a forecast. It is actually a mirror. Before I take this apart, let me state plainly what is being claimed. The article asks three large language models to predict how Bitcoin behaves around the US midterm elections. It folds in a forecast from Ali Martinez, a widely followed on-chain analyst, who notes that Bitcoin fell 72%, 65%, 52%, and 27% following the 2010, 2014, 2018, and 2022 elections respectively. It cites the prediction market Kalshi, which prices Republican control of Congress at under thirty percent. It references a set of Trump-era promises: make America the global crypto capital, establish a strategic Bitcoin reserve, prohibit a central bank digital currency, and deliver a friendlier regulatory framework. And it concludes, through the mouths of three models, that a Democratic win would produce a short, contained pullback rather than a structural break. There is a surface logic here that deserves respect. If the sitting administration is crypto-friendly, a legislative loss for that administration is, on paper, a negative for the asset. That is the whole of the argument. The rest is decoration. But the surface is where the analysis stops, and that is the problem. The piece is a survey of narratives, not a measurement of markets. It contains no funding rates, no open interest, no spot-versus-perpetual basis, no ETF flow decomposition, no realized volatility term structure. It contains, in fact, no primary market data at all. It contains three language models agreeing with each other, which is not corroboration, and one analyst's historical chart, which is not causation. Everything else is vibes dressed as rigor. I want to be precise about what I am not saying. I am not saying elections do not matter. I am not saying Bitcoin is immune to policy. I am saying that the specific causal claim on the table, that midterm elections drive Bitcoin into a bear market, is almost certainly false, and that the data being used to support it is contaminated by a variable that no one in the article mentions. That variable has a name, it runs on a four-year clock, and it has been the dominant driver of Bitcoin's cycle for over a decade. Before I get to it, I want to give the AI-generated forecasts their due, because dismissing them entirely would be its own form of laziness. The models were asked a narrow question and gave a narrow answer. Conditional on a Democratic sweep, conditional on the result being a surprise, conditional on no other macro shock, a short-term pullback is a reasonable base case. The problem is not the answer. The problem is that the question was framed so tightly that the answer could only ever be noise around a narrative. Ask a model what the market will do after an election, and it will tell you what the training data says markets do after elections. That is not a prediction. That is autocomplete with a Bloomberg terminal's aesthetic. Here is the finding that reorganizes everything. Every United States midterm election falls in an even-numbered year. And Bitcoin's most powerful structural cycle, the halving, also runs on a four-year clock. Line them up and the coincidence becomes impossible to ignore. The 2012 halving was followed by the 2013 top and the 2014 bear. The 2016 halving was followed by the 2017 top and the 2018 bear. The 2020 halving was followed by the 2021 top and the 2022 bear. Now look again at the so-called election drawdowns. 2014. 2018. 2022. Three of the four data points in Martinez's chart land exactly on the year after a post-halving peak. They are not election bears. They are halving bears wearing election costumes. This is a textbook confounding variable, and it is the single most important thing the original article missed. When two variables move together because they share a hidden common cause, you cannot attribute the effect to either one by observation alone. The midterm calendar and the halving calendar are both four-year cycles. They overlap by construction. To claim the election caused the drawdown, you would first have to subtract the halving effect, and you cannot do that with a sample of four overlapping observations. The correlation is real. The causation is an artifact. Let me push further, because the pattern is cleaner than the article lets on. Bitcoin's cycle has a recognizable anatomy: a halving cuts new supply issuance in half, a bull market builds over roughly twelve to eighteen months as the supply shock propagates through a market that is still thin, a euphoric top forms, and then a bear market follows as leverage unwinds and the marginal buyer disappears. The bear phase typically lands in the year after the peak. For the 2024 halving, the arithmetic points to a potential top in 2025 and a bear phase in 2026. And 2026 is a midterm election year. The overlap is not a coincidence that happened once. It is a structural feature of how these two cycles are timed, and it is about to happen again. This is why I find the article's conclusion so dangerous. It tells readers that post-election declines are excellent buying opportunities, which may turn out to be directionally correct for 2026. But it arrives at that conclusion through the wrong door. If Bitcoin does fall in 2026, the article will be able to say, see, the election, and the market will nod, and the real driver, the maturation of the post-halving distribution phase, will be buried under a political story for another four years. Being right for the wrong reason is not a strategy. It is a trap that pays out once and bankrupts you the next time the pattern breaks. Now let me address the sample itself, because even setting aside the confounding, the data is weak. Four observations is not a dataset. It is an anecdote with error bars. In statistics, a sample of four tells you almost nothing about the underlying process, especially when the four observations are not independent, which here they are not, because they are all drawn from the same cyclical structure. Anyone who has run a backtest knows the rule: the fewer the observations, the more the result is dominated by the specific period you happened to sample. Four overlapping points from one repeating cycle is the weakest possible foundation for a forward-looking claim. And the weakest of the four is 2010. In 2010, Bitcoin was a curiosity traded on a forum, with liquidity so thin that a single large order could move the price by double digits. Including 2010 in a post-election drawdown study is like including the first week of a startup's existence in a study of corporate governance. The asset had no market structure, no derivatives, no institutional participation, no macro sensitivity. The minus 72% figure is not a data point about elections. It is a data point about a market that barely existed. Strip it out and your sample of four becomes three, and your correlation weakens further. Let me now turn to the variable the article did discuss but misread: pricing. The article notes that Kalshi prices Republican control of Congress at under thirty percent. It treats this as context. I treat it as the single most actionable number in the entire piece. If the market already assigns a high probability to a Democratic outcome, then a Democratic outcome is not news. It is the baseline. And when an event resolves in line with the baseline, the marginal price impact is small, because the information was already in the price. The models' forecast of a five to ten percent drop assumes the market is surprised by something it is already expecting. That is not how efficient pricing works at the margin. The real risk, the one nobody is modeling, is the opposite tail. If Republicans hold Congress against the prediction market's odds, the crypto-friendly outcome is confirmed, and the market could sell off anyway, because the good news was already priced into the Trump trade that has been running for nearly two years. This is the classic expectation gap. The article and its three models spend all their energy on the Democratic scenario and almost none on the Republican one, even though the Republican scenario is the one with the larger surprise component. They have the asymmetry backwards. Hold the line when the world screams to sell, and the world is currently screaming about an election result that has not happened, based on a chart that does not say what it claims to say. That is not a reason to panic. That is a reason to look at the clock instead of the calendar. I want to dwell on the method, because I have spent the last year integrating machine learning into my own workflow, and I know exactly where these tools are strong and where they lie to you with perfect confidence. When I built AI-driven models into my trading in 2026, I learned a hard rule: a model is only as good as its inputs, and a language model predicting markets has almost no inputs at all. It cannot read the live order book. It cannot see the funding rate flip from positive to negative in real time. It cannot watch spot ETF flows print in the morning and infer the marginal buyer. It cannot detect a whale moving coins to an exchange. What it can do is reproduce the dominant narrative in its training data with grammatical elegance. When I ask a language model about markets, I am asking it what the internet already believes. That is occasionally useful as a sentiment thermometer. It is worthless as a forecast. The three models agreed with each other. The article presents this as a strength, a kind of triangulation. It is the opposite. Three models trained on overlapping corpora will converge on overlapping answers, not because they have independently verified the truth, but because they share the same priors. Agreement among correlated sources is not evidence. It is an echo. If I asked three traders who all read the same newsletter, I would get one opinion, not three, no matter how many traders I asked. This matters because of a specific cognitive failure that fluent output induces. Language models produce authoritative prose. The syntax is clean, the hedges are calibrated, the tone is measured. The reader's brain, trained over a lifetime to treat confident, well-structured writing as a signal of expertise, cannot help but assign credibility. The models return a prediction that is simultaneously vague and unfalsifiable, short-term pullback, not a trend reversal, which is the safest possible thing to say, because almost any outcome except a straight line up or a total collapse can be retrofitted to fit it. That is not a forecast. That is a statement designed to be right no matter what happens. I hold no ill will toward the models. They did what they are built to do. My concern is with the framing that turns their output into a market signal. The article is not a research product. It is a content product, and three AIs analyze is a marketing frame, not a methodology. It signals novelty and impartiality while delivering neither. When I read it, I do not see analysis. I see a template. Here is a distinction the article blurs, and it is the kind of distinction that separates a trader from a commentator. A United States midterm election changes control of Congress. It does not change the occupant of the White House. The president remains in office until 2029 regardless of what happens in the midterms. This is not a minor procedural detail. It is the difference between two entirely separate channels of policy influence. The crypto-friendly posture attributed to the current administration flows largely through the executive branch: personnel at the SEC and CFTC, executive orders, enforcement priorities, and the general tone set from the top. None of that reverses on a midterm result. The legislative branch controls the passage of new statutes, things like stablecoin frameworks and market structure bills, and that is the channel a midterm can actually redirect. So the article's implicit claim, that a Democratic Congress would meaningfully roll back crypto-friendly policy, conflates two channels with very different sensitivity to the election. When I worked with a legal team in London in 2025 to draft internal compliance guidelines for a mid-sized crypto fund, I learned to think in terms of regulatory structure rather than regulatory mood. The structure of who has authority over what determines what can actually change and when. A shift in legislative control affects the pace of new laws. It does not rewrite the rules already in force, and it does not reverse executive appointments. To model a midterm as a binary switch on crypto policy is to misunderstand the machinery. The article does not misunderstand it because it is complex. It misunderstands it because the article did not look. There is a subtler problem with the whole framework, and it is about time. The Trump trade in crypto has been running for nearly two years. Narratives have life cycles: a promise phase, a build phase, a fulfillment phase, and a fatigue phase. A trade that has been priced for two years is no longer a fresh catalyst. It is a crowded position. And crowded positions are dangerous precisely when their narrative catalyst finally arrives, because that is when the holders who bought the promise start looking for the exit. The article itself admits, almost in passing, that other decisions have aggravated geopolitical tensions and weighed on markets. That admission is the tell. It signals that the clean story, crypto-friendly administration equals higher prices, is already fraying at the edges. The market is transitioning from pricing the promise to grading the delivery. And delivery is where promises go to die. Consider the strategic Bitcoin reserve. On paper, a sovereign buyer is the most bullish structural demand shock Bitcoin could receive. It is the kind of thing that would justify a repricing far beyond anything an election could produce. But the article treats it as an established fact rather than an aspiration with a long, uncertain path through fiscal and legal machinery. Sovereign purchases of a volatile asset require appropriations, legal review, and political will that survives contact with a budget process. The gap between a promise and a purchase is where narrative meets reality, and reality usually wins. If the reserve materializes at a fraction of the promised scale, the disappointment will be felt not as a political event but as a repricing of an overpromised narrative. The same logic applies to the central bank digital currency prohibition. The absence of a CBDC is theoretically supportive of Bitcoin and stablecoin demand, but the article never argues the mechanism. It asserts the conclusion and moves on. That is not analysis. That is decoration. I want to widen the lens here, because the article makes a second mistake that is easy to miss and expensive to ignore. It treats the United States as the only jurisdiction that matters. It does not mention the European Union's MiCA framework, or Hong Kong, or Singapore, or the quiet regulatory maturation happening across Asia and the Gulf. This is not a small omission for a piece about the future of a global asset. The MiCA story is instructive. Europe's framework gives the market apparent clarity, which sounds like a gift. But clarity is not the same as permissiveness. The stablecoin reserve requirements and the compliance costs imposed on crypto asset service providers are structured in a way that favors large, well-capitalized players and squeezes small projects out of existence. When I helped draft those London guidelines, the exercise was not about understanding the rules. It was about calculating whether compliance was economically survivable. For a mid-sized fund, it is. For a small protocol, it is often not. Regulation, done this way, is not a level playing field. It is a consolidation mechanism wearing a safety jacket. The article's singular focus on American electoral politics misses this entirely. If a Democratic Congress slowed US legislation, capital would not vanish. It would reallocate toward jurisdictions with clearer, if costlier, rules. The crypto market is not a single-country market with a single political sensitivity. It is a global market with a mosaic of regulatory regimes, and the marginal dollar flows to wherever the structure is most legible. To model Bitcoin's future as a function of one election in one country is to shrink a global asset into a local story. That shrinkage is convenient for a headline. It is useless for a trade. There is a deeper tension running under all of this, and it is the one I find most worth watching. Bitcoin is being treated, simultaneously, as a safe-haven reserve asset and as a high-beta risk asset that drops five to fifteen percent on a single country's legislative election. Those two identities are incompatible. A true reserve asset does not flinch at a midterm. If Bitcoin is digital gold, then a United States election should be a rounding error, not a five-percent move. The fact that the entire article, and the models within it, assume Bitcoin will react sharply to the election is itself an admission that, in practice, the market still trades Bitcoin as a risk asset. This identity confusion is not a bug in the article. It is a feature of the asset's current stage. Bitcoin is in the middle of a transition, from a speculative instrument priced by retail sentiment to an institutional asset priced by macro liquidity and ETF flows. And the ETF era has changed the asset in a way that the original article never confronts. When BlackRock and its peers wrapped Bitcoin into a regulated, tradable, custodied product, they did not just open a door. They changed what Bitcoin is on the margin. It is now, functionally, a Wall Street risk asset that trades alongside equities, reacts to rate expectations, and is bought and sold by allocators who treat it as a volatility sleeve in a portfolio. Satoshi's peer-to-peer electronic cash is a footnote. The marginal price is set by institutional flow, not by peer-to-peer settlement. That transformation is why the sideways chop we are living through is so instructive. During a transition, the asset carries both sets of sensitivities, and the market oscillates between pricing it one way and pricing it the other. That oscillation is the range. And in a range, the job is not to predict the direction. The job is to position for the range and wait for the structure to resolve. Hold the line when the world screams to sell. In chop, the scream is constant. That is the point. The noise does not tell you which way the structure will break. It only tells you that most people are guessing. Let me say something about order flow, because that is where the real answer lives, and it is where the article is silent. When I look at a market like this, I do not start with the calendar. I start with who is buying and who is selling, and at what price they are willing to transact. The relevant question is not what will the election do to Bitcoin. It is at what price is the marginal institutional buyer willing to absorb supply, and how does that price interact with the levels where leverage is stacked. In a post-halving distribution phase, the pattern is predictable in shape if not in timing: leverage builds during the expansion, gets flushed during the contraction, and the flush produces the lows that become the next cycle's accumulation zone. The election, if it matters at all, matters as a timing trigger, a headline that gives an already-fragile market permission to flush. The headline is the spark. The fuel was stacked months earlier by everyone who levered into the promise. This is the difference between retail and smart money that the article never draws. Retail reads the headline and reacts to it. Smart money positions ahead of the headline, anticipates the reaction, and uses it as liquidity. If the market is braced for a Democratic win and a five-to-ten percent dip, then the dip, if it comes, will be bought, because the braced position is already the trade. And if the market is not braced for a Republican surprise, that is where the volatility will concentrate, in the direction nobody prepared for. The article's models are watching the wrong door. I will add one more layer, drawn from my own trading. During the 2024 spot ETF approval, I executed fifteen trades and netted one hundred twenty thousand dollars on a two-hundred-thousand-dollar base. I did not do it by predicting the news. I did it by waiting for the technical setup to align with institutional volume, and then acting on the rules I had battle-tested. The news was not the signal. The flow was the signal. The news was just the noise that gave the flow somewhere to go. Every disciplined trade I have ever made followed that pattern: ignore the story, read the flow, respect the level. And the flow, right now, is telling a story about cycle maturity that no election headline can overwrite. Here is the counterintuitive angle, and it cuts against almost everyone in this conversation. The article, the three models, and the analyst all assume the election is a source of risk to be feared or a catalyst to be traded. I think the more likely truth is that the election is largely irrelevant to Bitcoin's trajectory over any horizon longer than a few weeks, and that treating it as central is a category error. The reason is structural. Bitcoin's price over the coming year will be determined by the interaction of three forces: the maturation of the post-halving cycle, the direction of global liquidity and interest rates, and the pace of institutional adoption through ETF flows. The midterm election touches none of these directly. It touches the legislative channel, which is the slowest-moving and least Bitcoin-specific of the policy channels. To elevate it to the primary driver of a market that trades twenty-four hours a day across every jurisdiction on earth is to mistake a headline for a mechanism. And yet the market will behave as if the election matters, because markets are reflexive and narratives are self-fulfilling at short horizons. That is the opportunity. When a crowded narrative points in one direction, the asymmetric payoff is usually in the other. If everyone is braced for an election-driven dip, the dip is either bought or it never comes. The trade is not to predict the election. The trade is to recognize that the crowd's forecast has already been absorbed into positioning, and to prepare for the outcome the crowd has not priced. The most important blind spot, and I will say it plainly, is the halving cycle. Every source in this article, the three models, the analyst, the editorial voice, ignores it. Not one of them mentions that 2026 sits in the window where the post-halving distribution phase tends to mature. This is not a small omission. It is the omission. It is like analyzing a company's earnings without looking at its revenue. The single most powerful structural variable in Bitcoin's price history is absent from a piece that purports to forecast Bitcoin's price. That absence is not an oversight. It is a symptom of how crypto media produces content: aggregate the loudest narratives, skip the boring structural work, publish. And because the loudest narrative is always the one with a face and a date attached, the election wins the headline while the clock, silent and indifferent, sets the price. So what do I actually do with this. I do not trade the election. I trade the levels, and I let the election be the noise that delivers price to them. If a Democratic outcome produces the feared pullback, I want to know where the leveraged longs get liquidated, because that is where I want to be a buyer. In a post-halving distribution phase, those flush lows are not the end of the cycle. They are the beginning of the next accumulation window. The dip, if it comes, is not a reason to abandon the position. It is a reason to add to it, carefully, in tranches, with the discipline that survival demands. Feel the trend, not the headline. The trend is the cycle. The headline is the weather. If a Republican outcome produces a relief rally, I want to be a seller into strength, because the good news was priced two years ago and relief rallies in fatigued narratives are for exiting, not entering. The asymmetry is in the exits, not the entries. And through all of it, I will keep my eyes on the three variables that actually move this market: the halving cycle's maturation, the direction of global liquidity, and the pace of ETF flows. Those are the load-bearing walls. The election is a coat of paint. It looks important because it is new. It is not new. It is the same four-year clock, wearing a different suit, and it is about to tick again. The question is not who wins in November. The question is whether you will be watching the calendar or the clock when the market finally decides which one it was always following.