Hold existing positions; do not add above 195 until NCO stabilization is confirmed; stop at 205

• 0.19 PEG ratio is unreliable — $24 forward EPS built on flat credit assumptions, with 90+ day delinquencies nearly doubled since 2022
• Forward EPS revisions historically come down 20-40% in credit deterioration, which could shift PEG from 0.19 to 0.40
• 2022-2023 vintages will continue migrating into charge-offs over next 2-3 quarters regardless of new origination quality
• New delinquencies being flat only stabilizes the front of the pipe — middle of the pipe is already stressed and rolling forward
• NCOs lag delinquencies by roughly two quarters; full flow-through not yet visible in reported numbers
• Lower Bollinger Band in credit-sensitive names is not a floor — often breaks at halfway point of move
• Actual downside if Q3 NCOs disappoint is re-test of March lows near 173, an 18% drawdown from current levels
• 200 SMA at 205 is a round-number magnet that fails in stress regimes, not a structural floor
• $28.5 billion goodwill equals 25% of stockholders' equity — partial or full write-down becomes almost inevitable in stress scenario
• Buybacks are discretionary and were suspended in 2008, 2009, and 2020 during stress events — capital return yield is a callable feature, not a floor
• Stablecoin consortium launch is H1 2027, more than 9 months away with uncertain revenue contribution
• Asset sensitivity (higher-for-longer tailwind) is the same mechanism causing consumer stress in the cardholder book
• AML and debanking legal saga is escalating with Senate involvement and active Trump team pushback
• Zero retail engagement and no contrarian community thesis — entirely dependent on institutional flows for support

→ vs aggressive: 0.19 PEG ratio assumes $24 forward EPS will hold — when credit deteriorates, forward EPS revisions historically come down 20-40%, which would push PEG to 0.40 and eliminate the apparent value

→ vs aggressive: New delinquencies being flat is not an all-clear; 2022-2023 vintages will continue rolling into 90-plus days for 2-3 quarters regardless of new origination quality

→ vs aggressive: Lower Bollinger Band is not a buy signal in credit-sensitive names — it marks the halfway point of a move, not the bottom; 200 SMA at 205 is not a structural floor

→ vs aggressive: $28.5 billion goodwill (25% of equity) is not 'fine on paper' — impairments wipe reported equity, trigger capital reviews, and signal management overpaid for Discover

→ vs aggressive: 11% capital return yield is not a floor — buybacks are discretionary and were suspended in every prior stress episode (2008, 2009, 2020); the mechanism disappears when needed most

→ vs aggressive: Stablecoin consortium is H1 2027, more than 9 months out with uncertain revenue — this is 2027 bull case material, not justification for adding exposure at 211 today

→ vs aggressive: Asset sensitivity being a higher-for-longer tailwind is the same mechanism causing the credit deterioration in the 90-plus day bucket — these are linked, not separable

→ vs neutral: Tighten the stop below 200 SMA to 205, but do not add above 195 until at least one quarter of confirmed NCO stabilization prints

HOLD with conditional add/trim parameters: mental stop at $200, lower bound for adds at $195 only on capitulation volume, measured adds above $218 with credit confirmation, and trim positions into rallies at $218-$222.

• 0.19 PEG ratio is analytically sloppy because the earnings denominator is the contested variable — if Q3 NCOs print higher than expected, EPS estimates could fall 15-25%, pushing PEG from 0.19 to 0.27-0.35
• Lower Bollinger Band argument is mechanical thinking unsuited for a fundamentally-driven credit name — COF is in a shifting distribution, not a stationary process
• "Risk is already in price" argument is unfalsifiable and contradictory — cannot claim market is mispricing COF while trusting it to have correctly priced in credit risk
• 11 percent capital return yield is overstated as a floor — buybacks are the most discretionary capital allocation item and get cut in stress scenarios
• 18 percent drawdown to $173 prices worst-case vintage roll-forward as base case — new delinquencies are flat, book has been re-underwritten for two years, Q2 2026 was strongest revenue quarter in company history
• "Do not add above $195" is overly restrictive — a clean Q3 credit print could push COF back to $225-$230 within weeks
• Goodwill impairment is a multi-year tail risk tied to sustained underperformance, not a near-term catalyst — Q2 2026 showed record revenue of $15.85 billion, NII of $12.37 billion, and net income of $3.02 billion
• AML/debanking concerns add volatility premium but not structural risk — headline noise has circulated for months without affecting fundamentals or buyback program
• Mental stop at $200 on closing basis below 200 SMA on volume changes the technical structure meaningfully
• Discipline is not paralysis — meaningful difference between waiting for confirmation versus waiting for capitulation

→ vs aggressive: PEG ratio argument is analytically sloppy because the earnings denominator is the contested variable subject to revision based on Q3 NCO results

→ vs aggressive: Lower Bollinger Band is mechanical thinking applied to a fundamentally-driven credit name in a shifting distribution, not a stationary process

→ vs aggressive: "Risk already in price" argument is unfalsifiable and contradictory — cannot claim market misprices COF while trusting it to have correctly priced credit risk

→ vs aggressive: 11 percent capital return yield is overstated as a floor — buybacks are the most discretionary capital allocation item and get pulled when CECL provisions build

→ vs conservative: 18 percent drawdown to $173 requires credit deterioration not confirmed in NCO data — worst-case vintage roll-forward is being priced as base case

→ vs conservative: "Do not add above $195" is overly restrictive — clean Q3 credit print could produce 10-15 percent upside to $225-$230 within weeks

→ vs conservative: Goodwill impairment is a multi-year tail risk, not near-term catalyst — combined entity is exceeding acquisition underwriting case

→ vs conservative: AML/debanking is headline noise adding volatility premium, not structural fundamental risk