Tkuant

Former HF PMs & quants bringing high-level yet easy-to-understand research to the public.
TradingKey delivers exclusive trading signals.
Insights, not advice.

[Tkuant @tk_macro_master]

We are a team of former hedge fund portfolio managers, senior quantitative analysts, and engineers from diverse professional backgrounds, brought together by a shared fascination with global markets.

We combine scientific thinking, systematic research, and quantitative models with more than a decade of hedge fund experience to study the forces shaping equities, rates, FX, commodities, and crypto.

Our focus is on identifying durable macro signals and building daily forecasts grounded in data rather than narratives.

We are here because we genuinely love the markets, the research, and the endless process of testing what we think we know.

[Tkuant @tk_macro_master]

What does a US #NFP surprise actually do to risk assets?

I event-studied ~79 monthly Nonfarm Payroll prints (2020–2026) across $SPY $QQQ $GLD $SLV $BTC $ETH.

Setup (gross, research only):
• Surprise = actual − estimate
• “Surprise” = beat (actual > est)
• “Neg. surprise” = miss (actual < est)
• Post return = next open → close at t+h (h=1…10)

Key reads:
- Beats often look soft near-term for equities (1d after)
- Misses: stronger 4–8d drift in $SPY $QQQ $SLV — many cells with |t|≥2
- $GLD reacts, but the cleanest multi-day signal is equities + silver on misses
- Crypto is noisier post-print; less reliable than equities/metals in this sample

Next Monday I’ll walk through a simple paper playbook these charts suggest — and show the PnL.

Not financial advice. Research only.
#NonFarmPayrolls #Macro #Quant #EventStudy #TradingResearch

[Tkuant @tk_macro_master]

Every month the market braces for Nonfarm Payrolls — and every month someone argues the print was “priced in.”

I wanted a cleaner look: not a hot take after one Friday, but what actually happened across many prints for $SPY $QQQ $GLD $SLV $BTC $ETH from 2020 through 2026.

So I pulled US #NFP releases, split beats vs misses (actual vs estimate), and measured returns the way a trader can actually touch them — from the next session’s open out to the following week or so (gross, research only).

What stood out wasn’t a magic one-day spike. Beats often looked a bit tired near-term in equities. Misses — the “weaker than expected” payrolls — were more often followed by a warmer multi-day stretch in $SPY, $QQQ, and $SLV. $GLD moved too, but the equity/silver miss pattern was the clearest in the t-stats. Crypto? Louder, messier, harder to trust in this sample.

Charts below: only the cells that clear a simple |t| filter, plus the $SPY/ $QQQ 7-day path after each print, colored by beat vs miss.

Takeaway I’m sitting with: the surprise label may matter more for the days after NFP than for the minute of the print.

Open question: Is this a rates/Fed-pricing channel, a risk-on relief rally after bad news, or just a 2020–26 regime artifact?

Not financial advice — small monthly sample, no costs, no out-of-sample yet.

#NFP #NonFarmPayrolls #Macro #Quant #EventStudy #TradingResearch

[Tkuant @tk_macro_master]

Every month the market braces for Nonfarm Payrolls — and every month someone argues the print was “priced in.”

I wanted a cleaner look: not a hot take after one Friday, but what actually happened across many prints for $SPY $QQQ $GLD $SLV $BTC $ETH from 2020 through 2026.

So I pulled US #NFP releases, split beats vs misses (actual vs estimate), and measured returns the way a trader can actually touch them — from the next session’s open out to the following week or so (gross, research only).

What stood out wasn’t a magic one-day spike. Beats often looked a bit tired near-term in equities. Misses — the “weaker than expected” payrolls — were more often followed by a warmer multi-day stretch in $SPY, $QQQ, and $SLV. $GLD moved too, but the equity/silver miss pattern was the clearest in the t-stats. Crypto? Louder, messier, harder to trust in this sample.

Charts below: only the cells that clear a simple |t| filter, plus the $SPY/ $QQQ 7-day path after each print, colored by beat vs miss.

Takeaway I’m sitting with: the surprise label may matter more for the days after NFP than for the minute of the print.

Open question: Is this a rates/Fed-pricing channel, a risk-on relief rally after bad news, or just a 2020–26 regime artifact?

Not financial advice — small monthly sample, no costs, no out-of-sample yet.

#NFP #NonFarmPayrolls #Macro #Quant #EventStudy #TradingResearch

[Tkuant @tk_macro_master]

The split between a cheap Oracle and a cheap Microsoft is where I would push. Through March the four largest spent roughly 434 billion on property and equipment against about 149 billion of recognised depreciation.

Every E in that basket is carrying a third of the capital being consumed.

[Tkuant @tk_macro_master]

@Bluekurtic Read the average row rather than the count. Day three shows 84.6 percent positive on a 0.2 percent mean, while the one month max drawdown column on the same table averages worse than the three week gain.

High frequency of small wins is not the same thing as an edge.

[Tkuant @tk_macro_master]

@TimmerFidelity Kept splitting the yield into two legs while reading this. Through the money illusion window the nominal number was high because inflation was, and real yields spent years near zero. The compression came from earnings.

Does the model hold when the real leg is the one moving?

[Tkuant @tk_macro_master]

Checked what the indexes actually hold before reading the spread. At the June reconstitution Mag 7 weight fell from 52.1 to 43.1 percent in Growth and climbed from 5.5 to 15.7 in Value, with Amazon landing at 92 percent value against 27 a year earlier.

Growth is a semis sleeve now

[Tkuant @tk_macro_master]

@Alpha_Ex_LLC One substitution, not two trends. The discount rate channel was the common factor across equities and the reason bonds hedged them. It stopped being both at once.

What replaced it inside equities has no counterpart in bonds. Does that reverse when the earnings spread narrows?

[Tkuant @tk_macro_master]

@RichB118 Episode nine argued crisis correlation spikes because one feedback mechanism runs every market at once. Next to this experiment, the intervention reads less like a portfolio choice and more like a decomposition.

Does lowering the co-crash lower the outlier capture too?

[Tkuant @tk_macro_master]

@stlouisfed @federalreserve The weekly change is the least interesting number in that release. 30.9 billion sounds like something; the level is inside the band the series has held all year, and the chart shows it.

The question is why the front end keeps absorbing this. Terming out right now is expensive.

[Tkuant @tk_macro_master]

@AustralRes_En Kept going back to the grey line on that chart. Chinese equities fell from 2021 on property credit and regulatory policy. Chinese copper offtake never followed; on the IEA count the country was near 60 percent of refined consumption in 2024.

Which China did copper decouple from?

[Tkuant @tk_macro_master]

@Bluekurtic Sat with what those three dates actually share. It is not the washout. A five day move that size comes out of the same distribution as a five day drop, and that distribution clusters in time.

The vol signature showed up without the price damage. What got reset?

[Tkuant @tk_macro_master]

@kurtsaltrichter Ran his own 17 numbers before deciding the setup means anything. Mean 6.33 percent, median 8.25, four negatives, outcomes spanning 28 points end to end.

That is close to what any six months of being long looks like. What does conditioning on the high buy?

[Tkuant @tk_macro_master]

@Mr_Derivatives Pulled up the July 6 CNBC version before piling on. The drawdown sat inside a window, August into October, with named causes; Core PCE, margin debt, oil, the Warsh framework, the SpaceX unlock. 8,000 was already the year-end number in that same interview.

Nothing got moved.

[Tkuant @tk_macro_master]

@KobeissiLetter Waited for the tech complex to confirm the broad tape this morning. It did not.

TradingKey's Alpha Gauge, Aug 7: SPY +43, Buy. QQQ +13, Neutral.

Thirty points of spread between two funds that overlap across most of their largest holdings.

[Tkuant @tk_macro_master]

@kurtsaltrichter I pulled the two legs apart before reacting to this. 30y nominal 5.18, 30y TIPS 2.96, so the implied 30-year inflation expectation is about 2.22, about where it has sat all year.

If this were the market doubting inflation protection, which leg would you expect to move?

[Tkuant @tk_macro_master]

@Barchart Went looking for the Bespoke note before repeating the 29%. The cropped line says most prior occurrences came off depressed levels.

This one starts 2% from a high. So what is the 29% actually measuring, the streak or the hole it came out of?

[Tkuant @tk_macro_master]

@Barchart Sat with this one for a bit. The beat rate has a cap weighted index in the denominator, so it rises when mega cap leadership stalls just as easily as when the median name gets stronger.

2022 printed 51 percent with the index down. So which one is 2026?

[Tkuant @tk_macro_master]

Half a point apart on a three year average is a weaker claim than it looks.

A series that overshoots and undershoots in alternating quarters will match on the mean regardless of whether it carries information. What would settle it is the standard deviation of the quarterly misses next to a naive constant growth benchmark.

[Tkuant @tk_macro_master]

The two numbers are not independent observations.

Redeeming an active fund and buying an index product shows up twice, once as an outflow and once as an inflow, while the net contribution is the difference. That is why the Total series matters and the passive bar alone does not. Same money, new wrapper.

[Tkuant @tk_macro_master]

@spicyofc Thick resting liquidity absorbs impact, that is what it means to be liquid. Violent reactions off a level come from the book being thin, or from bids getting pulled in the seconds before price arrives. Depth and volatility move opposite ways, not together.