Michael Corley, MBA, LSSBB

How do you estimate 1-in-200-year events when you barely have 100 years of data? Hugh Miller's latest Normal Deviance column tackles this question with a statistical experiment. Using 100 years of portfolio loss data, the 95% range of estimates for the 99.5th percentile spans $1.03m to $6.18m — one actuary's estimate could plausibly be six times another's based purely on statistical noise. Regulations require estimates that carry enormous uncertainty. Part of an actuary’s skillset is communicating uncertainty about our uncertainty — and why no one else is better positioned to do it. Read Hugh's analysis: https://lnkd.in/gusi7vMz #ActuariesInstitute #DataScience #RiskManagement

21

This chart explains Structure/Correction Trading, showing how price action transitions between higher highs (HH), higher lows (HL), lower highs (LH), and lower lows (LL)—all within demand and supply zones. It highlights how traders can spot market shifts early by recognizing break-of-structure and failed highs/lows. More Technical Analysis Here: https://t.me/fxsignalgroup #ForexTrading #USDJPY #GBPUSD #TechnicalAnalysis #ForexSignals #PriceAction #TradingStrategy #DayTrading #SwingTrading #ForexForecast #LiquidityGrab #SmartMoney #ChartAnalysis #ForexMarket #TradeSetup #ForexEducation #MarketUpdate

1

The quality of economic data coming out of Washington--especially BLS--is deteriorating at an astonishing pace. Just the most recent news: We found out earlier this week they are basically guessing at inflation, and now a massive 90% downward revision to jobs from 147k to 14k. In other words, inflation is being under-estimated and jobs are being over-estimated. Are we headed toward a "stealth stagflation?" #inflation #interestrates #federalreserve #economy #stagflation

21

8 Comments

Per Arbor, most-watched supply chain metrics have been rolling over and signaling less stress, evidenced by aggregate index (black line) moving back towards 0 

19

Synthetic Energy Datasets with Gemini / Claude: Building Better Forecast Models Without Customer Data Energy companies today face a paradox: they need massive datasets to train forecasting models, yet real customer load data is tightly protected under privacy regulations. That’s where synthetic energy data comes in. By combining generative models like Gemini 1.5 Pro and Claude 3.5 Sonnet, utilities and researchers can now create realistic, statistically accurate energy traces without touching a single line of personal data. https://lnkd.in/gPg9VeNn

2