One question a day: is Bitcoin about to have a big day. We put a number on it before the day opens, and the next morning the market answers. Right or wrong, the answer stays on the board.
Not "will Bitcoin go up". Nobody calls that reliably, us included, and we print how badly we do at it. The question here is whether today is going to be a big day at all. That one can be called.
Yesterday we said 23%. Bitcoin had to swing 2.41% to count as a big day. It swung 0.82%, so it was not, and the call went on the board as right.
14 indicators read the same day and each one votes. No single one of them matters. What matters is how many point the same way, and today 63% did. Watch that number, because it is usually lower than you would expect: most days they disagree.
The 14 are not a secret: every one of them has an article on this site and the simulator draws them for you. What is not published is the part that took the work: how a reading becomes a score, and what agreement is allowed to be worth. That part is ours, and it stays in the building.
Four readings go in and one number comes out. Every morning they get marked against what happened, and whichever one was furthest off loses some of its say. Keep being wrong and you stop being listened to. Nobody makes that call. The scoring does.
14 indicators on the daily candle, the only one of the four with measured skill of its own.
What the perpetual market is paying to stay long. Crowded longs lean the day slightly the other way.
Sentiment from the curated stream, bounded so it can nudge and never steer.
Trend state against the 200-day. Up or down, nothing in between.
Swipe for act two
Every forecast gets marked the next morning against what Bitcoin actually did. A day that did not resolve cleanly is thrown out rather than guessed. Here is the whole record, misses included.
The number is a pool of forecasters whose weights follow their own graded record, scored every day from 2026-09-26.
The Brier score is how far our number sat from what happened, so lower is better. 0.25 is what you score by saying fifty percent every single day and never thinking about it. Threshold is the middle of the last ninety days: half of them moved more than that, half less. It moved is what the day did, high to low as a share of the open. The record is young; judge it when the count is in the dozens.
The forecast above is about whether today is a big day. This is the other half of the same question: when in the day the movement actually lands. It is two years of hourly Bitcoin prices and one more day every morning, in New York time, because that is where it shows up.
Mean absolute hourly move, New York time · 17,520 hourly prices · one day at a time, no pooling · rank correlation across all 24 hours 0.77 between the two halves · loud against flat on weekday hours t = 20.4 · computed 2026-11-01 · source CoinGecko
It is not the American session. Bitcoin is loudest from 10:00 to 13:00 in New York, every weekday, and then it goes flat. Not when New York goes home, three hours later, but when London does. You are looking at the part of the day when Europe and America are both at their desks, and it ends when Europe leaves. The dead hours are early morning, 05:00 to 09:00 New York, after Asia has gone to bed and before America is up. Friday is the odd one out and empties in the evening instead. Sunday evening is busy, because that is Monday morning in Asia. Saturday has almost no rhythm at all. Seven dials and not one, because a Saturday looks nothing like a Tuesday.
We also went looking for the hours it peaks and bottoms out, across 1512 different windows. There is nothing there. The time of day tells you how much Bitcoin moves, not which way it goes.
A forecast nobody acts on is an opinion. So four paper accounts trade this one, five thousand dollars each, Bitcoin only, and every fill is published.
Reads the Barometer at full resolution: the pool behind the volatility number with each expert's weight, the four direction forecasters, the fourteen indicators behind the composite, and the intraday clock for the day of the week it actually is. Then it decides for itself: open, rest a bracket, close or hold. It is given no rule about what any of it means, and it is told plainly that the direction call is measured near a coin flip. Long and short.
The same reading of the Barometer as StormAI, from the same snapshot at the same instant, with one thing added: it is shown its own past decisions, the thesis it wrote for each one, the falsifier it named, and how that day was graded. Nothing else differs. So the gap between the two is exactly what it is worth to a model to be able to see its own record.
The control. It copies StormAI exactly: the same instant, the same size, the same stop, the same instrument. The only thing it decides for itself is which way, and it decides that at random. It exists to isolate one question. If reading the forecast is worth anything, StormAI has to beat a coin that copies every other choice it makes.
No model at all. It trades only when the Barometer calls a STORM or SEVERE day and stands flat otherwise: a fixed rule, kept as the benchmark it was registered as. Direction comes from the price, never from the forecast; the position closes at the end of the UTC day. It answers the question that was actually asked, whether the forecast is tradeable, without a language model anywhere in the loop.
Trading Trades from the first trigger 23 to go 6.92 graded a week done 2026-11-25
It is not on a schedule. Over the last seven days the watcher looked 670 times and acted 76 of them, standing still for the other 594. Every one of those looks is recorded, because otherwise patience and a broken job look identical.
The band is CHOP at pHigh 0.287, and only 0.37 of the 2.41% threshold has been used, with a 0.88% range so far. The direction score is noise, so I have no basis for an outright long or short, and a breakout bracket would be betting against the day's size forecast. The loud 19:00-21:00 New York block is still ahead, but loudness has no direction and doesn't justify paying for a position by itself. CHOP's 4-for-4 record is too small to lean on and the overall Brier of 0.2524 is no better than a coin, so this is an absence of edge rather than a confident fade, and I also won't short volatility. The falsifier is 79305, the day low of 77438 plus the 2.41% threshold: a print there would make it a high-volatility day and show that an armed long should have been on; the snapshot doesn't state the day's cutoff, so the deadline of midnight New York time is my assumption.
Wrong if: a print at or above 79305 before 2026-11-02T05:00:00.000Z
The price-break wake names no level, and the 77,438 to 78,123 range is unchanged since my 09:35Z hold, with spot at 77,699 mid-range, so a market position would be a direction bet with no edge (AUC 0.52). The 0.88% range is 37% of the 2.41% threshold, and I estimate about 0.75% of typical movement left before 00:00Z from the clock's hourly means, assuming they add like a random walk, so a big day is still unlikely on this CHOP call (pHigh 0.29). The real case for a bracket is Sunday's loud block, 19:00 to 22:00 New York, which starts at 00:00Z, but the next forecast should arrive around 00:20Z, so I'd rather decide then with Monday's call than commit now without it; CHOP's 4 for 4 is too few days to lean on, and the pool's 0.2524 Brier is no better than a coin. My record shows a pattern: every settled hold scored right, but those levels were mostly too far away to test much, so this falsifier is deliberately closer. If 76,900 prints before 00:00Z, a short bracket under the low near 77,350 with about a 0.6% target would have paid, and holding was the wrong call.
Wrong if: a print at or below 76900 before 2026-11-02T00:00:00.000Z
the treatment did not act, so there is nothing to mirror
Wrong if: None of its own. This arm has no thesis, which is the point of it.
Band is CHOP. This arm trades STORM and SEVERE only.
The pool says CHOP at pHigh 0.287, with 58% of its weight on har3 and the Sunday weekday effect, and the day has used only 37% of its 2.41% threshold after yesterday's 0.82% range. The price break that woke me is a small move in a quiet hour (rank 20 of 24), and today's loud block, 19:00-21:00 New York, starts just as this UTC day ends, so a bracket has little reason to pay today. The evidence for the band is thin: CHOP is 4 for 4, but that is four days, and overall the pool's Brier of 0.2524 is no better than a coin at 0.25, so I am not treating it as an edge for fading the range either. Direction is noise at a 5.5 score, and the news forecaster is saying nothing, so there is no basis for picking a side. The falsifier is 79305, the day low of 77438 plus the 2.41% threshold, because printing it would make this a high-volatility day and mean a long breakout bracket should have been armed.
Wrong if: a print at or above 79305 before 2026-11-02T00:00:00.000Z
The price-break that woke me looks like a new Sunday high of 78,123, up from 78,092 at my 08:50Z hold, but spot has already slipped 0.26% below it to 77,922, in a quiet hour (rank 20 of 24, about 0.19% mean move). The 0.88% range is 37% of the 2.41% threshold, so the day clears only if the high reaches about 79,305 (1.77% above spot) or the low about 76,283 (2.10% below), and a bracket that pays needs the big day this CHOP call (pHigh 0.29) rates unlikely. A market position here would be a direction bet with no edge (AUC 0.52). I don't lean hard on the forecast, since CHOP's 4 for 4 is a small sample, the pool's overall Brier of 0.2524 is no better than a coin at 0.25, and har3 and clim hold 87% of the weight, but doubting it isn't a reason to pay for a breakout. My long run of 'right' hold verdicts proves little because those levels were far away and I have made no equity from them, so if this level prints, this hold missed a big day that a long bracket would have caught.
Wrong if: a print at or above 79310 before 2026-11-02T00:00:00.000Z
the treatment did not act, so there is nothing to mirror
Wrong if: None of its own. This arm has no thesis, which is the point of it.
Band is CHOP. This arm trades STORM and SEVERE only.
The price-break wake is a new day high at 78092, but the day's range is only 0.84%, or 35% of the 2.41% threshold, and it came in the quietest hour on the Sunday clock. The pool says CHOP with pHigh 0.287, led by har3 and the Sunday base rate at about 0.25; CHOP is 4 for 4, but that is too small a sample to lean on, and the overall Brier of 0.2524 is no better than a coin. Direction is noise (score 5.5, pCloseAboveOpen 0.4995), so I won't take a side, and a bracket has little support when the forecast gives under a 30% chance of the range expansion a break needs to pay. The loud 19:00-21:00 New York block could still produce a move, but the clock doesn't say which way, so I'm not paying for it in advance. A print at 79306 would take the range from today's 77438 low past the threshold, which would mean a long or an armed upside break was the right call and CHOP was wrong.
Wrong if: a print at or above 79306 before 2026-11-02T00:00:00.000Z
The price-break that woke me is a new Sunday high of 78,092, which is also spot, but the 77,438 to 78,092 range has used only 35% of the 2.41% threshold, and it came in the quietest hour of the day (rank 24 of 24, 0.14% mean move). A long here would be a direction bet at the day's high with no edge (AUC 0.52), and the day clears the threshold only if the high reaches about 79,305 (1.55% above spot) or the low about 76,253 (2.35% below), so a bracket that pays needs the big day this CHOP call (pHigh 0.29) rates unlikely. I don't lean hard on that number: CHOP's 4 for 4 is too few days, the pool's overall Brier of 0.2524 is no better than a coin at 0.25, and har3 and clim hold 87% of the weight while DVOL-aware har2 reads 0.43 on 6% of it, but doubting the forecast is not the same as having a reason to pay for a position. My record is a run of holds with no equity change, and their 'right' verdicts prove little because I keep naming far-off levels that rarely print, so this hold is a no-edge decision and I am not claiming it as skill. If 79,310 prints before the deadline, the day cleared the threshold upward and sitting out cost me a breakout I could have bracketed.
Wrong if: a print at or above 79310 before 2026-11-02T00:00:00.000Z
the treatment did not act, so there is nothing to mirror
Wrong if: None of its own. This arm has no thesis, which is the point of it.
Band is CHOP. This arm trades STORM and SEVERE only.
The band is CHOP at pHigh 0.287, with har3 and the Sunday climatology carrying about 87% of the pool's weight, and only 0.19% of the day's range is used against a 2.41% threshold. A bracket needs a large day to pay and this forecast doesn't expect one, while the direction read (5.5, neutral, an input I'm told is noise) gives no reason to open a side. The wake-up came from entering a loud clock block, but the block's mean hourly move is only about 0.21%, which is size rather than opportunity and too small to justify paying for a whipsaw. CHOP's 4 for 4 is a tiny sample and the overall Brier of 0.2524 is no better than a coin, so I'm not treating the band as proven, but an unproven quiet call is still no reason to trade. The falsifier is 79305, derived from the day low of 77438 plus the threshold range: if that prints today, the day has gone high-volatility and a long-side break would have been the right position.
Wrong if: a print at or above 79305 before 2026-11-02T00:00:00.000Z
The loud block that woke me (21:00 to 24:00 New York) averages only 0.21% per hour, and a clock reading says nothing about direction, so it gives no reason to act; spot at 77,586 is the day's high, and a long here would be a direction bet with no edge (AUC 0.52). The day has used 8% of the 2.41% threshold and clears it only near 79,310 (2.2% above spot) or 75,760 (2.35% below), so a bracket that pays needs the big day this CHOP call (pHigh 0.29) rates unlikely. I don't lean hard on that number: CHOP's 4 for 4 is too small a sample, the pool's overall Brier of 0.2524 is no better than a coin, and har3 and clim hold 87% of the weight while har2 reads 0.43. My record also shows that falsifiers this far out nearly always settle 'right', even on a 3.23% day, so those verdicts say little and the holds have earned nothing, but that alone is no reason to pay for a trade with no edge. This falsifier covers the upside, while my 00:20Z hold's 75,650 still covers the downside, and if 79,310 prints by the deadline, sitting out was the wrong call.
Wrong if: a print at or above 79310 before 2026-11-02T00:00:00.000Z
the treatment did not act, so there is nothing to mirror
Wrong if: None of its own. This arm has no thesis, which is the point of it.
Band is CHOP. This arm trades STORM and SEVERE only.
The pool calls CHOP at pHigh 0.287, led by har3 and the Sunday climatology rate, and yesterday's range was only 0.82% against a 2.41% threshold, so a bracket has little reason to expect room to run. The band's 4 for 4 is too small a sample to lean on, and the overall volatility Brier of 0.2532 is no better than a coin's 0.25, so this is weak support for a quiet day, not an edge to sell. The direction inputs are noise by measurement, and the news forecaster is saying nothing, so I have no basis for an outright long or short. Sunday's loudest block does average 0.47% an hour, but that measures size and not direction, and it does not justify paying for whipsaws on a day forecast as sub-threshold. The falsifier is a print above 79,350, about 2.4% above the day open, which would mean the day reached a high-volatility range and a long breakout bracket would have paid; a similar break down near 75,600 would count against this read too, but only one level can be recorded.
Wrong if: a print at or above 79350 before 2026-11-02T00:00:00.000Z
Sunday's new CHOP call (pHigh 0.29) gives me no size view to trade: the day has used only 2% of the 2.41% threshold, and it clears only near 79,305 (2.41% above spot) or 75,650 (2.31% below), so a bracket out there needs the big day this forecast calls unlikely. CHOP is 4 for 4, but that is too few days to lean on and the pool's overall Brier of 0.2524 is no better than a coin at 0.25; har3 and clim hold 87% of the weight at 0.27 and 0.25, while DVOL-aware har2 reads 0.43 on 6% of it, so I doubt the number a little without having a case for the opposite. A market position 0.05% from the open would be a direction bet, and the direction call has no edge (AUC 0.52, score 5.5, neutral lean). Sunday's loudest block, 19:00 to 22:00 New York at about 0.47% an hour, should start around 00:00Z Monday after today's clock change, which is outside this forecast's day if that day runs to 00:00Z as the 00:20Z open suggests, so it doesn't support a bracket on today's call. If the low prints 75,650 before that time, the day has cleared the threshold and a bracket would have been the right call.
Wrong if: a print at or below 75650 before 2026-11-02T00:00:00.000Z
the treatment did not act, so there is nothing to mirror
Wrong if: None of its own. This arm has no thesis, which is the point of it.
Band is CHOP. This arm trades STORM and SEVERE only.
The band is CHOP at pHigh 0.227, and with about 55% of the UTC day gone the range is 0.66%, only 27.5% of the 2.41% threshold, at $77,404 inside a $511 range. That supports neither a direction, since the direction read is noise at 5.5, nor a breakout bracket, since a level near this range is likely to whipsaw on a quiet Saturday and a level far from it leaves little room to run. I don't fully trust the low number: it leans on har3's Saturday feature (0.06 at 52% weight) while har1 and har2 sit at 0.53 to 0.58, the prior day's range was 3.23%, and last Saturday's CALM call missed narrowly at 2.44%, while CHOP's 3 for 3 is too small a sample to carry much weight. That disagreement is a reason for caution, not an edge worth paying a spread for, so I stay flat. A print above $78,950 today would put the range past the threshold measured from today's $77,076 low, showing that a long bracket should have been armed and this hold was wrong; the snapshot gives no reason to prefer the upside, and this is simply the one side the falsifier format lets me name.
Wrong if: a print at or above 78950 before 2026-11-01T00:00:00.000Z
The wake doesn't name a level, but the day high is now 77,587, up from 77,542 at my 12:05Z hold, while spot has slipped back to 77,404, 0.24% under it. The 77,076 to 77,587 range is 0.66%, 27.5% of the 2.41% threshold, so the day clears only if the high reaches about 78,934 (1.98% above spot) or the low about 75,760 (2.12% below). A bracket out there needs a big day against a CHOP call with pHigh 0.23; I doubt that number somewhat, since har3 and clim hold 79% of the weight on weekday effects while har1, har2 and persist read 0.53 to 0.65, and CHOP's 3 for 3 is too few days to lean on, but that doubt doesn't amount to a positive case for paying for a break. A market position here would be a direction bet with no edge (AUC 0.52), in a below-average hour (rank 16 of 24, about 0.17% mean move). My record shows a run of holds, and Friday's STORM day cleared its threshold with nothing on, so that is a cost worth watching on a high-pHigh day, but it isn't a reason to force a trade on a low-pHigh Saturday.
Wrong if: a print at or above 78950 before 2026-11-01T00:00:00.000Z
the treatment did not act, so there is nothing to mirror
Wrong if: None of its own. This arm has no thesis, which is the point of it.
Band is CHOP. This arm trades STORM and SEVERE only.
The wake-up is a new day high at 77,542, but the range so far is only 0.60%, about a quarter of the 2.41% threshold, and hour 8 ranks 18th of 24 for loudness. The pool says CHOP at 0.227, so a breakout bracket would be paying to bet against the forecast, and fading the high would be a direction call with no measured edge. That 0.227 leans heavily on har3 and clim (79% of the weight), while har1 and har2 sit above 0.5 after yesterday's 3.23% range, and last Saturday's 0.10 call was wrong at 2.44%, so CHOP's 3-for-3 record is too thin to trade on its own. With no edge in either direction and a pool that overall is no better than a coin on Brier (0.2523 vs 0.25), I stay flat. A print at 78,940, which would take the range from today's 77,076 low past the threshold, would mean a long or a long-side bracket was the right call.
Wrong if: a print at or above 78940 before 2026-11-01T00:00:00.000Z
The price-break that woke me is a new Saturday high of 77,542, which is also spot, but the 77,076 to 77,542 range is only 0.60%, 25% of the 2.41% threshold. The day clears the threshold only if the high reaches about 78,934 (1.8% above spot) or the low about 75,717 (2.4% below), and a bracket set at those levels needs a big day on a CHOP call with pHigh 0.23, while only 2 of the last 12 Saturdays cleared it. A long here would be a direction bet at the day's high with no edge (AUC 0.52), in a below-average hour (rank 18 of 24, about 0.16% mean move). I don't fully trust the low number: har3 (0.06) and clim (0.17) hold 79% of the weight, har1, har2 and persist read 0.53 to 0.65, CHOP's 3 for 3 is a tiny sample, the overall Brier of 0.2523 is no better than a coin, and last Saturday's CALM call at 0.10 missed a 2.44% range. That doubt is grounds for a smaller view rather than a trade, and I note that my recent falsifiers have mostly sat far from price, so their 'right' verdicts show little; this one is the actual threshold-clearing level.
Wrong if: a print at or above 78950 before 2026-11-01T00:00:00.000Z
the treatment did not act, so there is nothing to mirror
Wrong if: None of its own. This arm has no thesis, which is the point of it.
Band is CHOP. This arm trades STORM and SEVERE only.
I was woken by a price break, but spot at 77133 is just a new day low inside a 0.44% range, only 18% of the 2.41% threshold. The pool gives a 0.23 chance of a high-volatility Saturday. That number is carried mostly by har3 (weekday-aware, weight 0.52) and the same-weekday base rate (weight 0.27); the two models that read the shock, persist and har1/har2, are given little weight. The intraday clock is in its quietest hours, and Saturday's loudest block is only 0.21% per hour. That leaves little room for a breakout bracket to run before costs and whipsaw eat it. The direction inputs sit at 0.50 and are noise. The overall vol record is no better than a coin (Brier 0.252 against 0.25), and CHOP's 3 for 3 is too few days to lean on, so I see no reason to pay to trade a quiet day. If the low reaches 75600 today, the day range clears the threshold measured from the 77471 high, and a short-side bracket would have been the right call.
Wrong if: a print at or below 75600 before 2026-11-01T00:00:00.000Z
The price-break that woke me is a new Saturday low of 77,133, which is also spot. The day's range is 77,133 to 77,471, or 0.44%, which is 18% of the 2.41% threshold. The day clears the threshold only if the low reaches about 75,600 or the high reaches about 78,993. The forecast is CHOP with pHigh 0.23, and it is carried by har3 (0.06) and clim (0.17), which hold 79% of the weight. CHOP is 3 for 3, but three days is too few to lean on, and har1, har2 and persist read 0.53 to 0.65, so a big day is not ruled out. Even so, a bracket needs a big day to pay, and this call says one is unlikely. Shorting the new low at market would be a direction bet, and the direction call has no edge (AUC 0.52). This is the quietest hour of the week's clock (rank 24 of 24, about 0.18% mean move), and the Saturday hours ahead are quieter still. I will hold, and the call is wrong if the low prints 75,600 before Sunday 00:00Z, because that is a threshold-clearing day a bracket would have caught.
Wrong if: a print at or below 75600 before 2026-11-01T00:00:00.000Z
the treatment did not act, so there is nothing to mirror
Wrong if: None of its own. This arm has no thesis, which is the point of it.
Band is CHOP. This arm trades STORM and SEVERE only.
The pool calls Saturday CHOP with pHigh 0.23, and the weight sits with har3 (0.06, which knows it is Saturday) and clim (0.17). CHOP is only 3 of 3 right, but it is the band's best Brier score. A breakout bracket on a quiet weekend is the kind of trade that gets stopped out in both directions, and the overall vol record (Brier 0.252) is no better than a coin, so the band does not justify spending equity that is already down $345. The direction inputs are noise, so I am not trading a side. If price prints above 79100, the day's range from the 77231 low would pass the 2.41% threshold, and an upside break bracket would have been the right call.
Wrong if: a print at or above 79100 before 2026-10-31T23:55:00.000Z
The price-break that woke me is a new Saturday high of 77,471. That puts the day's range at only 0.31%, 13% of the 2.41% threshold. The low is the 77,231 open, so the day clears the threshold only if the high reaches about 79,093 or the low drops to about 75,600. The forecast is CHOP with pHigh 0.23, and CHOP is 3 for 3 so far, though three days is too few to lean on. Pool weight is still heavy on weekday effects (har3 at 0.06 and clim at 0.17 hold 79%), while har1, har2 and persist read 0.53 to 0.65. Buying a 0.31% rise off the open in one of the quietest hours (rank 22 of 24, about 0.18% mean move) would be a direction bet with no edge (AUC 0.52), and a bracket would need the big day this forecast says is unlikely, so I'm holding and keeping my 00:20Z falsifier: 79,100 printing before 00:00Z Sunday would mean the day was big and a long would have been the right call.
Wrong if: a print at or above 79100 before 2026-11-01T00:00:00.000Z
the treatment did not act, so there is nothing to mirror
Wrong if: None of its own. This arm has no thesis, which is the point of it.
Band is CHOP. This arm trades STORM and SEVERE only.
The pool says CHOP at pHigh 0.227 for a Saturday, there is no directional edge (score 5.5, AUC 0.52), and with no size edge either a bracket is more likely to be whipsawed than to run. The low reading is not well supported: it rests on har3's weekday-driven 0.063 at 52% weight, while har1 and har2 read 0.53 and 0.58, the prior two days cleared the threshold, last Saturday's CALM call at 0.103 missed, CHOP's 3 for 3 is only three days, and overall Brier (0.25225) is no better than a coinflip. That disagreement argues against fading the day as well as against paying for a breakout, so I see no trade worth taking. The falsifier is the open minus the 2.41% threshold, about 75369; a print there means the day was large enough that a downside bracket would have been the right call, and the choice of the lower side is arbitrary because I have no directional view.
Wrong if: a print at or below 75369 before 2026-11-01T00:00:00.000Z
Saturday's new CHOP call (pHigh 0.23) offers nothing to trade: a bracket needs a big day, this forecast says one is unlikely, and only 2 of the last 12 Saturdays cleared the threshold (the 0.167 clim rate). A market position 0.05% from the open would be a direction bet with no edge (AUC 0.52), and this hour is quiet (rank 21 of 24, about 0.19% mean move). I don't fully trust the low number: har3 (0.06) and clim hold 79% of the weight on weekday effects, while har1, har2 and persist read 0.53 to 0.65, CHOP's 3 for 3 is only three days, the pool's overall Brier of 0.2577 is worse than a coin's 0.25, and last Saturday's CALM call at 0.10 cleared its threshold. That doubt isn't enough to justify a bracket, because the day's clock is also quiet (loudest block about 0.21%) and my Friday bracket sat outside the levels that cleared the threshold, so it earned nothing on a day the size call was right. The snapshot still doesn't show whether that bracket's 76,600 short filled after the low hit 76,502, and the 24.95 gap between cash and realised P&L remains unexplained, so I'm not sizing up on an account I can't fully reconcile. The falsifier is 79,100, about where the day clears 2.41% upward if the 77,231 low holds; it tests one side only, and the downside mirror is roughly 75,450.
Wrong if: a print at or above 79100 before 2026-11-01T00:00:00.000Z
the treatment did not act, so there is nothing to mirror
Wrong if: None of its own. This arm has no thesis, which is the point of it.
Band is CHOP. This arm trades STORM and SEVERE only.
The STORM call has already been met: today's range of 2.74% is 1.156 times the 2.37% threshold, so the forecast no longer implies more room from here. Spot at 77241 sits mid-range, about 1% from the 76502 low and 1.8% from the 78598 high, so there is no edge in a market order, and the direction inputs are noise by the experiment's own measure. The clock points to the quietest stretch ahead, with 19:00 to 23:00 New York hours averaging about 0.18% to 0.21% and Saturday's blocks flatter than Friday's, so a bracket around the day's extremes would mostly risk a false break with little room to run. The pool's live record also doesn't justify paying for this: its vol Brier of 0.268 is worse than 0.25 for a coin, and STORM days are 6 right of 10. A print below 76000 by 14:00 UTC would mean the day low broke and ran about 0.7% further, so a short bracket at that low would have been the right call; the field takes one side only, and this is not a directional view.
Wrong if: a print at or below 76000 before 2026-10-31T14:00:00.000Z
The price-break alert doesn't name a level, and the day's 76,502 to 78,598 range is unchanged since my 20:05Z hold; spot has only recovered 0.47% to 77,241, mid-range. The 2.74% range is already 1.16 times the 2.37% threshold, so Friday's size call is answered and a position opened here would be a direction bet with no edge (AUC 0.52), in a quiet hour (rank 20 of 24, about 0.21% mean move, with quieter hours next). I'm not arming a new bracket because the snapshot has no forecast for Saturday, so there is no size view to trade yet, and I'll decide when one arrives. My record shows many holds settled 'right', but most of those falsifiers sat far from spot, so they say little about skill and earned nothing, which is why this level is closer, 0.98% above spot. The snapshot still doesn't explain whether the 76,600 short trigger filled when the low went through it, or the 24.95 gap between cash of 4,874.02 and the 4,898.97 implied by 5,000 less the -101.03 realised.
Wrong if: a print at or above 78000 before 2026-10-31T06:00:00.000Z
the treatment did not act, so there is nothing to mirror
Wrong if: None of its own. This arm has no thesis, which is the point of it.
Already holding. This arm carries one position at a time.
The STORM call has already been used up: the day's range is 2.74% against a 2.37% threshold, so the band says little about what is left of the day. The clock now runs through Friday's quieter hours, with mean moves falling from 0.28% to about 0.18% per hour, which gives a breakout less room and makes a whipsaw through a tight bracket more likely. The vol record does not justify forcing a trade either, with STORM right on 6 of 10 days and an overall Brier slightly worse than a coin, and the direction score is noise, so I have no basis for picking a side. A print below 75,900, about 0.8% under the 76,502 day low, by the deadline would mean the expansion was still running and a short bracket under the low would have been the right call.
Wrong if: a print at or below 75900 before 2026-10-31T08:00:00.000Z
The day's low reached 76,502, under my 76,600 short trigger, yet the snapshot shows no position, no pending bracket and no trade on my record, so it doesn't tell me whether that bracket ever filled; cash of 4,874.02 is also 24.95 below what 5,000 less the -101.03 realised implies, and nothing here explains the gap. The 2.74% range is already 1.16 times the 2.37% threshold, so the size call is answered and spot at 76,879, 0.49% above the low, offers only a direction bet, which has no edge (AUC 0.52). I won't arm a new bracket: this hour's 0.28% mean move falls to about 0.18% by New York's late evening, Saturday's loudest block is only 0.21%, and there is no Saturday forecast in the snapshot to base a size view on. My record shows two mistakes: today's triggers sat past the levels that cleared the threshold, so the size call earned nothing, and my falsifiers at 79,600 were so distant that their 'right' verdicts prove little. This one is closer on purpose, 1.14% below spot in quiet hours, and if 76,000 prints before 04:00Z then a short or a fresh bracket would have been the better call.
Wrong if: a print at or below 76000 before 2026-10-31T04:00:00.000Z
the treatment did not act, so there is nothing to mirror
Wrong if: None of its own. This arm has no thesis, which is the point of it.
Already holding. This arm carries one position at a time.
The forecast goes out every day whether anyone reads it or not. If you want to find out whether you can do better, the simulator runs on the same live prices with a paper account.