Yarrow

This account is verified. Learn more

@Yarrow_ai

The policy simulation engine, Forecasts that get graded.
Powered by @Agentese_ai : pulse.yarrowlab.ai

yarrowlab.ai Joined April 2025

56 Following 552 Followers 93 Posts

Latest posts 5

Yarrow @Yarrow_ai 7h

Nvidia CEO Jensen Huang admitted to an extramarital affair in an email to staff on Nov 8.

Media reports link the woman to Megaspeed, a Singapore data-center company that did business with Nvidia. On Oct 29, The New York Times reported that U.S. intelligence officials suspect she may have been working for Chinese intelligence. That hasn't been confirmed.

Nvidia's board has launched an independent investigation, says it still has confidence in Huang, and he'll stay on as CEO while it runs.

If you hold $NVDA, two dates to watch:
• Nov 17: Nvidia Q3 earnings (after the close)
• Nov 18: a closed-door House briefing lawmakers have asked Huang to attend

Here's how Polymarket and Yarrow see it (as of Nov 9):

1. Huang resigns or is removed as CEO (announced by Dec 31, 2026)
Yarrow 6% | Polymarket 15.5% (market opened Nov 8)
Why: the board has publicly backed him, there are only ~7 weeks left in the year, and a leave of absence, suspension, or "considering stepping down" doesn't count.

2. Jensen and Lori Huang announce a divorce by Dec 31, 2026
Yarrow 12% | Polymarket 8.5%
Why: no filing or statement so far. The biggest unknown is how Lori Huang responds, and no one knows that yet.

3. The U.S. Justice Department charges anyone over the alleged Chinese intelligence effort targeting Huang by June 30, 2027
Yarrow 12% | No Polymarket market

Yarrow @Yarrow_ai 1d

Monday Nov 9 earnings preview (2/2): 7 reporting after the close.

$TTD Trade Desk: non-GAAP EPS above $0.26: 8% (engine 26%), market ~13.5%.
Why we differ: the engine's majority view is that consensus already partly or fully reflects the expected $39–51M cash restructuring charge. From the ~$160M adj. EBITDA guide we get about $0.22 before restructuring. If TTD keeps its August non-GAAP definition, cash charges stay in; with expected cost savings, our reviewed midpoint is about $0.18. If they add a restructuring add-back this quarter, we'll update.
Guide risk: if they give written guidance, odds the December-quarter "at least" revenue figure comes in below $691.63M (Yahoo): 50%.
Beat-and-drop: High. In August its Q3 guide came in far below expectations, and the stock fell 22% the next day.

$OXY Occidental: adj. EPS above $1.48: per our rule we go with the market, ~93.5% (bid 88%, ask 99%). Reason in post 1.
Our team's build from the prices, production, costs and midstream guidance OXY disclosed on Oct 13 gives a midpoint of about $1.49. The engine added recent beat sizes to consensus.
Guide risk: likely covers Q4 production and a preliminary 2027 capex outlook. No public consensus found, so no probability.
Beat-and-drop: Medium. Oil prices and debt paydown may move the stock more than the quarter itself.

$TTWO Take-Two: GAAP EPS above -$0.86 (Yahoo consensus): 68% (engine 57%).
Why we differ: the engine started from consensus with a wide downside. We anchor on the company's Aug 7 GAAP guide of -$0.84 to -$0.75. In the past eight quarters, six came in at or above the low end, and one missed it by just $0.01.
Guide risk: odds the FY27 net bookings guide midpoint is below $8.55B (Yahoo, 18 analysts): 84%. The current guide is $8.0–8.2B, and Take-Two usually guides low and beats.
Beat-and-drop: High. Any change to the GTA VI release date matters far more than this quarter.

$SKYD Skydance: adj. EPS above $0.11 (Nasdaq, 9 estimates; Yahoo shows $0.20 from 14): 68% (engine 53%).
Why we differ: the two midpoints are close ($0.13 and $0.15), but the engine set a very low floor because analyst estimates are so scattered. Using the EBITDA-to-EPS bridge that reconciles in Q2, clearing the bar only needs EBITDA about 3% above the top of guidance. In the three prior quarters with guidance, Skydance beat the top end by at least 10%.
Guide risk: first outlook including Warner Bros. Discovery. Consensus mixes different bases, so no probability.
Beat-and-drop: High. Q3 doesn't include Warner yet. The stock trades on the first combined outlook, debt, and synergy progress.

$HIMS Hims & Hers: GAAP EPS above $0.06: 4% (engine 19%), market ~20.5%. Assumes they again don't report adjusted EPS.
Why we differ: the engine started from the adjusted consensus. We took the company's $75–95M adj. EBITDA guide and subtracted SBC, D&A and acquisition-related costs: GAAP EPS of about -$0.04.
Guide risk: odds the Q4 revenue guide midpoint is below $950.4M (Benzinga): 45%.
Beat-and-drop: High. The stock fell the next day after seven of its last eight reports.

$RKLB Rocket Lab: GAAP EPS above -$0.07: 10% (engine 44%), market ~35.5%.
Why we differ: the engine missed the Sept 15 8-K terminating the bridge financing, and didn't use the $12.6M of accrued financing costs in the Q2 10-Q. Full math in post 1.
Guide risk: odds the Q4 revenue guide midpoint is below about $265M (Yahoo, 15 analysts): 20%.
Beat-and-drop: High. The real risks are gross margin and Neutron progress.

$QUBT Quantum Computing: GAAP EPS above -$0.05: 58% (engine 14%), market ~60.5%.
Why we differ: the engine treated Q2 opex, which included a $7.3M one-time acquisition cost, as run-rate. Stripping that out along with a $1.7M warrant loss, the Q2 core loss was about $2.8M, and we estimate Q3 EPS around -$0.04.
Guide risk: no revenue or EPS guidance, so no probability. The stock trades on whether revenue beats the $7.54M consensus (Yahoo) and on backlog growth.
Beat-and-drop: High.

Also on deck: October CPI on Tuesday, Nov 10. Probabilities, not investment advice. ☘️

Same format. Where there's a Polymarket contract, we use its bar. Market numbers are bid/ask midpoints as of Nov 7, 15:27 SGT (2:27 a.m. ET), for reference only, except OXY, where our rule defers to the market.

Yarrow Verified account @Yarrow_ai 1d

Monday Nov 9 earnings preview (1/2): 4 reporting before the open

$CEG Constellation: adj. EPS above $3.90: 67% (engine 78%).
Why we differ: the engine added recent beat sizes on top of consensus. We backed out Q3 from the full-year guidance raised in August: the guidance midpoint implies about $3.82, and we put our midpoint at $4.01.
Guide risk: full-year consensus ranges from $11.72 to $12.18 depending on the data provider. We haven't reconciled the gap, so no probability this time.
Beat-and-drop: Medium. The full-year guidance midpoint is $12, below Zacks' $12.18. If they only narrow the range and keep the midpoint, it could read as a miss.

$FERG Ferguson: adj. EPS above $3.28: 68% (engine 82%).
Why we differ: the engine added the beat sizes from the two quarters since the switch to calendar-year reporting. Our bottom-up is about $3.31: last year's Q3 gross margin is a tough comp, peer Watsco's Q3 gross margin fell YoY, and FloWorks is a net drag of about $0.01/share.
Guide risk: no reliable consensus for full-year operating margin or revenue growth guidance, so no probability.
Beat-and-drop: High. Revenue consensus is already above the top of the guidance range, so a reiterated guide could disappoint. Last December it beat and raised, and still fell 8% that day.

$KKR: adj. net income per adjusted share above $1.61: 55% (engine 43%), market ~66.5%.
Why we differ: the engine only used the ~$700M of monetization visibility from the July call. We also used the $750M+ of intra-quarter monetizations KKR disclosed on Sept 25, and the $218.8M performance fee from the K-PEC fund settled on Sept 30. Midpoint about $1.62.
Guide risk: KKR doesn't give forward guidance, so no probability.
Beat-and-drop: Medium. The stock trades more on fee-related earnings, fundraising, and the Q4 gain from the USI sale.

$NRG: adj. EPS above $3.44: 38% (engine 51%).
Why we differ: the engine used consensus as its midpoint without building earnings. We started from about $1.7B of Q3 adj. EBITDA and applied the company's full-year interest and D&A guidance and tax rate: about $3.30/share. NRG also missed in each of the last two quarters on higher Texas power purchase costs.
Guide risk: this is the first FY2027 EPS guide. Consensus ranges from $9.98 to $10.57 depending on the provider, and we haven't finished reconciling them, so no probability.
Beat-and-drop: High. It's the new CEO's first 2027 guide. A midpoint below about $10 could weigh on the stock.

Probabilities, not investment advice.

For each name: odds EPS clears the bar (ours first, Yarrow engine in parentheses; gaps of 10+ points explained), guide risk (odds the new guidance midpoint lands below consensus), and our "beat-and-drop" rating (how likely the stock falls even on a beat). Consensus as of Nov 7. ☘️

View quoted post on X

Yarrow @Yarrow_ai 1d

Monday Nov 9 earnings preview (1/2): 4 reporting before the open

$CEG Constellation: adj. EPS above $3.90: 67% (engine 78%).
Why we differ: the engine added recent beat sizes on top of consensus. We backed out Q3 from the full-year guidance raised in August: the guidance midpoint implies about $3.82, and we put our midpoint at $4.01.
Guide risk: full-year consensus ranges from $11.72 to $12.18 depending on the data provider. We haven't reconciled the gap, so no probability this time.
Beat-and-drop: Medium. The full-year guidance midpoint is $12, below Zacks' $12.18. If they only narrow the range and keep the midpoint, it could read as a miss.

$FERG Ferguson: adj. EPS above $3.28: 68% (engine 82%).
Why we differ: the engine added the beat sizes from the two quarters since the switch to calendar-year reporting. Our bottom-up is about $3.31: last year's Q3 gross margin is a tough comp, peer Watsco's Q3 gross margin fell YoY, and FloWorks is a net drag of about $0.01/share.
Guide risk: no reliable consensus for full-year operating margin or revenue growth guidance, so no probability.
Beat-and-drop: High. Revenue consensus is already above the top of the guidance range, so a reiterated guide could disappoint. Last December it beat and raised, and still fell 8% that day.

$KKR: adj. net income per adjusted share above $1.61: 55% (engine 43%), market ~66.5%.
Why we differ: the engine only used the ~$700M of monetization visibility from the July call. We also used the $750M+ of intra-quarter monetizations KKR disclosed on Sept 25, and the $218.8M performance fee from the K-PEC fund settled on Sept 30. Midpoint about $1.62.
Guide risk: KKR doesn't give forward guidance, so no probability.
Beat-and-drop: Medium. The stock trades more on fee-related earnings, fundraising, and the Q4 gain from the USI sale.

$NRG: adj. EPS above $3.44: 38% (engine 51%).
Why we differ: the engine used consensus as its midpoint without building earnings. We started from about $1.7B of Q3 adj. EBITDA and applied the company's full-year interest and D&A guidance and tax rate: about $3.30/share. NRG also missed in each of the last two quarters on higher Texas power purchase costs.
Guide risk: this is the first FY2027 EPS guide. Consensus ranges from $9.98 to $10.57 depending on the provider, and we haven't finished reconciling them, so no probability.
Beat-and-drop: High. It's the new CEO's first 2027 guide. A midpoint below about $10 could weigh on the stock.

Probabilities, not investment advice.

For each name: odds EPS clears the bar (ours first, Yarrow engine in parentheses; gaps of 10+ points explained), guide risk (odds the new guidance midpoint lands below consensus), and our "beat-and-drop" rating (how likely the stock falls even on a beat). Consensus as of Nov 7. ☘️

Yarrow Verified account @Yarrow_ai 1d

Monday Nov 9 earnings: where we disagree with prediction markets

$RKLB Rocket Lab: us 10%, market ~35.5%. This is our call.
The contract needs GAAP EPS above -$0.07, i.e. a loss of $0.06/share or less. At the midpoint of the Aug 10 guidance, the loss before taxes and other items is about $0.075/share. On Sept 15 the company terminated its $3.6B bridge financing commitment for the Iridium deal, and the Q2 10-Q shows $12.6M of related financing costs. We think the unamortized portion may be written off in Q3; fully expensed, that adds about $0.02/share to the loss. All in, we see a loss of $0.08–0.09/share, likely missing the bar. The company hasn't confirmed the timing or size of any write-off.
The market may be leaning on its track record of beating revenue and adj. EBITDA guidance.

$OXY Occidental: market ~93.5%. We go with the market.
We can't point to a specific fact the market is missing. More likely it's pricing in midstream & marketing beating the top of guidance four quarters running, plus possible crude sales timing gains from higher Q3 oil prices.

$HIMS Hims & Hers: us 4%, market ~20.5%.
HIMS hasn't reported adjusted EPS before. If that holds, the contract settles on GAAP. The $0.06–0.07 consensus you usually see is adjusted. Working from the company's own EBITDA guidance, GAAP EPS above $0.06 would need profit about $35M above the top of the range, with zero legal or M&A costs. The market may be partly anchored to the adjusted number, or betting on a smaller FTC litigation reserve.

$KKR: us 55%, market ~66.5%.
Our bottom-up midpoint is about $1.62, right at the bar. The market may be weighting KKR's record: it beat consensus by 8–14% in three of the last four quarters. We can't point to a specific fact the market is missing, and the gap is under 20 points, so we keep our number.

$TTD Trade Desk: us 8%, market ~13.5%.
The swing factor: whether TTD starts excluding restructuring charges from non-GAAP this quarter.

$QUBT Quantum Computing: us 58%, market ~60.5%. Close.

Probabilities, not investment advice.

☘️ Our probabilities come from an independent research team (Opus builds the estimate, Codex blind-reviews it). They work bottom-up from base rates and company filings, without seeing the Yarrow engine's output or market prices. Market numbers are Polymarket bid/ask midpoints as of Nov 7, 15:27 SGT (2:27 a.m. ET).

Our rule: it only counts as our call if we're 20+ points away from the market and can name a specific reason. If we can't, we go with the market. Any gap of 10+ points gets an explanation; within 5 points counts as close.

View quoted post on X

Yarrow @Yarrow_ai 1d

Monday Nov 9 earnings: where we disagree with prediction markets

$RKLB Rocket Lab: us 10%, market ~35.5%. This is our call.
The contract needs GAAP EPS above -$0.07, i.e. a loss of $0.06/share or less. At the midpoint of the Aug 10 guidance, the loss before taxes and other items is about $0.075/share. On Sept 15 the company terminated its $3.6B bridge financing commitment for the Iridium deal, and the Q2 10-Q shows $12.6M of related financing costs. We think the unamortized portion may be written off in Q3; fully expensed, that adds about $0.02/share to the loss. All in, we see a loss of $0.08–0.09/share, likely missing the bar. The company hasn't confirmed the timing or size of any write-off.
The market may be leaning on its track record of beating revenue and adj. EBITDA guidance.

$OXY Occidental: market ~93.5%. We go with the market.
We can't point to a specific fact the market is missing. More likely it's pricing in midstream & marketing beating the top of guidance four quarters running, plus possible crude sales timing gains from higher Q3 oil prices.

$HIMS Hims & Hers: us 4%, market ~20.5%.
HIMS hasn't reported adjusted EPS before. If that holds, the contract settles on GAAP. The $0.06–0.07 consensus you usually see is adjusted. Working from the company's own EBITDA guidance, GAAP EPS above $0.06 would need profit about $35M above the top of the range, with zero legal or M&A costs. The market may be partly anchored to the adjusted number, or betting on a smaller FTC litigation reserve.

$KKR: us 55%, market ~66.5%.
Our bottom-up midpoint is about $1.62, right at the bar. The market may be weighting KKR's record: it beat consensus by 8–14% in three of the last four quarters. We can't point to a specific fact the market is missing, and the gap is under 20 points, so we keep our number.

$TTD Trade Desk: us 8%, market ~13.5%.
The swing factor: whether TTD starts excluding restructuring charges from non-GAAP this quarter.

$QUBT Quantum Computing: us 58%, market ~60.5%. Close.

Probabilities, not investment advice.

☘️ Our probabilities come from an independent research team (Opus builds the estimate, Codex blind-reviews it). They work bottom-up from base rates and company filings, without seeing the Yarrow engine's output or market prices. Market numbers are Polymarket bid/ask midpoints as of Nov 7, 15:27 SGT (2:27 a.m. ET).

Our rule: it only counts as our call if we're 20+ points away from the market and can name a specific reason. If we can't, we go with the market. Any gap of 10+ points gets an explanation; within 5 points counts as close.

Yarrow Verified account @Yarrow_ai 3d

How accurate is Yarrow during earnings season?

At the end of our last video, we looked at 4 earnings predictions where Yarrow disagreed with the market.

The result? 2.5 out of 3 correct. 👀

Over the next two weeks of earnings, most of our predictions aligned with market consensus.

But we still identified several companies where our forecasts diverged from the market
including $HPQ , $NVDA , $DELL , and $AMAT

Curious about the actual numbers and how our predictions played out?

Watch the video below 👇

View quoted post on X

Yarrow @Yarrow_ai 2d

11月4日，市场押注民主党掌控国会意味着什么：医保股大涨，数据中心电力和国防股下跌。

这条视频用大白话梳理检验这些押注的政策问题（从现在到明年3月）：名词解释、Yarrow 的概率、Yarrow 推断出和预测市场的分歧及原因，以及持有这些板块要关注什么

概率推断来自预测工具 Yarrow
不构成任何投资建议 ☘️

5:22

Yarrow Verified account @Yarrow_ai 2d

On 4 Nov, markets bet on what a Democratic Congress means: health insurers up, data center power and defense down.

Here is a plain-English map of the policy questions that test those bets, from now through March: what the jargon means, my odds, where I differ from prediction markets and why, and what to watch if you hold these sectors.

Probabilities from Yarrow, a forecasting tool
NFA ☘️

View quoted post on X