PL

Semis like to roll lower from Sep-Oct into year-end: solana:5XqMbyZbtse5UQUzgrYMpeWLWLR6BWFqUm5H3VAWpump 1995 its 2Oct break below 20wk EMA, 13Nov 1995 clear breakdown .. 11Dec 52wk MA breakdown confo. initial bottom wk of 18Dec (only mini), then first multi week bottom wk of 15 Jan 1996. solana:5XqMbyZbtse5UQUzgrYMpeWLWLR6BWFqUm5H3VAWpump 1997 13Oct, 1st bottom 15Dec 1997, slighty undercut 1by wk of 2Jan '98 low, which kick started 1st proper multi week rally. solana:5XqMbyZbtse5UQUzgrYMpeWLWLR6BWFqUm5H3VAWpump 2000 dotcom: wk of 25 Sep 2000 clear breakdown. 1st choppy volatile bottom 27Nov 2000 tricky to trade. 1st clear rally started wk of 2Jan 2001 to wk of 29Jan. solana:5XqMbyZbtse5UQUzgrYMpeWLWLR6BWFqUm5H3VAWpump 2018 clear break wk of 1st Oct '18, 1st low 29Oct, final low 24Dec'18, clear turn higher 1st wk of Jan 2019 -> huge rally to new highs into Apr.

if you take the current episode, advance over 60-75wks of this kind of magnitude 150%+, we only have 3 priors in solana:5XqMbyZbtse5UQUzgrYMpeWLWLR6BWFqUm5H3VAWpump 1995 +181% over 69w take several significant swing lows at the top and draw a line. break of that line ie clear range break first leg down -36% over 15wks from the breakdown level. 1997 (also started in Oct) +192% over 57w led to -34% decline to the initial low which held for at least 5 weeks 2000 dotcom +640% over 75w led to -40% decline over 9 weeks to the initial multi week low 2026 if the clear breakdown level on the solana:5XqMbyZbtse5UQUzgrYMpeWLWLR6BWFqUm5H3VAWpump is 10900 that would put us at wk 2 currently and down 3% or so, while still holding the rising 52wk MA and hanging around the prior significant swing low at 10445 (Leopold liquidation low)

If you think pure mathematics is too complex, you don’t appreciate how simple it is when compared to the world it is attempting to model. For every question, including trading edge there exists an optimal level of “zoom” or scale, at which you are able to best describe the phenomena in question to determine the most precise level of order or signal from Noise. The RIGHT level of zoom matters: - too fine grained and your signal becomes noisy due to detail which actually gets smudged away at the scale at which you’re operating. Too much detail takes away. It can leave your strategy overfitted to noise (esp if you are working with sparse samples) and lead to instability in performance out of sample. - too coarse grained and your signal is just too smudged to deliver meaningfully high fidelity “edge”. Reductionism is a useful step to remove elements or features which are not meaningfully contributing to the signal. The goal is to isolate to stable and robust principles, and work out their deviations from long term baseline. Are the deviations predictable and can they be made stable relative to some solid first principles assumptions? Then you end up with a robust edge that’s likely to perform well out of sample. Every situation in the market has an “input” from descriptions at varying levels of “zoom”. Confluence from those descriptions can predict higher edge. You cannot reverse engineer the edge into first principles because datasets you’re working with are too sparse at medium frequency to just number crunch everything and “prove”. You can only start from good “taste”, guesses rooted in solid first principles and stack observables at varying levels of zoom and from different aspects to get to an edge that incredibly high grade.

$QQQ 2018 analogue was fundamentally different to now, but timing and occurence of midterms was coincident. here's a brief revisit of the action at the time. 4 Oct start -> 24Dec bottom, -23% deep initially US10yr yields broke higher, but in the final leg in Dec yields nosedived as stocks collapsed in the final leg over worries re weaker growth and dovish Fed. AAPL missed guidance 1-2Nov, Texas Instruments and semis warned of slowing demand. Main flavour was market concerned that Nasdaq stock valuations had become too high and vulnerable, US-China trade tensions and some soft semis reports helped confirm, and Fed was tightening and market was worried it would continue doing so into slow down. Market bottomed ahead of any positive news into year-end, around 24Dec: 19 Dec Fed hiked and Powell gave a very hawkish press conf which triggered a 4 day parabolic extension to the downside, 21 Dec Fri was Dec OPEX last of the year. next day market made the low. 4 Jan 2019 Powell indicated Fed would be dovish on rates and balance sheet policy, which helped the market trigger a "wedge pop" thru 20day EMA

Stock market is a not great predictor of recessions, contrary to popular opinion, but a great predictor of economic bottoms. Re the former it becomes more obvious when you consider that everyone has called 20 of the last 2 recessions at the very least. Sentiment and positioning drive short term corrections. Sometimes they may happen ahead of the economy going into a recession and more often than that they do not. Earnings growth deceleration + clear decline has led both the stock market and the economy during the last major recessions dating back to dotcom. When Profits and profit margins are declining, a recession began within the next year about 27% of the time, compared to 5% when margins were rising. Profit margins are currently rising even backing out contribution from private companies holdings in megacaps 10Q-s, roughly at 14.5%. Credits @EPBResearch.

Commanding rejection of the rally at the 20day EMA in $QQQ as Friday gap down followed is probably the most significant takeaway from the start of October downtrend.

To caveat, Stock market decline and recession are independent variables. Stock market can have deep corrections without a recession and huge rallies within a bear market. My tweet was just addressing the claim “stock market is predicting a recession here” is false in general based on statistical evidence at the very least, and is contradicted on this occasion by economic indicators that predict the recession better than the stock market.

Profit margins still rising

As you start to walk on the way, the way appears - Rumi.