NAS100 Long on May 4: A C+ VWAP Pullback That Opens the Month

The macro tape on May 4

The May 4 New York AM open carried a divided macro signature. The 10-year yield at 4.422 percent was the headline factor. It sat 3.6 basis points above its 5-day EMA at 4.386 percent, had broken above yesterday's session high at 4.402 percent, and was trading 0.8 basis points away from the 5-day peak at 4.430 percent. The framework treats the 10-year yield as the primary driver for NAS100 trades. Yields trending above their EMA and approaching multi-day highs is a structural headwind for any long, and the system's hard rule is that yields breaking above the 5-day high triggers an immediate ban on new long positions and a forced exit on any open long. We were 0.8 basis points away from that switch.

The cross-asset tape did not confirm the yield signal cleanly. The VIX at 17.42 was slightly below its 5-day EMA at 17.56, which the system reads as neutral rather than bearish. The DXY at 98.282 sat below its 5-day EMA at 98.427, also neutral. Breadth was actively bearish: the NYSE advance-decline line at -397 ran well below its 5-day EMA at -73.4 and below yesterday's session low at -215. The picture was a yield headwind on a soft dollar with neutral volatility and broadly negative breadth, the kind of mixed read where the system declines trend-continuation longs and only entertains structural mean-reversion entries at well-defined levels.

The Macro Agent read this configuration as lean-bull at 56 percent confidence. The bullish factors cited were technical: a capitulation bounce of 754 points off the recent low and price holding 325 basis points above the 5-day EMA. The agent did not cite rates as a factor. That detail matters in the framework. A macro lean-bull read that does not cite the primary driver carries reduced weight, and the system applies a confidence haircut that pulls the effective conviction toward the bottom of the lean-bull band.

What the chart was showing

By 14:00 UTC the 60-minute timeframe presented a technically bullish but momentum-fading picture. The fast EMA at 27,751 sat above the slow EMA at 27,620 with a 131-point spread. RSI on the 60-minute was 58.3, neutral-bullish but not extended. MACD line was above zero, but the histogram had been negative for four consecutive bars, indicating the bullish impulse from the overnight rally had stalled. Price at 27,792 was holding above the session VWAP at 27,775 by a thin margin. ATR on the 60-minute was 60.6 points, normal volatility, no expansion regime.

The 15-minute timeframe was transitional. The fast EMA at 27,761.4 sat 5 points below the slow EMA at 27,766.6, a bearish stack but compressing toward an imminent crossover. RSI was 55.1, just above the neutral midline. MACD line was -2.13 below zero, but the histogram was +8.74 and expanding positive, the signature of a bullish momentum cross developing inside a bearish stack about to flip. Price had reclaimed VWAP on the 15-minute. The structural picture across timeframes was: 60-minute bullish but tired, 15-minute mid-cross from bearish to bullish, 5-minute already turned with a confirmed bullish EMA crossover at 13:55.

The setup grade returned 5 of 7 confluences cleared. The two missing factors were both rate-related: the 10-year yield did not support longs, and the Macro Agent did not align at 60 percent or higher with a rate citation. Three structural factors cleared: the Trend Agent at 63 percent bullish, the 60-minute EMA stack confirmation, and the 5-minute reaction at the VWAP cluster. Two execution factors cleared: 15-minute RSI above 50 with expanding MACD, and a clear calendar with no high-impact events for 24 hours. The grade landed at C+ in the lower band of medium-high, with sizing reduced to 0.75 percent equity risk versus the standard 1 percent for the rate-driven downgrade.

The setup the Trend Agent flagged was a NAS100 Long VWAP Pullback Buy in a confirmed intraday uptrend with macro caveats. It is one of the most-traded patterns in NY AM session intraday work, and walking through it explains why the system declined nine times before the second enter finally cleared.

What the pattern is

Price establishes an intraday uptrend on the 60-minute timeframe: fast EMA above slow EMA, price holding above both, momentum confirmed by RSI in healthy bullish territory. Inside that uptrend, price retraces against the move into a structural support zone. The cleanest version of that zone is the session VWAP itself, the volume-weighted average price computed from the cash open. The entry is not the touch of VWAP. The entry is a five-minute bullish reaction inside the VWAP zone: a rejection candle, a close back above the touched level, RSI lifting above its midline, MACD histogram turning positive on the next bar. Without that reaction, the touch is just a touch.

How professional traders actually use it

This is a staple of NY AM session continuation. The math favors confirmed entries over chases. Buying the upper extension after a 200-point rally exposes the position to the first mean-reversion bar. Buying the VWAP retest with a rejection body inside the zone places the entry at the bottom of the next leg, with the stop sitting just below structural invalidation. The R per unit of risk improves dramatically, often shifting a setup that would be 0.3R from extension to 1.0R or higher from VWAP.

Volume is the tell. A quiet pullback into VWAP means thin participation, and the level is being grazed rather than defended. A pullback that arrives on average-or-better volume and bounces with above-average volume is the level holding because real demand is stepping in. The 5-minute volume signature on the rejection bar is the difference between a tradeable VWAP retest and a continuation lower that just paused at the average.

Why it works

VWAP is the most-watched intraday reference in institutional execution. Algorithms benchmark against it, desks fill orders relative to it, and discretionary participants size off its slope. When price approaches VWAP from above and holds, the level is absorbing supply at a price the average market participant agrees is fair for the session. When price approaches VWAP from above and breaks through, the average has shifted, and the prior leg is invalidated. The pattern works because the level itself is functional, not symbolic. Real orders sit there.

It fails in the wrong regime. A NAS100 Long VWAP Pullback Buy in a confirmed bear trend, or against an active yield spike at the highs, will see the VWAP retest fail and price continue lower. That is why the Macro Agent's regime read gates the pattern before the Trend Agent is allowed to score it. On May 4 the macro read was specifically conflicted: rate signal bearish, dollar and volatility neutral, breadth weak. The gate cleared at reduced confidence, capping the setup at C+ with reduced sizing rather than blocking the trade outright.

How the system reads this, dynamically not dogmatically

SkyAnalyst does not favor the NAS100 Long VWAP Pullback Buy as a strategy. The same morning, the agents fired two short setups on the dollar's other side: a XAUUSD short that ran to TP2 for +4.9R after waiting through five WAIT evaluations, and a EURUSD short on a single VWAP rejection evaluation that closed at TP2 for +2.22R. Two of those three trades on May 4 were shorts. Each was a different playbook with a different logic and a different edge.

The system reads the tape first and fits the pattern to what is actually there. It does not show up to the chart with a playbook and look for opportunities to run a preferred setup. The four agents running in parallel, trend, macro, cross-asset, risk, each contribute a different lens on what kind of market this is. When they agree, we trade. When they do not, we sit out. On May 4 the agents agreed three times in three different shapes: a multi-evaluation gold short on a confirmed rejection cluster, a single-evaluation euro short on a VWAP rejection, and an eleven-evaluation NAS100 long on a VWAP retest that took two enter signals to fill. The system does not have a favorite setup. It has a process.

What this trade teaches

The structural premise for the NAS100 long was bullish at every evaluation across the thirty-minute window. EMA stack aligned on the 60-minute, MACD histogram expanding positive on the 15-minute, RSI rising back above the midline on the 5-minute. The system declined nine times and aborted one enter before the second enter at 14:36 finally fired and held. The only thing that changed at 14:36 was that the confirmation candle printed inside the entry zone with the volume and momentum signatures the pattern requires, and the next bar did not invalidate it.

That sequence is what discipline looks like in code on a divided macro tape. A discretionary trader watching the same chart would have felt the pull to enter at 14:13, when the level was holding and the structural picture was clear. The nine wait cycles and the aborted enter at 14:15 are not the system being indecisive. They are the system refusing to act on a setup until the specific confirming evidence prints, and then refusing to fill the entry when the next bar after the signal walks the trigger back. The 14:36 enter cleared because the 14:35 bar printed the rejection body with above-average volume and the 14:36 bar did not invalidate it. Two of the eleven evaluations saw a tradeable signal. One filled.

A C+ setup that closes at TP1 is the median trade. May opens with the same discipline the system ran through April. The framework does not size up because a new month started, and it does not skip a qualifying setup because the prior month was uneven. - From the desk - May 4, 2026

The trade then ran 30 points to TP1 at 27,810 over fourteen minutes with no recorded drawdown, closing at +0.4R (TP1) and +$800 (TP1) on the hypothetical $100,000 account at 2 percent risk. The same C+ grade on a less cooperative tape would have stopped at 27,705 in roughly the same window. TP2 at 27,855-27,865 and TP3 at 27,920-27,940 would have paid 1.15R and 2.0R respectively if the move had extended. The market did not deliver more than the first level on this session, and the position closed on the broker's TP1 fill.

What did not happen on this trade

Three counterfactuals matter here. The system did not enter on the first ENTER signal at 14:15 because the next bar walked back the trigger before fill, and the framework treats trigger-then-fade as a non-entry. The system did not pass on the trade when the Macro Agent confidence sat at 56 percent without citing rates, the lean-bull read held the gate open at reduced sizing rather than blocking the trade. The system did not run the trade at full size, the rate-driven downgrade pulled equity risk to 0.75 percent versus the standard 1 percent. Each of these decisions was made by a different agent before the position entered. None required the operator to override the rules.

The May month-to-date entering this trade was zero, since this was the first NAS100 trigger of the month. Adding +0.4R (TP1) here is the first credit on the May ledger. The April month-to-date that fed into this trade is documented in the April monthly recap, and the structural arc through the same April closing week is in the Apr 27-May 3 drawdown report.

From the desk

The interesting thing about this trade is not that it ran. The NAS100 Long VWAP Pullback Buy is a textbook NY AM session setup, and a clean execution to TP1 on a 0.4R outcome is exactly what the pattern produces when the inputs are right and the market does not extend further. The interesting thing is that May opened with three triggers across two directions on a single morning, and the system sized each one according to the conditions at entry without reference to the rolling tally.

The XAUUSD short on May 4 ran at half normal sizing because the Macro Agent flagged tradeability at 25 of 100. The EURUSD short ran at 0.5 to 0.75 percent equity risk because the Trend Agent flagged the regime as TRANSITIONING. This NAS100 long ran at 0.75 percent equity risk because the rate-driven macro downgrade pulled the setup grade into the lower band of medium-high. Three trades on the same morning, three different sizing decisions, all made before the positions entered. None required hindsight. The system does not size up after a winner and size down after a loser. It sizes based on the regime call at entry time, and it accepts that the same morning can produce a +4.9R outlier, a +2.22R clean continuation, and a +0.4R median trade depending on what the chart needs to deliver.

A reasonable question by now is whether a retail trader with ChatGPT and a TradingView chart could reproduce this. They cannot, and not because of model quality. On May 4 the Macro Agent had written its 56 percent lean-bull read with no rate citation into the shared state at 13:30 UTC and had not updated it since. The Trend Agent, on its eleven evaluations, read that value and used it to gate the setup at C+ with reduced sizing rather than the higher conviction the structural picture alone might have justified. If the Macro Agent had been chatting in prose about "yields are concerning but technicals look constructive," the Trend Agent would have had to interpret the tone. It does not, so it did not. The coordination between the four agents is the product. That is what a chat interface cannot simulate, and it is what this case study shows in practice on a setup that took eleven evaluations and two enter signals before the trigger conditions assembled inside a single five-minute bar without invalidation.

Through May 4, 2026, the cumulative ledger reads roughly +14.9R YTD across 88 trades from January 12 inception. This trade contributes +0.4R (TP1) to the ledger at the credited TP1 level, the simulated $100,000 account at 2 percent risk per trade tracks +$800 (TP1) on this single trade in dollar terms. At the reduced 0.75 percent applied to this trade, the dollar return scaled proportionally lower than the headline figure. For the same-session companion shorts that ran the macro tape's other side, see the May 4 gold short case study and the EURUSD VWAP rejection short.

The next case study will be filed when its position closes. We will work through May the same way we worked through April: one trade at a time, the median trade the same as the outlier.

From the SkyAnalyst Team.