My Methodology: Elliott Wave + Wyckoff + Fibonacci Applied to SPX/SPY

After years of analyzing the S&P 500 (SPX500 Future and SPY) day by day, I ended up developing my own approach that combines several classic technical analysis tools, refined over time through trial and error. This post is the starting point of this blog: here I explain how I think about the market, so that every analysis I publish from now on has this framework behind it.

It's not a magic formula or a crystal ball. It's a multi-confirmation process: I don't make a directional call until several different frameworks point in the same direction.

1. The foundation: Elliott Wave

I use Elliott Wave counting as the backbone of market structure. But instead of applying textbook rules without questioning them, I added three rules of my own that have consistently worked for me, and that I always respect before confirming a count:

Rule 1 — Wave 2 must retrace more than 50% of Wave 1.
If a correction I'm labeling as "wave 2" is very shallow (retraces less than 50%), I distrust that it's really a wave 2. The zones I watch most closely are the 61.8% and 78.6% Fibonacci retracements.

Rule 2 — Wave 4 must not retrace more than 50% of Wave 3.
In my experience, a healthy wave 4 stays below that 50% retracement of the prior wave 3. If it clearly exceeds it, I start questioning the count — it might be a complex correction, or I could simply be looking at the wrong degree of structure.

Rule 3 — Wave 4 must stay within the Fibonacci Speed Resistance Fan.
This is perhaps the most distinctive part of my methodology. I draw the Fibonacci fan from the origin of wave 1, and as long as wave 4 stays inside that fan, I consider the impulse still valid. If it clearly breaks out of the fan, the count loses credibility and needs to be rethought.

2. Wyckoff as structural confirmation

I overlay the Elliott Wave count with the Wyckoff schema — accumulation/distribution, levels like AR, UT, UTAD, Spring, and Test. I don't use them in isolation: I look for the Wyckoff phase I'm seeing (for example, a distribution) to be consistent with the Elliott wave I'm counting at that same moment. When both frameworks agree, my conviction in the thesis rises considerably.

3. "Slide then accelerate": my way of detecting the start of a wave 3

This is a methodology I developed myself to anticipate the start of a wave 3 (the strongest and most "profitable" wave of an impulse), instead of reacting once it's already underway.

It has two phases:

• "Slide" phase: ATR is compressed, below its moving average, and RSI is moving in a neutral zone (40-55). It's the "calm before the storm" phase — the market is indecisive, with no clear direction.
• Acceleration trigger: ATR expands sustainably across several candles, the MACD histogram also expands, a volume spike appears, and a relevant structure breaks (support, resistance, channel). When these four signals converge, I interpret that wave 3 is starting.

I automated this logic in TradingView with my own Pine Script indicator ("Wave 3 Detector - Slide & Accelerate"), which includes a cooldown filter to avoid repeated signals on the same move.

4. The "Exhaustion Doji" pattern

I identified a recurring pattern that tends to precede structural breakdowns: an indecision doji candle, with contracted volume, near a relevant resistance, followed by a "false impulse" candle that fools anyone reading it as continuation. When this pattern appears together with overbought or weakening RSI, and MACD already showing weakness, it tends to anticipate a turn.

I also automated it in a custom indicator ("Exhaustion Doji + MACD/RSI Confluence"), which includes a win-rate tracker for confirmed breakouts at N bars, to objectively measure how well it performs in practice.

5. Liquidity sweeps and liquidity zones

I complement the analysis with visually identified liquidity zones: levels where price touched and rejected repeatedly over weeks or months. When price returns to that zone with a wick that pierces the level but closes back inside, together with elevated volume and an RSI divergence, I interpret it as a liquidity sweep — often coinciding with the end of an Elliott wave.

6. How I use Claude (AI) as part of my process

Some time ago I added one more piece to my workflow: I use Claude, an AI assistant by Anthropic, as a kind of "second pair of eyes" on my own charts.

The dynamic is simple: I share screenshots of my TradingView charts — with my Elliott counts already marked, my custom indicators, my liquidity zones — and I discuss each decision with it in real time. It's not that the AI tells me what to do; on the contrary, I'm the one proposing the hypothesis (for example, "this could be the end of green wave 1") and Claude helps me check consistency: does this reading respect my own Fibonacci rules? Is there any tension with the count we'd been tracking before? Which levels should I watch to confirm or rule out the idea?

Several times this back-and-forth led me to correct a count in time. One concrete example: for several sessions I was working with a downward A-B-C structure, until reviewing in detail exactly where correction "B" had ended — millimetrically below the origin of wave 1, not above it — we realized together that it was actually a 1-2-3 impulse, not a simple correction. That kind of cross-checking, session after session, is where I find the most value in this process: it doesn't replace my judgment or my methodology, but it does act as a constant quality control on my own decisions.

I also use it for more operational things: documenting and keeping the state of my active count up to date, reviewing the logic of my Pine Script indicators when something doesn't trigger as expected, or simply thinking out loud about whether two scenarios are compatible with each other or contradict one another.

Going forward, in several of the follow-up posts I'll be sharing screenshots of these conversations directly, so the full process is visible: not just the final conclusion, but how it was built.

In summary

None of these frameworks is used in isolation. The golden rule of my methodology is multi-confirmation: Elliott Wave (structure and wave rules) + Wyckoff (market phase) + validated support/resistance levels + momentum (RSI/MACD/ATR) + volume. Only when several of these pieces converge in the same direction do I feel comfortable with a directional thesis.

In upcoming posts I'll be sharing live updates on my current SPX/SPY count, applying exactly this framework.

---

Disclaimer: The content of this blog is for informational and educational purposes only and does not constitute investment advice. The ideas expressed are the author's personal opinions, shared for entertainment and learning purposes.

 

 

Applying the Wave 4 Rule: How I Discarded an Invalid Count in Real Time

One of the pillars of my methodology is refusing to accept any Elliott Wave count that breaks my structural rules. In this post I want to walk you through, step by step, how I applied the wave 4 rule on SPX500 (TVC, 15m) during the July 17-20 session — and why I ended up discarding one of the variants I had been considering.

The rule I apply:

Wave 4 must not retrace more than 50% of wave 3, and it must remain within the Fibonacci Speed Resistance Fan drawn from wave 1's origin. If it breaks either condition, the count loses credibility and I look for alternatives.

In this case, the anchor for the fan and the Fibonacci retracements is the top of orange wave 2 (~7,583.62) — the point where the decline I'm subdividing internally begins.

---

Chart 1 — The base structure

Starting point: the decline from the orange wave 2 top (7,583.62), subdivided into five waves of blue degree (1-2-3-4-5). Within blue wave 3, an additional purple-degree subdivision appears (1-2-3-4-5), and in the final leg a smaller green 1-2 structure is forming, with green 1 completed near 7,440-7,446 and green 2 projecting as the ongoing corrective bounce.

At this first chart I hadn't yet drawn the Fibonacci or the fan — it's the visual base on which I'll validate (or discard) the following variants.

---

Chart 2 — First Fibonacci anchor adjustment

Here I apply the Fibonacci from the correct origin: the orange wave 2 top at 7,583.62 as point "1," and the low of the structure as point "0" (7,431.02). The 0.5 level falls at 7,507.32.

With this anchor, purple wave 4 stays below the 50% retracement and inside the fan — the count passes the filter.

---

Chart 3 — Fine-tuning point 0

A small precision adjustment: I move point 0 to 7,446.59, which shifts the 0.5 level to 7,515.10. This doesn't change the validity of the count, it just refines exactly where the structural low I'm using as reference sits.

Purple wave 4 still respects both conditions — retracement under 50% and containment within the fan.

---

Chart 4 — The variant I discarded

This is the most important exercise in this post. When I recalculated purple wave 4 using a different reference point, the retracement exceeded 50% of wave 3 and price closed outside the Fibonacci fan.

When a wave 4 breaks both conditions at once, I don't consider it a valid wave 4 under my methodology — no matter how good it "looks" visually on the chart. That's why I discarded this variant and kept the structure validated in charts 2 and 3.

---

Why this matters

This exercise isn't just a technicality. It illustrates something central to how I operate: a wave count isn't an opinion, it's a hypothesis that has to survive objective rules. When a structure doesn't meet them, I don't force it to fit what I "want to see" — I discard it and stick with the one that holds up.

With the structure validated, the current map stands as follows: blue wave 5 completing the decline from the orange wave 2 top, with green 1 marking the recent low (~7,440-7,446) and green 2 developing as the ongoing corrective bounce.

 

SPX/SPY Wave Tracking: From an A-B-C Read to a 1-2-3-4-5 Impulse

This is a chronological log of how my SPX/SPY count evolved over the course of about a week — from an initial complex A-B-C correction read, through several real-time updates, to the eventual reclassification into a clean 1-2-3-4-5 impulse. I'm publishing the full sequence, in order, so the process is visible, not just the final conclusion.

Each entry below reflects what I posted on X at the time, expanded here with a bit more context.

---

July 11 — Wave C (red) of 4 (black), starting down?

Original post: "Wave C (Red) of 4 (Black), maybe has started to the downside, after a nice ending diagonal of B (red), on Friday...."

At this point I was tracking a larger-degree wave 4 (black), with wave B (red) completing as an ending diagonal — a terminal structure typically found at the end of B-waves or 5th waves. Once that diagonal wrapped up, price started rolling over, which I read as the possible start of wave C (red) — the final leg down to complete wave 4 (black).

This was the starting hypothesis: a straightforward three-wave (A-B-C) correction, with C now getting underway.

---

July 14 — Wave 3 (orange) down?

Original post: "Wave 3 (orange) down?"

A few days later, the internal structure had refined itself. Zooming into the 5-minute chart, I could now see a smaller-degree 1-2-3-...-5 sequence forming inside what I'd been calling wave B — with green waves 1 through 5 visible, followed by an A-B-C move back up (orange wave 2), capped by a small ending diagonal.

The key question at this point: was the next leg down actually wave 3 (orange) — a full impulsive wave — rather than just another leg of a simple corrective C? The chart was starting to hint that the internal structure was more impulsive than corrective.

---

July 14 (later same day) — Ending diagonal of C of 2 complete? Wave 3 down starting soon?

Original post: "Ending diagonal of C of 2 up complete?. Wave 3 down starting soon?"

Later that same evening, the micro-structure clarified further: what looked like wave C of orange wave 2 had completed as its own ending diagonal (labeled i-ii-iii-iv-V on the chart), right at a resistance zone matching the earlier highs. That combination — diagonal completion + resistance rejection — was the first strong signal that wave 3 (down) could be starting very soon.

This is where the shift in thinking really began: instead of a simple corrective bounce, the up-move was behaving like a textbook orange wave 2, meaning the next leg down should be an impulsive wave 3, not just wave C of a correction.

---

July 15 — Multiple Liquidity Sweep (LS)

Original post: "Multiple Liquidity Sweep (LS)"

The following day, price returned to the same resistance zone (~7,610–7,625) not once but four separate times — each producing a liquidity sweep (LS1 through LS4/5) with wicks piercing the zone and closing back below it. This kind of repeated, multi-touch liquidity sweep at the same level is a strong technical signature: it shows the market repeatedly testing and rejecting the same price, consistent with a resistance zone that's absorbing liquidity before a reversal — exactly what you'd expect at the top of a wave 2.

---

July 17 — Reclassifying the count: Waves 1 and 2 are complete

Original post: "Everyone is waiting for Wave 4 to end so the SPX can continue its rally toward 8,000. But this chart suggests a different scenario. The price action over the past few months looks like the completion of Waves 1 and 2. If that's correct, the next move is likely a..."

This is where the count formally changed. Going back over the full multi-week chart, the wave B top (~7,631.59) and the wave 2 top (~7,631.43) landed millimetrically apart — with wave 2 topping just below the origin of wave 1, never overlapping it. That's the Elliott Wave overlap rule doing its job: it ruled out the simple A-B-C read entirely and confirmed this was a genuine 1-2 impulsive sequence instead.

Contrary to the popular narrative at the time — that the market was still finishing a corrective wave 4 before continuing its rally toward 8,000 — this chart pointed to a different scenario: waves 1 and 2 already complete, with wave 3 (down) the next expected move.

---

What changed, and why it matters

Looking back at this sequence, the shift wasn't a random change of mind — it was the accumulation of evidence across several days: an ending diagonal completing exactly where expected, repeated liquidity sweeps confirming the same resistance zone, and finally the overlap rule confirming that wave 2 could not have exceeded the origin of wave 1. Each new chart added one more piece of confirmation, until the full picture pointed clearly away from the original A-B-C read and toward a 1-2-3-4-5 impulsive structure.

This full back-and-forth — testing the count against price action, checking it against my own Fibonacci rules, and working through the logic with Claude along the way — is exactly the kind of process I described in my [methodology post]. I'll keep posting these updates in real time, warts and all, so the whole sequence stays visible.

---

Disclaimer: The content of this blog is for informational and educational purposes only and does not constitute investment advice. The ideas expressed are the author's personal opinions, shared for entertainment and learning purposes.