When: RSI sustained >40

Then: Re-enter 40-50% tranche

REDUCE 15% (not 50%); fix stop loss to 83.80; require multi-signal confirmation for re-entry; revisit thesis post-Q3 earnings

• Entry at 85 with stop loss at 90 creates inverted risk-reward on a short position—stop triggers above entry, guaranteeing a forced loss on any bounce
• Proper stop loss should be 87.00-87.50 based on ATR of 1.31 and volatility conventions
• Dividend yield of 3.55% below 4.5% ten-year Treasury is contextual, not catastrophic—SO has 63% total return over 5 years and 65 consecutive years of dividend growth
• Forward payout ratio is 77% (not 88% as aggressive analyst claims)
• Beta of 0.316 makes SO portfolio ballast, not a capital appreciation vehicle—selling defensive positions on 20bp rate moves destroys long-term portfolio construction
• RSI of 29.96 indicates selling exhaustion, not a momentum trap; path of least resistance shifts toward mean reversion
• Price held lower Bollinger band at 85.08, a setup where disciplined risk reducers take profits on shorts
• 15 million share volume spike (2.5x 30-day average of ~7 million) is a data point, not a regime change
• Polymarket 20pp jump in two-hike probability is reflexive and reversible; building 50% reduction on sentiment is top-calling
• Data center moratorium probability fell 15pp—a structural tailwind for SO via Georgia Power exposure
• Negative FCF during major capex cycle is standard operating model for regulated utilities; construction in progress up 55%
• PG&E at 9.5x earnings is not a valid comparable due to wildfire liability and regulatory compact questions absent in SO

→ vs aggressive: Stop loss at 90 on a sell entry at 85 is not risk management—it's a guaranteed loss trigger that forces buying back higher than sold

→ vs aggressive: Dividend yield of 3.55% framed as a 'millstone' ignores 63% five-year total return and 65 consecutive years of dividend growth—the dividend IS the reason to own this name

→ vs aggressive: Forward payout ratio is 77%, not 88% as claimed, making the dividend sustainability argument materially stronger

→ vs aggressive: RSI at 29.96 signals selling exhaustion and mean reversion setup, not a smoking gun for new shorting

→ vs aggressive: 15 million share volume spike is 2.5x average—not a regime change but a notable data point being overweighted as certainty

→ vs aggressive: Building a 50% position reduction on Polymarket sentiment is top-calling that reverses just as fast; same mechanism can push two-hike probability back to 40%

→ vs aggressive: Aggressive analyst ignores the 15pp drop in data center moratorium probability—a meaningful structural tailwind for SO

→ vs aggressive: PG&E comparison is invalid due to wildfire liability and regulatory overhang that SO does not face

→ vs aggressive: 50% trim 'panic math' sacrifices half the income stream on a 5-7% drawdown from August highs without evidence of fundamental regime change

Trim 25% of SO exposure at $85-$87; maintain dividend-sized core; set trailing stop at $83.80; do not initiate new short-side tactical exposure due to mean reversion risk; reassess after Q3 earnings

• Stop at $90 is fundamentally broken for a short position—should be $87-$87.50 (1.5-2x ATR where ATR is 1.31)
• 25% trim base case is appropriate; 50% trim escalation trigger is confirmed close below lower Bollinger band at $85.08 on expanding volume
• Re-entry requires 2 of 3 confirmations: close above $87.88 on volume PLUS (contracting MACD histogram OR bullish RSI divergence from sub-30)
• Aggressive Analyst valid on: yield inversion (3.55% vs 4.5% 10-year), negative FCF (~$4B annually), 4.5% annual share count creep, 55% YoY CIP increase, 9% debt increase, uniformly bearish technicals (price below 10 EMA, 50 SMA, 200 SMA; MACD below signal with expanding negative histogram; RSI near 30; VWMA falling), September 18 distribution session at 15M shares
• Conservative Analyst valid on: negative FCF is standard for regulated utility build cycle, SO earns above-allowed ROE at 11.48%, PG&E comparison is apples-to-oranges due to wildfire liability overhang, dividend Aristocrat status provides price floor
• Polymarket data is informative but not conclusive—can move back on next CPI print; prediction markets are not dot plot or Fed Funds futures
• AI data center moratorium probability falling 15 points is a structural positive for SO given Georgia Power data center exposure—a multi-year rate base growth driver
• Dividend is structural reason stock trades at premium to PG&E; trimming too aggressively destroys the feature that justified original allocation
• PEG ratio is elevated at 2.02; 5-year cumulative return is 63% (rich)
• RSI sub-30 plus support at lower Bollinger band creates asymmetric mean reversion risk against fresh short initiation
• Reassess after Q3 earnings in late October (approximately 4-5 weeks)

→ vs aggressive: Overstating macro certainty—Polymarket moving 20 percentage points in a week is reflexive sentiment, not regime fact; treating sentiment signal as established fact is analytical error

→ vs aggressive: Conveniently skipped the AI data center moratorium probability falling 15 points, which is a structural positive and multi-year rate base growth driver for SO

→ vs aggressive: Framing dividend as 'millstone' is emotional language inappropriate for rigorous analysis; dividend is the structural reason stock trades at premium to PG&E

→ vs aggressive: Inherited the same flawed $90 stop structure from the trader's plan without correcting it despite correctly identifying the bearish setup

→ vs conservative: 'One data point, not a regime change' is exactly what bag-holders said in 2007, 2000, and every other secular shift—dismissing convergence as 'yellow flag' rather than 'red flag' is asymmetric framing that lets analysts stay bullish into regime changes

→ vs conservative: 15% trim may be too small—if bear case plays out to lower band break at $85.08, a 15% trim leaves firm exposed to further 8-10% drawdown on remaining 85% of position; that's not protecting assets, that's hoping harder

→ vs conservative: Using PG&E at 9.5x earnings as de-rating comparable is misleading—PG&E carries wildfire liability overhang that SO doesn't