satoshimacro

$BTC | Is September really Bitcoin's worst month, or is August the one to watch? The two answers disagree depending on which number you use. September has the lower average return. August has the better average but a far worse median and the worst win rate of any calendar month. That gap matters more than the label.
## What does "worst month" actually mean?
On the desk we never trusted an average without checking the median behind it. SatoshiMacro's monthly returns heatmap tracks every calendar month of Bitcoin price history in AUD terms since 2013, and across 2014 to 2025 the headline numbers for the two weakest months read almost the same at a glance. September averages -1.9 percent with a -1.3 percent median and closes positive in 6 of 12 years, exactly 50 percent. August averages a slightly milder -1.0 percent, yet its median is -8.3 percent and it has closed positive in only 3 of 12 years, a 25 percent hit rate. One of those months is consistently mediocre. The other is bimodal.
## Why the mean and the median disagree so sharply in August
Here is the specific reason. August 2017 returned plus 67.4 percent, the second best single month in the entire AUD dataset behind only May 2017's plus 70.4 percent. That one outsized rally drags August's twelve year average most of the way back toward zero while doing nothing for the typical August, which the median shows was a losing month three years out of four. September has no equivalent outlier in either direction. Its distribution is narrower, and its average and median tell roughly the same story. If you are sizing a position around "the weak month," August is the one where the average lies to you.
## What this means for cycle top timing
The heatmap's own FAQ makes an honest admission worth repeating: prior Bitcoin cycles peaked in November or December, but the current cycle broke that pattern. Its highest daily close so far sits at A$187,780 on 13 August 2025, not in the traditional fourth quarter window, and the price has since drawn down heavily from that level. Seasonality predicted a Q4 top. It did not show up on schedule. My read is that calendar patterns describe what usually happens across three visible cycles, not what will happen in this one, and three cycles is a small sample for any timing conclusion.
## Should October still get credit as Uptober?
Yes, with a caveat. October is the strongest calendar month by every measure the heatmap tracks: plus 16.8 percent average, plus 15.9 percent median, and a 75 percent win rate across 2014 to 2025, the best combination of any month. But "mostly" is the honest word. The page flags three negative Octobers in that window, in 2014, 2018 and 2025. When I size anything around a calendar effect, I size it like background noise sitting underneath the real catalysts (halvings, ETF flows, macro liquidity), never as the thesis itself.
## Does any of this change what you actually pay tax on?
This is where the Australian angle matters, and where seasonality data quietly runs out of road. The heatmap shows raw AUD price returns, not after tax returns, and the two are not the same thing. If you are holding past the ATO's 12 month mark to qualify for the 50 percent CGT discount, a weak September or a strong October changes your paper position, not your tax timing. SatoshiMacro's CGT calculator models the discount separately from price action for exactly this reason, because the two questions, when did the price move and when does my holding period reset the discount, have different answers and get confused constantly in year end planning.
## The bottom line
Across 144 months from 2014 to 2025, 80 closed positive, 55.6 percent, for an average monthly return of plus 5.55 percent. Nine of the twelve calendar months carry a positive average. That is a genuinely favourable base rate, not a reason to treat any single month as predictable. September's average is weak but stable. August's average flatters a month that is actually worse than it looks. October earns its reputation in most years, though not every year. Past performance here is exactly that, past performance: a statistical tendency with a small sample caveat attached, not a signal you trade on its own.
https://satoshimacro.com/tools/crypto/cycle-indicators/bitcoin-monthly-returns-heatmap/?utm_source=binance_square&utm_medium=social&utm_campaign=autopilot_article
#SatoshiMacro #BitcoinSeasonality #CryptoTax #Bitcoin

$BTC | If you missed the 30 June window to harvest a crypto loss in Australia, is the deduction gone?
No. A realised capital loss on crypto does not expire. If you did not sell an underwater position before 30 June 2026, you lost the chance to offset gains in the 2025-2026 year specifically, not the loss itself. Under section 102-15 of the ITAA 1997, any loss you do realise carries forward indefinitely until you have gains to use it against.
This matters right now because the calendar is confusing people. The 31 October deadline sitting ten days out is a lodgment deadline for the 2025-2026 return, not a second chance at tax-loss harvesting for that year. On the desk we used to get this exact question from junior traders every October: "can I still do something about last year's tax." For ordinary income the answer is almost always no once the year has closed. For capital losses, the answer is more useful than that, and most retail crypto holders do not realise it.
## Why the 30 June window closing does not erase the loss
Tax-loss harvesting in Australia works in a fixed order. You realise a loss by disposing of a position (selling to AUD, swapping to another token, gifting it, or sending it somewhere you no longer control), then that loss nets against capital gains realised in the same financial year, and only the remaining net gain gets the 50 percent CGT discount if held over 12 months. If 2025-2026 is already closed and you never disposed of the losing position, you have not harvested anything yet. Nothing has crystallised. The loss is still sitting unrealised on your balance sheet, and whether to realise it now is a decision for 2026-2027, not a missed deadline for 2025-2026.
Where people do lose the benefit is a different mistake entirely: selling at a loss and buying the same coin straight back with no real change in position. The ATO does not need a US-style wash-sale rule to catch that; it has Part IVA, the general anti-avoidance provision, and a sell-and-instant-rebuy pattern is close to the textbook example of what it targets.
## What the ATO actually lets you carry forward
Once a loss is realised, there is no cap and no time limit on carrying it forward. SatoshiMacro's tax-loss harvesting calculator models this directly: enter unrealised losses against this year's realised gains and it outputs the carryforward balance alongside the immediate tax saved. Take a trader with AUD 8,000 of short-term gains already locked in and AUD 25,000 of unrealised losses on positions they are exiting anyway. Harvesting now fully offsets the AUD 8,000 (saving roughly AUD 2,400 at a 30 percent marginal rate) and leaves AUD 17,000 to carry into next year. That carryforward balance is not a consolation prize. It sits on the tax account at full value until a future gain consumes it.
## The case with zero gains this year is the one people get wrong
Say there are no realised gains at all in 2025-2026, just AUD 12,000 of losses on positions already decided against. Harvesting produces no saving this year, full stop, because a capital loss can only reduce capital gains, never salary or other ordinary income. But if a large long-term gain is expected later, say when a long-held BTC position eventually gets sold, that AUD 12,000 carries forward and applies before the 50 percent discount, before marginal rate. My read is that the "no current benefit" framing is what stops people from harvesting a position they already want out of. The benefit is deferred, not absent, and an unrealised loss costs nothing to wait on either way.
## Where the calculator's estimate is honest about its own limits
The tool treats all of a year's gains as one cohort matching a single holding-period selection, long-term or short-term. Real portfolios usually mix both, and the ATO's actual ordering rule applies losses against non-discounted gains first because they are taxed at the full rate. That means a real mixed portfolio will equal or beat the calculator's estimate, never fall short of it. It is a planning tool, not a lodgment-ready number, and SMSF holders need the separate SMSF calculator since fund-level gains sit at 15 percent in accumulation phase with a one-third discount rather than the individual 50 percent rate.
## What I would actually do with 10 days to 31 October
Lodge the 2025-2026 return on what already happened; do not try to retrofit a harvest into a year that is closed. Then use the next eight months properly: track unrealised losses as they appear rather than waiting for a May scramble, and keep the recordkeeping clean. ATO record-keeping rules run five years from lodgment, covering date, AUD value, counterparty and fees for every disposal, loss or gain. That discipline is what makes next year's harvest decision a five-minute calculation instead of a guess.
One more thing worth flagging before 31 October: a crypto-to-crypto swap is a disposal in its own right. Swapping ETH for SOL, or USDC for a different stablecoin, can realise a loss (or a gain) without ever touching AUD. If any of this year's swaps sat on a loss versus your AUD cost base at the time, that loss should already be in this return, not next year's. It is an easy one to miss when nothing ever hit a bank account.
Full methodology, the worked examples above, and the Part IVA detail live on SatoshiMacro's calculator page.
https://satoshimacro.com/tools/crypto/calculators/tax-loss-harvesting-calculator/?utm_source=binance_square&utm_medium=social&utm_campaign=autopilot_article
#SatoshiMacro #CryptoTax #TaxLossHarvesting #Bitcoin

$BTC | Yes. Bitcoin's Pi Cycle Bottom Indicator fired for the fourth time since 2014 on 1 September 2026, when the price closed at A$108,493 and the 150-day moving average crossed back above the 471-day moving average times 0.745. SatoshiMacro's tracker still has the ratio at 1.066 as of 18 October 2026, sitting inside the "Bottom signal forming" band. It is not a buy signal. It is a late confirmation that typically shows up months after the actual low.
## Why does a crossing moving average count as a bottom signal?
The mechanics are simple even if the reasoning behind the exact numbers is not. Take a 150-day simple moving average of Bitcoin's price and a 471-day simple moving average multiplied by 0.745. When the shorter line crosses back up through the longer, scaled line, the model calls a cycle bottom. Philip Swift built this as the mirror image of his more famous Pi Cycle Top (111-day MA versus 350-day MA times two). SatoshiMacro runs both off daily BTC/AUD closes going back to January 2013, blending Bitstamp's USD series converted at the daily AUD/USD rate with Kraken's native AUD market for the most recent two years.
On the desk, we always treated these curve-fitted ratios as descriptive, not predictive. The 150 and 471-day windows and the 0.745 multiplier were chosen because they fit past bottoms well, not because they were derived from first principles. There is a cute mnemonic floating around that 150 times pi is roughly 471, which is a fun coincidence rather than a real explanation. SatoshiMacro's own page is upfront about that, and so am I.
## What happened the previous three times it fired?
History here is short. But it is consistent in one respect. It is always late.
- 21 July 2015, BTC at A$375. The actual low landed around 14 January 2015 near A$210.
- 22 May 2019, BTC at A$11,017. The actual low landed around 15 December 2018 near A$4,401.
- 24 March 2023, BTC at A$41,145. The actual low landed around 21 November 2022 near A$23,595.
Each firing arrived four to six months after the real bottom, not before it. There was no firing at all during the March 2020 COVID crash, which is the kind of gap that should temper anyone's confidence in a four-data-point model.
## So does the September firing mean the bottom is already behind us?
My read is that this is confirmation of a pattern already in progress, not a new signal to act on. SatoshiMacro's page is explicit that the September 2026 firing has no 12-month forward record yet, so we genuinely do not know how this one resolves. Past 12-month returns after a firing have been positive but mixed, including a period in 2020 where price fell back below the firing level before the next leg up. If you are waiting for this indicator to tell you the all-clear has sounded, you will be waiting for a confirmation that, by design, only ever arrives after the fact.
## What's the catch with trusting this?
A few honest limitations worth sitting with before anyone leans on this. The sample is small. Four firings in twelve years is not enough to build real statistical confidence. The parameters are heuristic, reverse-engineered to fit history rather than grounded in market structure, so they can stop working the moment that structure changes. And the lag is structural, not incidental, since both moving averages are trailing by definition. SatoshiMacro treats it as one input among several cycle-bottom signals, not a standalone call, and weights it inside Tier 1 of the SatoshiMacro Model at 25 percent of that tier's composite rather than letting it drive the whole read on its own.
## How does this change what I would actually do with it?
What I would actually do is use the firing as a prompt to check the rest of the cycle-bottom picture rather than as a trigger on its own. If you are an Australian investor sizing entries around a cycle low for CGT purposes, the practical layer still matters more than the single best price: the ATO's 12-month discount clock only starts running from the date you actually acquire the asset, so front-loading entries during a confirmed bottom window, rather than chasing the exact low tick, is what sets the next disposal up to clear that discount threshold cleanly. That is a tax-timing argument, not investment advice, and it is estimate-level thinking, not a substitute for your own accountant.
Four firings is not much to build a thesis on. Yet each one has, so far, landed inside a period most traders would now call a genuine cycle low in hindsight. That is exactly why it still earns a place in a wider model, not a seat at the head of the table.
https://satoshimacro.com/tools/crypto/cycle-indicators/bitcoin-pi-cycle-bottom-indicator/?utm_source=binance_square&utm_medium=social&utm_campaign=autopilot_article
#SatoshiMacro #BitcoinCycleAnalysis #OnChain #Bitcoin

$BTC | Can you short Bitcoin to protect a spot holding you already own, without taking on a second, separate bet?
Yes, but only if you treat the short as its own leveraged position with its own stop and its own cost, not as free insurance on the coins you already hold. In Australia an ASIC-regulated crypto CFD is the practical way to do it, because a CFD never requires you to own the coin, so you can open a sell position directly against a holding you are not ready to sell outright.
## What does hedging a Bitcoin holding with a short actually mean?
If you hold spot Bitcoin and expect a near-term pullback but do not want to sell (a tax event, a cold-storage withdrawal you do not want to redo, or simple conviction in the long-term thesis), a short CFD lets you open a sell position that profits if the price falls, offsetting some of the paper loss on your coins. SatoshiMacro's guide to shorting Bitcoin frames this as one of only two reasons to short at all: betting on a fall, or hedging a holding you intend to keep. The short is a second, separate position. It does not touch your spot coins.
## How much hedge does the ASIC leverage cap actually buy you?
On the desk, sizing a hedge always started with the cap, not the conviction. ASIC caps crypto CFD leverage at 2:1, the lowest ratio it applies to any retail asset class, and requires negative balance protection so a client cannot lose more than their account balance. At that cap, a 500 dollar deposit controls a 1,000 dollar short position, meaning you need roughly half your intended hedge notional sitting in margin before you open it. If your spot holding is worth 2,000 dollars and you want to hedge the full amount, you are tying up around 1,000 dollars of fresh capital to do it. That capital earns nothing while the hedge is open.
## What happens if the hedge moves against you?
This is the part traders skip. If Bitcoin rallies 20 percent instead of falling, that 1,000 dollar short loses 200 dollars, which is 40 percent of the 500 dollar margin behind it, even though your spot coins just gained value. The two legs are meant to offset. In practice they rarely offset cleanly: different venues, different fee structures and a funding cost on the CFD leg mean the hedge will not match your spot position dollar for dollar.
## Is a short CFD actually a clean hedge?
No. That is the honest limitation worth sitting with before you open one. A short CFD is not a perfect hedge. It is a second leveraged position with its own margin call, its own stop loss and its own cost of carry, run on a different venue from the coins it is supposed to protect. It can take the edge off a pullback, but it cannot remove the risk, and a badly sized hedge can lose money on both legs in a choppy, sideways market where neither side moves far enough to offset the other.
## What would I actually do?
My read is that a short CFD earns its place as a hedge only when there is a specific reason for the timing: a known selling event, an EOFY date you want to defer past, or a technical level you genuinely expect to break. Not as a standing policy against every coin you own. Size it the way you would size any leveraged short. Risk a small, fixed percentage of account equity, set the stop above your entry before you open the trade, and treat the margin as capital at risk, not as a cost-free add-on. ASIC's negative balance protection means the worst case is losing your account balance, not more. A stop loss is what actually decides whether that hedge costs you a little or a lot.
Full breakdown of the 2:1 cap, the negative balance protection rule and the ASIC-regulated crypto CFD brokers available in Australia, on SatoshiMacro:
https://satoshimacro.com/guides/forex/how-to-short-bitcoin/?utm_source=binance_square&utm_medium=social&utm_campaign=autopilot_article
#SatoshiMacro #BitcoinHedge #CryptoCFD #Bitcoin

$BTC | Bitcoin's risk metric reads 0.37 today, or 37.2 out of 100, sitting in the below fair value band, the second lowest of five zones. That is nowhere near the extremes that have marked every prior Bitcoin cycle top or bottom across thirteen years of AUD price history.
## What Is Bitcoin's Risk Metric, Exactly?
It is a single number from 0 to 1 that tells you where Bitcoin's price sits relative to its own long run trend. Take the price, compare it to a logarithmic regression fitted across the whole history, measure the gap in standard deviations, clamp it to plus or minus three sigma, and map that onto a 0 to 1 scale. A reading near 0 means the price is deep below trend. A reading near 1 means it is deep above trend. On the desk we used to call this kind of thing a mean reversion gauge, and that is roughly the right mental model, except Bitcoin's mean is a moving target that itself grows exponentially.
SatoshiMacro's version of this indicator is AUD native, built from a January 2013 start on Bitcoin's own logarithmic regression, with a fitted power law exponent of about 5.41 on the AUD data. The same mathematics is published under a different name by physicist Giovanni Santostasi, who calls it the Power Law Oscillator. Different label, same residual. If you have seen either term thrown around on crypto Twitter, they are describing the same calculation.
## Where Does Bitcoin Sit Right Now?
The current reading is 0.37, which the model classifies as below fair value, the second lowest of five bands. For context, the 2024-2025 cycle's high point registered only 0.63 back in January 2025, well short of the 0.77 threshold that marks the cycle top zone. The most recent local low printed 0.31 in June 2026. None of this looks like either extreme of the historical range.
That gap between 0.63 and the 0.77 cycle top threshold matters. My read is that this cycle has simply not produced the kind of euphoric overshoot that 2013, 2017 and 2021 all did, at least not yet on this particular measure.
## What Do The Zone Boundaries Actually Mean?
The five bands are not arbitrary round numbers. They are percentile cut offs computed from the metric's own 166 monthly readings since January 2013, recalculated on every data refresh:
- 0.77 and above: cycle top zone, the top 7 percent of months
- 0.64 to 0.77: caution zone
- 0.43 to 0.64: near fair value, the most common band
- 0.29 to 0.43: below fair value, where the reading sits today
- Below 0.29: deep value, the bottom 10 percent of months
76 percent of all months on record have landed between 0.3 and 0.7. Extremes are, by construction, rare.
## How Have The Zones Called Past Cycle Tops And Bottoms?
The highest readings on record are November 2013 at 1.00, the only month to ever hit the clamp, followed by December 2017 at 0.89 and the 2021 first peak in March at 0.83. Each visit to the cycle top zone has come within months of a cycle high and been followed by a drawdown of 50 percent or more. The lowest readings cluster around the Mt Gox collapse in 2015 and the FTX collapse in December 2022, both inside or near the deep value band.
What the metric has never done is call a precise date. It flags a statistically rare zone, not a top tick or a bottom tick. Treat it as a forecast for the season, not tomorrow's temperature.
## How Would You Actually Use A Reading Like This?
The practical framework splits into two behaviours. Below 0.29, long horizon investors treat it as an accumulation regime, not a single buy signal, because visits to that band have historically lasted months rather than days. At or above 0.77, the same investors treat it as a distribution regime, often laddering exits rather than trying to sell the exact top.
For Australian holders specifically, the zone framework interacts with tax timing. If you are sitting on an unrealised gain and the reading climbs into caution or cycle top territory, the ATO's 50 percent capital gains discount for assets held past 12 months is worth checking before you act on the signal. A reading that says consider trimming and a tax position that says six weeks until the 12 month discount applies can pull in different directions, and the after tax math usually wins that argument.
At 0.37 today, neither behaviour applies. The honest answer to what you should do with this reading is nothing dramatic, because the metric itself is not flagging anything dramatic.
## Where Does The Metric Break Down?
It is a one factor model built entirely from price versus time. It says nothing about liquidity, regulation, halving cycles or ETF flows, all of which move Bitcoin independently of where it sits on a regression line. The bands are statistical associations drawn from three completed cycles, not a law of physics, and they will shift slightly as a fourth and fifth cycle add data. The regression also refits on every update, so a reading computed for a given month today is not identical to what it printed in real time back then.
That last point is the one I would flag hardest to anyone treating this as a trading signal rather than a positioning tool. It is built for multi year decisions, not entries this week.
Full methodology, the zone table and the live gauge sit on SatoshiMacro's page below.
https://satoshimacro.com/tools/crypto/cycle-indicators/bitcoin-risk-metric/?utm_source=binance_square&utm_medium=social&utm_campaign=autopilot_article
#SatoshiMacro #BitcoinRiskMetric #PowerLawOscillator #Bitcoin