AMZ.US Amazon - Trade Plan — 2026-05-06 | trading-OS

TL;DR

LONG AMZN via a two-leg staged limit entry into the $248–$258 pullback zone (anchor fill at $250.00), with a hard stop at $234.00 below the swing-pivot cluster. First target $264 (prior ATH / S-turned-R), second target $279 (BB upper / pattern extension), horizon 84 trading days (~4 months) carrying through the 2026-07-31 Q2 earnings catalyst. Total target size 4.0% of NAV, filled in two halves: 2.0% NAV at $252.00, 2.0% NAV on confirmation at $250.00 or below. No initial buying at $273.55 — chasing is explicitly prohibited by the verdict (research-verdict.md §1, §7).

Source verdict

"LONG, conviction 3/5. Directional view: long. Entry discipline: staged into the $248–$258 pullback zone, not chased at $273.55… The directional verdict is LONG; the conviction reflects that the current price is not the entry price." — research-verdict.md §1.

"Within the horizon the entry zone is $248–$258 (preferred: $250–$252 on a constructive pullback to flag support); the structural stop is $234–$240 (below the swing-pivot cluster at $236.74); the trade is invalidated on a daily close below $227.41 (SMA50). This is not a market-on-open long at $273.55 — both researchers explicitly rejected that framing by the close of round 2." — research-verdict.md §7.

Conviction: 3/5. Horizon: 3–6 months. The verdict is the input; this plan operationalises it.

Direction & instrument

LONG AMZN common stock (US listing).

No options overlay at initiation. AMZN HV20 is 27.0% (technical.md, Volatility table) — not extreme — and sentiment.md shows no euphoria flag, so put hedges are unnecessary alpha-drag at this entry. Calls are not used as a substitute for shares; the verdict's discipline is about entry price, not about capping notional, and replacing shares with calls would re-introduce the chase at a different layer.
No pair short. The bull's case rests on AMZN-specific re-acceleration (fundamentals.md, Bull Case bullet 1; news.md items #1–#2), not on hyperscaler-relative trading.
US large-cap, deeply liquid, no borrow or capacity constraint considerations. Direction is unambiguous.

Entry

Method: scaled limit (two legs)
Validity: GTC, valid for 8 weeks (through 2026-07-01) — the pullback may take time to develop and the verdict explicitly accepts waiting (research-verdict.md §7). Cancel and re-evaluate if not filled by 2026-07-01 (≈30 days before Q2 earnings).

Leg	Price	Size	Trigger	Rationale
1	$252.00	50% of position (2.0% NAV)	Limit GTC	Top of preferred zone (`research-verdict.md` §7); sits at EMA21 ($252.19, `technical.md` Trend table) and BB mid ($253.32) — a structurally meaningful pullback that does not require deep weakness.
2	$250.00	50% of position (2.0% NAV)	Limit GTC, only after Leg 1 fills	Anchor of preferred entry (`research-verdict.md` §7, $250–$252 on constructive pullback to flag support). Avoids over-committing if the pullback is shallow.

Reserve / contingency leg (not initiated automatically): if price trades to $248.28 (swing-pivot low, technical.md Support table) on non-distribution volume and leg 2 has filled, the trader may add a third 1.0% NAV tranche subject to risk-committee re-approval. This is a contingent action, not part of the base plan.

Why staged, not single-shot: conviction is 3/5, not 4/5; the position-sizing skill discipline (see Size below) and the verdict's explicit "do not chase" both argue against a single all-in fill. Scaling also smooths fill quality through the LVN gap at $248–$256 (technical.md, S/R note) where price can move quickly.

Why not a market or breakout entry:

Market entry at $273.55 is forbidden by the verdict.
The breakout-entry alternative (technical.md, Timing View, "$280 on close above BB upper on >1.2x avg vol") is explicitly inferior at ~1.5:1 R/R vs the pullback's ~2.0:1, and the research-manager already chose the pullback path. We do not second-guess that.

Stop

Price: $234.00
Type: hard stop on a daily close basis (intraday wicks below $234 do not trigger; only an end-of-day close below $234 exits the position).
Justification:
- Sits below the $236.74 swing-pivot (technical.md Support table, 2026-01-29 low) and just above the top of the $232–$236 HVN cluster (technical.md, Volume HVN table: $230.26, $222.68, $220.79 are the heaviest-volume nodes in the 252-day profile).
- Aligns with the technical-analyst's structural-stop band of "$234–235" derived from a 2×ATR stop on a $250 entry: $250 − 2×$7.43 = $235.14 (technical.md, Timing View, "ATR-derived stop suggestion for long entries at $250: $235.14, just above the HVN cluster at $232–$236").
- Below the flag-invalidation level of $248 (technical.md, Pattern, "A close below $248 cancels the pattern"), so a stop at $234 does not trigger on flag-invalidation alone — it requires the deeper retest the verdict explicitly contemplates (research-verdict.md §6 invalidation #2: "opens the LVN gap from $248–$256, and projects a retest of the $232–$236 swing-pivot / HVN base").
- Validates the noise-floor test: ATR14 = $7.43; daily 1σ on a $250 entry ≈ $4.20 (HV20 27% / √252). Stop distance from anchor fill = $250 − $234 = $16, or 3.8σ of expected daily move — well above the 1.5σ minimum, so noise-only trips are unlikely.
- Stop distance as % of anchor entry = 6.4%. Combined with 4.0% NAV size, position-level risk is 0.26% of NAV — inside the sizing skill's hard ceiling of 1.0% per-trade NAV-at-risk.

Soft-stop overlay (event-driven, see Thesis-invalidating events below): the stop above is the price invalidation. The thesis also exits regardless of price on specific catalyst conditions — operationally a "close-the-day" instruction to discretionary execution even if $234 has not been touched.

Trail rule:

After Target 1 ($264) prints, raise stop to entry-anchor minus 1× ATR = $250 − $7.43 = $242.50 (rounded to $242.00 for cleanliness). This locks in a small loss-or-flat position while preserving room for the second leg to develop.
After Target 2 ($279) prints, raise stop to $264.00 (TP1 level), converting the residual into a free-roll for the pattern-projection target.

Targets

Tier	Price	Action	Rationale
TP1	$264.00	Trim 50% of position (2.0% NAV)	Prior all-time high pre-2026-05-05 / S-turned-R; explicit Target 1 in `technical.md` Timing View pullback row. From anchor entry $250.00: reward $14, risk $16, R/R = 0.88:1 on TP1 alone — but TP1 is a partial exit, not the full case. Including TP2 weighting the blended R/R is materially better (see below).
TP2	$279.00	Trim a further 25% (1.0% NAV); leave 1.0% NAV runner	BB upper extension + ATH extension (`technical.md` Resistance: $279.90 BB upper); explicit Target 2 in `technical.md` Timing View. From anchor entry: reward $29, risk $16, R/R = 1.81:1.
TP3 (stretch)	$295–$310	Trim final runner in stages	Breakout-entry continuation targets per `technical.md` Timing View row 2 ($295 / $310). Not the high-tight-flag mechanical projection of $347–$352, which `technical.md` Pattern explicitly tags as "mechanical projection only … less reliable than in classic base-breakout scenarios." We do not target prayers.

Blended R/R analysis (entry $250, stop $234):

50% out at $264 → +$14 × 0.5 = +$7.00
25% out at $279 → +$29 × 0.25 = +$7.25
25% runner at $300 (mid of TP3 band) → +$50 × 0.25 = +$12.50
Expected gross win = +$26.75 per share
Loss case = $16.00 per share at full size
Blended R/R ≈ 1.67:1, conservative because it assumes only partial trail capture on the runner. Acceptable for conviction-3.

Horizon

84 trading days (≈4 months), valid through 2026-09-01.

The verdict gives 3–6 months (research-verdict.md §7). 84 days sits at the lower-middle of that range.
The decision-relevant catalyst is Q2 2026 earnings on 2026-07-31 (research-verdict.md §7; news.md Catalyst Calendar). The horizon explicitly carries through earnings — we accept gap risk because the thesis is about the AWS print itself (fundamentals.md "What Would Change My Mind" #1; news.md "Key earnings watch for Q2 2026").
Secondary calendar items inside the horizon (news.md Catalyst Calendar): AWS re:Invent preview ~2026-05-21; June FOMC 2026-06-04 / 2026-06-18; CPI 2026-06-11; Prime Day ~2026-06-15. None are thesis-binding; they are tape-risk to size through, not to plan around.
If the position has not begun filling by 2026-07-01 (i.e., neither staged leg has triggered), cancel orders and re-evaluate. We do not initiate new long exposure within 30 days of the catalyst at this conviction level.

Size

Method: minimum of fixed-fractional, vol-targeted, and quarter-Kelly, rounded to round lots (per the position-sizing skill's default rule).

Method	Inputs	Output
Fixed-fractional	0.25% NAV-at-risk per trade × $16 stop ÷ $250 entry	3.9% of NAV
Vol-targeted	10% portfolio vol target ÷ 27% HV20 × 1.0 weight	~37% NAV (gross), capped to single-name limit
Quarter-Kelly	win-rate ~52% (conviction-3 prior), R/R 1.67:1, k=0.25	~4.0% NAV
Single-name cap	per CLAUDE.md / sizing skill	10% NAV (not binding here)
Per-trade risk cap	per sizing skill	1.0% NAV-at-risk (not binding)

Selected size: 4.0% of NAV (the minimum of the three method outputs after capping the vol-targeted figure at the single-name limit; quarter-Kelly is the binding constraint).

Leg	%NAV	Trigger price	NAV-at-risk per leg
1	2.0%	$252.00	0.144% (stop $234)
2	2.0%	$250.00	0.128% (stop $234)
Total	4.0%	blended $251.00	0.272%

Position-level NAV-at-risk of 0.27% is well inside the 1.0% per-trade ceiling, leaving headroom for the contingent third tranche described under Entry. Round-lot rounding is portfolio-size dependent and applied at execution; for a $10M NAV portfolio this is roughly 800 shares total (400 + 400) — the round-lot rounding is immaterial at this size.

Why not 5–6%: conviction is 3/5, not 4/5. The verdict explicitly warns that "a LONG verdict at conviction 3 with an explicit 'do not chase' instruction is materially different from a conviction-4 chase- LONG, and the portfolio-manager's eventual five-tier rating should reflect that" (research-verdict.md Memory Note). Sizing reflects that. The risk committee may push back; that is the protocol working as intended.

Critical assumptions

AWS Q2 2026 (2026-07-31) prints AWS revenue growth ≥ 22% YoY. This is the bull's stated assumption-that-must-hold (research-verdict.md §6.1). Below 22% flips verdict to NO-TRADE; below 20% flips to SHORT.
The $248–$258 pullback zone is reached within the 8-week order window (i.e., by 2026-07-01). If price does not pull back and instead breaks above $279.90 on >1.2x volume, the plan does nothing — we accept the missed trade rather than chasing. The verdict explicitly contemplated this.
The high-tight flag at $256–$278 holds — i.e., no daily close below $256.16 on >1.2x average volume in the next 4–6 weeks (research-verdict.md §6.2). If this breaks before our entries fill, cancel orders — the setup has changed and the new structure needs re-analysis, not the same plan executed lower.
AR/Revenue ratio does not expand for a third consecutive quarter in Q2 (research-verdict.md §6.3 / fundamentals.md "What Would Change My Mind" #3). Two quarters is yellow; three converts to red and downgrades the fundamentals verdict.
No material new geopolitical escalation against AWS data centres beyond the existing UAE/Bahrain repair situation (sentiment.md Theme 3). The current exposure is priced; an escalation that affects AWS revenue (vs current infrastructure-repair) is a thesis hit, not just a sentiment hit.
Sentiment regime stays NEUTRAL or below (sentiment.md, Crowded-Trade Check: all five danger signals currently absent). A move into euphoric/crowded territory before our entry fills would re-introduce chase-risk under a new label and require re-sizing.

Thesis-invalidating events

These cause an immediate exit regardless of price level — the $234 hard stop catches price invalidation; this list catches event invalidation.

Q2 2026 AWS YoY growth < 22% (announced 2026-07-31). Exit on open of 2026-08-01 regardless of price. (research-verdict.md §6.1; fundamentals.md "What Would Change My Mind" #1; news.md "What Would Change This View" item 1.)
Daily close below $256.16 on >1.2x average volume within 4–6 weeks of this plan's filing date — even if the position has filled at a price below $256.16. This invalidates the high-tight-flag structure that the entry zone depends on (research-verdict.md §6.2; technical.md Pattern Invalidation). Operationally: if leg 1 has filled at $252 and price then closes at $254 on 1.5x volume the next session, the stop at $234 is not yet hit but the structure is broken — exit on the next open.
FY2027 capex guidance disclosed at Q4 2026 earnings remains above $150B combined with continued AR/Revenue expansion (research-verdict.md §6.3). This is outside the 84-day horizon for most cases but listed for completeness if horizon is extended.
FTC obtains an injunction or structural remedy before trial conclusion (news.md "What Would Change This View" item 2). Low probability but high impact; an injunction is binary.
A second AWS-data-centre kinetic event with documented service disruption beyond the UAE/Bahrain incident (sentiment.md Theme 3 escalation). The current incident is a sunk cost; a second re-prices cloud-sector geopolitical risk.
CEO Andy Jassy departure or reverses his $10M open-market purchase (news.md, Form 4 2026-04-20). The Jassy buy is referenced in the bull's pillar of confluence (research-verdict.md §3 indirectly); reversal would be a meaningful insider signal against.

What I'm explicitly not doing

Not buying any size at $273.55, the current price. Both bull and bear converged on this; the verdict made it binding.
Not using options as a chase substitute. Buying calls now achieves the same chase at a different layer; passed.
Not pre-emptively trimming for earnings. The horizon deliberately carries through Q2 2026 earnings on 2026-07-31 because the print is the thesis. Cutting before the catalyst monetises the wait but discards the asymmetric outcome the staged entry was designed to capture. If risk-committee challenges this, the alternative is to not enter at all — exiting before the catalyst is the worst of both worlds.
Not chasing the high-tight-flag $347–$352 mechanical projection. technical.md itself flags this as "less reliable than in classic base-breakout scenarios"; targeting it would be a prayer, not a plan.
Not adding on weakness below $234 (the structural stop). If the stop is hit, the trade is wrong on price; re-evaluation requires a new analyst pack, not "averaging down with conviction."
Not reducing size based on the AR/Revenue yellow flag pre-Q2. The fundamentals analyst graded it yellow with B2B-mix-shift interpretation (fundamentals.md Red Flag Scan); the verdict treats Q2 AR ratio as an invalidation trigger rather than a sizing governor. Acting twice on the same signal is double-counting.

Position management decision rules (operational)

Situation	Rule
Price reaches $252	Leg 1 fills automatically (limit GTC).
Price reaches $250 after Leg 1 has filled	Leg 2 fills.
Price reaches $250 without Leg 1 having filled	Leg 1 still has higher trigger; Leg 2 will not pre-empt. (Both are GTC limits, not OCO.)
Price reaches $248 on non-distribution volume after both legs fill	Optional contingent third tranche of 1.0% NAV — requires risk-committee re-approval, not auto-executed.
Price closes below $234 (any session before TP1)	Hard stop; full exit at next-day open.
Daily close below $256.16 on >1.2x average volume (any session)	Soft thesis-stop; exit even if price stop not hit.
Price reaches $264 (TP1)	Trim 50% (2.0% NAV); raise stop to $242.
Price reaches $279 (TP2)	Trim further 25% (1.0% NAV); raise stop to $264.
Price between $279 and $295	Hold runner; no further trims.
Price reaches $295	Trim half of remaining runner (0.5% NAV).
Price reaches $310	Exit final 0.5% NAV; trade complete.
Q2 earnings 2026-07-31 prints AWS < 22%	Full exit on 2026-08-01 open regardless of price.
2026-07-01 reached with no fills	Cancel orders; re-evaluate against fresh analyst pack.
2026-09-01 reached (horizon end) with position still open and no TP/SL hit	Re-evaluate against fresh analyst pack; do not roll mechanically.

References

data/reports/AMZ.US/2026-05-06/research-verdict.md — verdict, conviction, entry zone, stop band, invalidation conditions, horizon.
data/reports/AMZ.US/2026-05-06/technical.md — entry/stop levels, ATR, S/R, HVN profile, regime, pattern, timing view.
data/reports/AMZ.US/2026-05-06/fundamentals.md — AWS / advertising margin runway thesis, AR-ratio yellow flag, "What Would Change My Mind" thresholds.
data/reports/AMZ.US/2026-05-06/news.md — Q1 2026 earnings beat, Jassy supply-constrained guidance, Jassy Form 4 buy, catalyst calendar, "What Would Change This View".
data/reports/AMZ.US/2026-05-06/sentiment.md — neutral regime, no crowded-trade flags, drone-strike risk narrative.
data/reports/AMZ.US/2026-05-06/memory.md — first tracked AMZ.US decision; conviction cap of 4 on first call; no streak adjustment.

This is research, not trading advice. The trading-ops system never auto-executes orders. This plan is delivered to the risk committee next; their stress-test may modify entry, stop, target, or size before the portfolio-manager renders a final five-tier rating.