Altair
Skip to content

Trading Analysis Report: UNH

Analysis date: 2026-10-21 · Model: claude-opus-5-5 · AI-generated, not financial advice.

  • Analysis date: 2026-10-21
  • Rating: Hold
  • Generated: 2026-10-21 13:20:03
  • TradingAgents 0.6.0: anthropic, deep claude-opus-5-5, quick claude-sonnet-5-5
  • Analysts: market, sentiment, news, fundamentals; research debate rounds 5, risk debate rounds 5
  • Data vendors: core_stock_apis yfinance, technical_indicators yfinance, fundamental_data sec_edgar,yfinance, news_data yfinance, macro_data fred, prediction_markets polymarket

V. Portfolio Manager Decision

Rating: Hold

Executive Summary: Hold UNH. Keep existing exposure up to a standard weight, but do not start or add at the provisional $417.49 close: reward to $433.68-$453.36 is about 1:1, only 0.8-0.9R against the blended 383/363 exit, and all three analysts put fair value near $402. If you are above standard weight, trim about half the excess now and the rest into $422-433; if you are flat or underweight, place no resting order, confirm Q3 is actually reported and score it against pre-written tiers, then buy at most 25% of a standard allocation on a daily close that reclaims the live 10 EMA near $402, with a second 25% clip only after Nov 4 and total exposure capped at standard weight. Tactical shares cut half on a daily close below the live daily supertrend ($383.01 today), everything exits on a weekly close below $362.97, and sizing must survive a repeat of this year's -19.6% gap through the late-January Q4 print and 2027 guide.

Current Price: 417.49

Price Target: 402.0

Confidence: Medium

Time Horizon: 3 months

IV. Risk Management Team Decision

Aggressive Analyst

Aggressive Analyst: Let me open, since nobody else has spoken yet. The word "Hold" makes this plan sound passive, but it isn't. It tells you to keep your winners, refuse a bad entry, and have capital pointed at a much better one. That's how you press a bet when the odds are about to improve.

Start with what the tape says. The weekly trend is up, the 10, 50 and 200 averages are stacked, and ADX went from 14 to 42.6 in seven sessions. RSI went from 33 to 71 in under three weeks. That's a violent regime change, and it came after a stock that closed at 255.97 in March is now 63% higher. The people who waited for a clean, risk-free setup after the January collapse missed the entire move. I'm not here to tell you to wait for certainty. I'm here to tell you the plan already prices in the uncertainty and sets triggers around it.

Take the math on why not to buy at 417. Against the 383.01 daily stop, the risk is about 34.5 points. The reward to the 433.68 closing high is about 16, and to the 453.36 monthly stop it's about 36, so roughly 1:1. Now buy at 400 with the same stop. The risk is 17 points, and the reward to 433.68 is nearly 2R, to 453.36 about 3.1R, and to the 458.79 July high about 3.5R. Waiting 17 points turns a coin flip into a payoff worth taking at size. Buyers already defended that shelf twice, at 400.00 and 399.20, and the 10 EMA at 400.09 sits right there. If anything, the 10 EMA keeps rising, so that zone is a moving target. I'd rather get there early and be ready.

Now the objections I expect from the conservative side. They'll point to the TD-9, the +1.97 z-score, RSI near 70, and price against the upper Bollinger band. All of those are daily-timeframe pause flags. The weekly TD count is -3 and the monthly is -1, so the higher timeframes show no exhaustion. Weekly z is +0.97 and monthly is +1.12. With ADX at 42, price can walk the upper band for a while. And the TD-9 rests on a provisional close from a bar where the open is above the high, which is impossible. I wouldn't build a sell case on that.

They'll say OBV is below its late-September level. But the technical report itself says the absolute OBV level isn't meaningful and the slope is what counts, and the slope confirmed the thrust, from -55.9M to -34.3M. They'll say the monthly supertrend is still down. Yes, and its stop at 453.36 is an upside target, not a reason to be scared. A month-end close above it would flip the monthly regime, and the plan keeps standard-weight holders in place through that zone and only trims excess.

On fundamentals, they'll lean on revenue growth of 0.4% and margins that remain below 2024's. That's fair, but I'd flip it. Operating margin is 7.1% against 8.0 to 8.6% in 2024, and on roughly $450B of revenue each point of margin is about $4.5B of operating income, around $5 a share before interest and taxes. The thesis is a margin recovery, not a growth story, and cash flow supports it: H1 free cash flow was $18.4B, about 1.56x net income. At about 18x forward earnings and a TTM FCF yield near 7% (some of it working-capital timing, I'll grant), the market isn't paying a growth multiple for a growth company. RBC just went to $498, which is 5.8R from a 400 entry.

They'll also bring up the lawsuit, the Democratic House at 92%, the lack of insider buying, and the gap risk. The suit is a headline with no sizing, and the stock rallied 10% through it. A 92% House outcome is consensus, not news, and the real surprise risk is the Senate at 66/34. If anything, election uncertainty clears on November 4. Conway's three small sales at an average of about $383 aren't distribution. The gap risk is real, with a -19.6% day in January, but the plan sizes for it. Volatility is also where the +9.4% and +7.0% days came from, and you don't get those without accepting the others.

To the neutral analyst, who I expect to say hold and see what happens: that's the version of Hold I reject. Without triggers it's just waiting. This plan has an entry, a size ladder tied to implied Q4 EPS (50% at 5.00 or more with a stable medical care ratio, 25% at 4.50 to 5.00, nothing below 4.50), a half-cut at a daily close under 383.01 on rising volume, and a full exit on a weekly close under 362.97. One thing the neutral view underrates is the data gap. Q3 rests on one unverified social media source that shows revenue of $112.7B (-0.4%), operating margin of 5.8% (down from 7.1% in Q2), and guidance raised to at least $20.50. If that margin number is right, it's a sequential step down, and a skeptic should say so. But it's also far better than the 3.8% a year earlier, and the stock rallied on it. That's exactly why the lookup comes before the money.

Here's where I'd push the plan harder. If UNH never pulls back and instead grinds into 422, I don't want to be locked out. I'd add a confirmation rule: a daily close above 422.40, then 433.68, with expanding volume and OBV recovering toward -24M, earns the second half of the position. Then the stock either gives us the cheap entry or proves the breakout. Either way we're positioned, and nobody is chasing a stretched, partly corrupted print.

I don't know your current holdings, so apply it this way. If you own a standard weight, sit tight and trim only the excess into 422 to 459. If you're flat, don't buy 417. Be ready at 400, and be bolder on confirmation than on hope.

FINAL TRANSACTION PROPOSAL: HOLD Aggressive Analyst: Let me give back what I should first, because it makes the rest of my case stronger.

You're both right that my 17-point stop was the weak link. It's 1.45 ATR on a stock whose ATR just jumped 36%, and the plan's real exit is 362.97. Blended risk on a 400 entry is about 27 points, which makes the reward about 1.25R to the 433 close and about 2R to 453. I sold 3.5R, which was the best case dressed up as the base case. You're also right that my breakout add was badly built. Buying a close above 433.68 against a 383 stop risks about 50 points for 20 to 25 of room, worse than the 417 entry I rejected, and 422.40 is under half an ATR away, so it proves nothing. "Either way we're positioned" was a rule about participation, not payoff. And Neutral is right that the ladder mixes bases. If you compare reported H1 EPS of 12.94 with adjusted guidance, you get an implied Q4 that is meaningless.

Now here is where I stop agreeing.

Conservative, you say the forward EPS of 23.26 assumes 50% growth over trailing, so the setup favors disappointment. Look at what the trailing number contains. It still holds Q3 2025's 2.59, and the feed's own arithmetic leaves almost nothing for Q4 2025, though the report flags that TTM EPS and net income don't reconcile. That "50% growth" is mostly the trough quarters rolling off. H1 2026 EPS is already 12.94, so 23.26 leaves 10.32 for the second half. That's about 5.16 a quarter against 6.47 a quarter in H1, a built-in step-down of roughly 20%. Whether that cushion is enough depends on the medical care ratio, which we don't have. But it isn't a heroic assumption, and "closer to 20x on the guide" compares a floor to a forward figure whose fiscal year the report says is unspecified.

On the data gap, I'd say the evidence is lopsided. The Oct 13 session had a gap open on 12.65M shares, Elevance's call has already happened, and the post's revenue matches the prior-year comp to a tenth. The print is almost certainly out. So the lookup is a ten-minute task, not a reason to wait two weeks. I agree there's no resting order until it's done. I'm with Neutral against the calendar gate. The Fed hold is 86% priced, beta is 0.6 and the House is 92% priced. Senate details won't be resolved by Nov 4 either. A limit near 400 only fills if the market hands us a lower price, and buying a dip into an event is how you get paid for event risk.

On OBV, Neutral already took 2.4M off the shortfall, because that's just the partial Oct 21 volume. The rest has context. OBV fell about 32M from Sept 28 to Oct 13 and has retraced about two-thirds of that in five sessions. The last two volume readings, 5.0M and 6.8M, sit inside the 3.3M to 7.5M range from before the breakout. That's normalizing, not drying up, though I'd want volume to expand on any break of 422.

On the fade pattern, July 16 faded and the stock drifted to 382. Oct 13 faded and the stock rose 10.5% in six sessions. One of each isn't a pattern. The 415.9 print is 0.8% under yesterday's close, on a bar whose open is above its own high. And the TD-9 is a pause flag, which is the pullback the plan is waiting for. It argues for the 400 order. Neutral is right that it stays valid above 406.67, so I won't dismiss it.

On margins, 2024 saw Q3 rise from Q2, but 2025 went 8.32 to 4.61 to 3.81. Two years, opposite patterns. Neutral is right that one unverified 5.8% can't carry weight for either of us.

On the lawsuit, the election and zero insider buying, none of these can be sized. That argues for small size and the 362.97 exit, not for abstaining. A repeat of January's gap costs about 5% of a standard allocation on a 25% tranche and about 10% on a 50% tranche. That's the price of admission, and the ladder already caps it.

On trimming, I'll take Neutral's version. The excess shares carry the same 1:1 math I used against buying at 417, so it would be inconsistent to keep them. Trim a third to a half now and the rest into 422 to 433.

Where I still differ is the "re-underwrite" language. A month-end close above 453.36 would put you 70 points from the 383 stop, which is the worst version of the payoff problem you both criticized in my rule. And a re-underwrite with no pre-committed outcome is the passive Hold I rejected. So here is the rebuilt rule. After the filing is verified, a close above 433.68 on expanding volume allows one quarter-size add. Its stop trails under the rising 10 EMA, so risk is about 20 points, not 50. The payoff comes from the 453.36 monthly flip and open air above the 461.62 52-week high, not from a fixed target. If the verified numbers fail, there is no add. I'd keep the ladder's sizes once it's restated on one EPS basis. I wouldn't halve them again, and I wouldn't make the election a condition for the second tranche. Price holding above 383 does that job.

So Hold stands. I don't know your holdings, so apply it this way. At standard weight, sit tight. If overweight, trim as above. If flat, place no order today. Do the lookup, then use a conditional entry near 400 on a daily close that holds the zone, sized by the ladder. I'd rather be early at a good price with small size than be right about everything and flat.

FINAL TRANSACTION PROPOSAL: HOLD Aggressive Analyst: I'll start with what I'm giving up, because it's a lot and I think it's right.

Four hundred is a fair price, not a discount. The plan's own valuation centers near 402, so the entry buys better exit geometry, not a margin of safety, and I'll stop calling it a good price. Neutral's July point also kills my 433.68 add. That was the closing high on July 21, and the stock was at 382.49 by August 20. I'll take Neutral's default instead: after a month-end close above 453.36, holders stop trimming, and anyone under standard weight can build toward a ladder tier with size set off the then-current stop. That is the pre-committed outcome I was asking for, and it doesn't pay up at a level with a failure history. I'll also take trimming about half of any excess now, a single clip at the 25% tier, and the point that the last two sessions' volume doesn't confirm anything. That's a reason not to buy here, and I never argued otherwise.

Now the disagreements. The first is the second tranche "no higher than the first fill." Say the first tranche goes in near 400 and the daily stop is 383. Then the second tranche can only be bought in a 17-point window above the level where the plan cuts half. It never fills in the scenario where the thesis is working, so the second rung exists only when the stock is deteriorating. That is averaging down into the stop. Conservative, you rightly attacked my rule for ignoring payoff, so let's use payoff. After November 4, allow the second tranche wherever the reward to 453.36 is at least 1.5R on the blended exit. At 400 that's 53 points over 27, about 2R. At 405 it's 48 over 32, 1.5R. At 410 it's about 1.2R. So the cutoff lands near 405, tied to the math instead of to wherever the first fill happened. That doesn't smuggle the breakout back in, because 435 isn't close.

Second, I disagree that the lookup can only veto. It sets the tier, in both directions, and the tier needs rewriting before anyone opens the filing. Subtract H1 EPS and a Q3 near 5 from the 20.50 floor and implied Q4 is about 2.5, roughly half the Q3 run rate. That is either a company burned in 2025 and sandbagging, or a seasonal cost step-up, and we can't tell which from here. A floor isn't a forecast, so keying a 5.00 threshold to it would say no position for the wrong reason. I'd write the rule now on what the filing can show, on the guidance basis:

  • Top tier: Q3 EPS about 5.00 or better and a medical care ratio flat or lower sequentially.
  • Middle tier: EPS of 4.50 to 5.00, or 5.00 or better with the ratio up but explained.
  • No position: EPS below 4.50, or a ratio that is up and unexplained.

The same basis problem hits the "needs a 13% beat" argument. If 23.26 is a 2026 number, Conservative is right that Q4 has to top Q3. If it's next year's, it is 13% growth over the floor, which isn't a stretch for a margin recovery. The report says the year is unspecified, so neither of us can use it as a hurdle.

On "the market already read the filing," that's true of the headline. But the stock is up 10.5% in six sessions and RBC is still raising targets this week. Estimate revisions and post-earnings drift usually play out over weeks, not in one day. That isn't a certainty and I won't size to it. Timing matters too. Right after a print is when scheduled gap risk is lowest, because the next one is a quarter away.

On the gap math, 2.1% against 5% still compares a modal gain with a tail loss and puts no probability on either. The window has one -19.6% day and two upside gaps of 7% or more. The unscheduled risks are real: the lawsuit, the Senate, and policy details. That's why the build is capped at 25% or 50% of standard. Conservative, the 125% stacking case also misreads the plan. The sizing line says existing holders keep up to standard weight, and the build is for flat or underweight books. The 399 stop cluster is exactly why entry waits for a daily close. We don't want to be the stop that gets run.

So here is the plan for your book, whatever it is:

  • Standard weight: sit tight.
  • Overweight: trim about half the excess now and the rest into 422 to 433.
  • Flat or underweight: place no order today. Do the lookup against the tiers above, and if it supports one, enter between 400 and 405 on a daily close that holds. Keep the second tranche inside the 1.5R test and after November 4. Hold total exposure at standard weight. Cut half under a daily close below 383.01 and exit under a weekly close below 362.97.
  • Month-end close above 453.36: apply Neutral's default.

Being a little early at a fair price with a small, capped position beats being right about everything and flat.

FINAL TRANSACTION PROPOSAL: HOLD Aggressive Analyst: Neutral, you won the 5.82% argument, and it cleans up a mistake I've been making. Your test holds. The 3.53% profit margin is exactly TTM net income over TTM revenue, and TTM EPS of 15.55 sits almost on top of Q3 2025 plus the two 2026 quarters. The same arithmetic leaves Q4 2025 near zero EPS but about $1.8 billion of net income, which is the reconciliation gap the fundamentals report already flagged. So the feed is messy. Still, neither reading gets you to the minus 2.4 a share you'd need if a Q3 near 5 were already in the snapshot. That makes 5.82% a trailing number and the match with the social post a coincidence. We're back to one unverified post that got revenue right.

I leaned on that post when I said the margin beat the 3.8% from a year ago. Conservative leaned on it to say the lookup will probably show a ratio that's up, and again to say the market cheered a sequential margin decline. All three uses come off the table. I'll also concede that I can't size post-earnings drift, so I won't size anything off it. But the filing does more than veto. The 20.50 floor less H1's 12.94 less a Q3 near 5 leaves roughly 2.5 for Q4, on mixed bases, so treat that as rough. Either the company is sandbagging or it expects a heavy seasonal cost step-up, and only the filing says which. That decides the tier.

On "explained," Conservative, fixing the definition before anyone looks is right, but yours would reject a quarter where costs landed where the company said they would. Neutral's version is better: test the medical care ratio against the company's own full-year guide, and check whether guidance rose on the same basis. I'd drop the year-ago comparison entirely, because the year-ago quarter was the 3.8% trough and anything beats it. I used that comp myself and it proves nothing. I'd also fix a tolerance now, say within half a point of the guide for the middle tier, so nobody moves it after seeing the numbers.

Top tier is adjusted Q3 EPS around 5.00 or better, a ratio at or inside the company's range, and guidance raised like-for-like. Middle tier is EPS of 4.50 to 5.00, or a ratio inside the tolerance with guidance still raised. Anything else is no position. If the post is right, EPS barely discriminates, so the ratio test does the work, and that's fine as long as it's written down.

On the second tranche I'll take Neutral's ceiling near 403, after November 4, on a close above the 10 EMA, no larger than the first, inside the cap. It's an option, and if it never fills we lose nothing. One thing we've all glossed over is that every level in the plan is today's. The daily supertrend only ratchets up while the trend holds, so a fill near 400 two weeks from now may face a daily stop well above 383.01. The rule should say "current daily stop," with the weekly stop doing the full exit, and the risk math should be rerun at the time of fill.

I agree a close at 392 is a failed shelf, not a bargain. On the lower re-look at 384 to 387, I'd note the payoff improves there, since the weekly stop is the real risk and the 50 SMA, middle band and daily line cluster. But a close under that daily line halves the position, so it needs fresh math, as Neutral said.

I'm retracting "early beats right and flat." At a fair price with no margin of safety, early is just exposure, and Neutral is right that flat has a cost we're knowingly accepting. What I'll keep is narrower. When the filing, the price and the daily close agree, nobody should argue themselves out of it with an unsized headline. The FOMC, the election, the lawsuit and open enrollment are real. They're priced through the 25 to 50 percent cap and the weekly exit, not through extra gates. Conservative has moved the election gate onto the second tranche only, and I'm fine with that. The first clip shouldn't wait for November 4.

One edit to the original plan that nobody flagged is that the half-cut is written as a daily close below 383.01 on rising volume. Both of you quietly dropped the volume qualifier in your restatements, and I'd make that official. A mechanical stop shouldn't have an escape hatch for a thin-volume close.

Applied to your own book, since I don't know your holdings: - Standard weight: sit tight. - Overweight: trim about half the excess now and the rest into 422 to 433. - Flat or underweight: place no order today. First confirm Q3 is actually out, then score it against the tiers fixed in advance. If a tier is met, buy on the first daily close inside 400 to 405. Enter the 25 percent tier in one clip and the 50 percent tier as two equal halves, with total exposure capped at standard weight. - Exits: cut half on a daily close under the live daily stop, currently 383.01, and exit fully on a weekly close under the weekly stop, currently 362.97. Size for the gap, not the 17-point line. - No add on strength. A month-end close above 453.36 stops the trimming and triggers fresh sizing off the then-current stops.

The boldness in this plan isn't paying up. It's that when the top tier, the price and the close all agree, we take the full size without a seminar.

FINAL TRANSACTION PROPOSAL: HOLD Aggressive Analyst: Let me start with what you two have fixed, because the plan is better than the one I opened with. Neutral's payoff-defined entry is the right construction, since it moves with the stops instead of hard-coding today's prices. Scoring the medical care ratio excluding prior-period development is a good catch, and only the filing can show it. I'll take "and" in the middle tier and ties going down a tier. I'll take a one-tier penalty for a steep implied Q4 step, but only one tier. A floor-only guide makes implied Q4 look steep almost by construction, so use the midpoint if the company gives a range. The tactical and core split on the daily half-cut also makes sense. At 417 the daily line is almost three ATR away. A sleeve bought near 400 has only about 1.45 ATR of room, so that's where the tight cut belongs, and the weekly stop should govern the core.

Now the problem, which I can show with arithmetic. The entry as written has one very narrow door. The first tranche needs a daily close above the live 10 EMA and inside the payoff ceiling near 402. The EMA is 400.09, and it rises whenever price stays above it. If the stock merely closes near 415 tomorrow, the EMA moves to about 402.8, already at the ceiling. After five sessions near 415 it's around 409 or 410. Even a steady glide of three points a day, from 414 down to 402 over five sessions, leaves the EMA near 404. The 402 close then sits under it and fails. In a pullback the stops don't ratchet, so the ceiling stays put while the EMA keeps climbing. By my arithmetic, the two conditions overlap only if the stock drops almost immediately. Neutral, you said the entry would fire often enough that the plan isn't passive. With both conditions in place, I don't think it does. That is the Hold with no path to buy that we all said we wanted to avoid.

The fix is in Neutral's earlier round and got dropped from the final list. Add a second door at 384 to 387, where the 50 SMA, the middle band and the daily line cluster. It requires a daily close, not an intraday touch, and both the daily and weekly supertrends still up. At that price the daily line is a few points away, so by Conservative's own sizing rule the real risk is the weekly stop, about 21 to 24 points. From 386 that's roughly 2R to the 433.68 close and just under 3R to 453.36, which passes the same test. A close at 392 stays a no-buy, a failed shelf, as Neutral said. So the rule has two doors: a prompt dip to about 402 above the EMA, or a deeper retrace into the cluster with the trend intact, and nothing in between. If you won't add the second door, then say plainly that the first one is a few-session option and otherwise the plan is to stay flat. Don't let anyone think they have a buying plan when they have a wish.

Conservative, on "full size without a seminar," I'll narrow the claim. I'm not skipping checks. Once the tests written in advance pass, a new objection should count only if it's new information: a change in the filing, the price or the stops. The FOMC, the election, the lawsuit and open enrollment were all on the calendar when the tiers were written. You argued that a fair price near 402 doesn't pay you to carry them. But you told a standard-weight holder to sit tight at 417 through the same events at a worse price, and nobody here proposes cutting to zero before November 4. The top tier is half a standard allocation, half the size of the position you'd have a holder keep at a higher price. A January-style gap costs about 10% of a standard allocation on that tier, with a written weekly exit. The amounts are bounded, and that's the price of admission.

On Q4, you're right that the question can't be dropped, and Neutral is right that management rarely narrates a floor, so the step test is the better tool. One caution: the feeds contain no Q4 quarters at all, so the comparison year has to come from the filing. 2024 wasn't entirely clean either, since the feed has that Q1 anomaly. Treat it as a rough yardstick, not a precise bar. I'd also keep the filing as a gate, not a formality. The world report still treats Q3 as upcoming, so step one is confirming it's out.

Here's how to apply it to your own book, since I don't know your holdings.

If you're at standard weight, sit tight. The weekly stop at 362.97 governs the core.

If you're overweight, trim about half the excess now and the rest into 422 to 433.

If you're flat or underweight, place no order today. Confirm Q3 is out, then score it against tiers fixed in advance. Those tiers use "and," send ties down, and include the ratio excluding prior-period development and the step test. If a tier is met, buy only through one of the two doors on a daily close. The 25% tier goes in as one clip, and the 50% tier goes in as two halves. The second half comes after November 4 and only if the payoff test still passes. Total exposure stays capped at standard weight.

Cut half of any tactical sleeve on a daily close under the live daily stop, 383.01 today, with no volume condition. Exit everything on a weekly close under the weekly stop. Don't buy strength. A month-end close above 453.36 means stop trimming and resize off the then-current stops.

The aggressive part of this plan isn't paying up. It's making sure there is a door to walk through when the price, the filing and the close all line up.

FINAL TRANSACTION PROPOSAL: HOLD