The Pyramid Navigator — the geometry of adding to a position

13 January 2025, just after half past three in the afternoon, the E-mini S&P 500. Two operators, called Thomas and Marco in the book, have the same morning behind them: the same pre-session analysis, the same convergence in the funnel, the same A-plus classification at the same level. Both go long. Both are stopped out — the market had other plans, it happens, that is what the stop is for.

By evening Thomas stands at −0.2 R. Marco stands at −2.0 R, in dollars: 1,400 lost instead of 200. No different market, no different analysis, no different setup. The entire difference is a single click: three minutes after his stop hit, Marco bought again three points lower. Full size, stop by feel, reason: "The market is recovering, I'll win the loss back."

The same gesture — enlarging a position — made Paul Tudor Jones a legend of risk management and dismantled Marco's trading day. That is the uncomfortable point of this tool: whether adding to a position builds a career or wrecks an account is decided by no forecast and no experience. It is decided by a calculation that fits on a beer mat — and that almost no one does before the click. This tool makes it visible, tranche by tranche.

Why this instrument

The WVPO method treats scaling as part of the risk-management pillar, not as a matter of style. Book III, chapter 35A sets out the three-tranche architecture for it: one tranche trades the hypothesis, two verify it, three pyramid the verified hypothesis. The word behind it is old — pyramiding, or scale-in — and so is the tradition behind it: you add to winners, never to losers. The book names this dividing line: anti-martingale.

What happens mechanically is pure arithmetic. Your cost basis is not a price but a weighted average: the sum of price times size across all tranches, divided by the total size. Every add sits above the old average — otherwise it would be averaging down — and therefore pulls it upward, trailing the current price. The distance between price and break-even is the only cushion the whole position has. And that cushion shrinks with every add.

From this follows the term this tool revolves around: the tipping point. That is the pullback in points your position can take at most before it stands underwater as a whole. As long as the tipping point lies outside the normal market noise — outside the pullback depth that a healthy trend produces regularly, without anything breaking — the pyramid can breathe. Slip it into this noise band, and an ordinary pullback is enough to destroy the trade: no structural break, no event, just statistics.

Whether the tipping point stays outside or slips in is decided by the geometry of the pyramid: how large are the follow-on tranches relative to the first, and how close to the high are they bought? Chapter 35A casts this into a four-condition definition, the risk-free add — primary tranche at least +1 R in profit, add stop below the structure, add size at most 50 % of the primary tranche, orderflow verification at the add level. Whoever meets all four builds out a position without ever exceeding the planned risk. Whoever skips one builds not a pyramid but a tower on sand.

The monster position, then, is not a question of courage. It is an arithmetic problem with four side conditions — and that is exactly what this tool lets you play through, with the rules from Book III and the sizing fundamentals from Book II, chapters 29 and 30.

Begin at the construction site: open the primary tranche at the breakout — and watch what every add does to the copper-colored break-even line.

The guided tour

Seven handles, from the first buy to your own account size. Each carries an insight the text alone cannot deliver.

1 · The construction site — the break-even trails the price

What: The construction site walks you through a model trend leg: base, breakout, two advance legs, two structural add stations. At the first station you open the primary tranche (at the default scale, 25 micro contracts at 5,800, stop 5,792 — eight points, one R). At stations A1 and A2 you decide: add or skip. The copper-colored line on the chart is your weighted cost basis — it moves with every add.

Why: You can read the formula in a minute. Watching the line jump is something else: after the first add at 5,812 the break-even no longer sits at 5,800 but at 5,803.9. After the second at 5,824 it sits at 5,806.7. The price has risen by 32 points, yet your cushion is only 25.3 points — the rest has gone into the adds. That is the fundamental insight of pyramiding: building out a position means spending cushion.

Exercise: Build the rule pyramid — take both adds, leave the geometry at "≤ 50 %" — and note the break-even after each step: 5,800 → 5,803.9 → 5,806.7. At the high, check the "cushion" line in the balance: 25.3 points. That single number is your tipping point.

2 · The balance magnifier — what is actually at stake

What: To the right of the chart your pyramid stacks up as a block diagram: block height equals buy price, block width equals size. Tap a block and you get its file — size, price, profit contribution in dollars, share of the position. Below it the aggregate balance runs in real time: cost basis, cushion, open result, worst case at the current stop.

Why: "I'm nicely in profit" is a feeling. The magnifier replaces it with numbers that carry a file reference — and they contradict the feeling regularly. At the end of the model run, for instance, the aggregate stands at +3.4 R while the third tranche, taken on its own, is 45 dollars underwater. Both at once is normal: the early tranches carry the late ones. Whoever has seen that once understands why a pyramid must grow heavier from bottom to top — and not the other way around.

Exercise: Run to the end and tap through all three blocks. Find the tranche with a negative contribution and read the positive aggregate next to it. Remember the picture — it is the argument against every heavy buy near the high.

3 · The geometry lab — the tipping point against the noise

What: Two sliders, one scale. The ratio sets the size of each follow-on tranche (50, 100 or 150 % of the previous), the placement sets the location (structural at the HVN upper edge and former VAH — or tight to the high, as the momentum impulse suggests). The scale below shows your tipping point against the red-marked noise band: 8 to 14 points, derived from the pullback to the former VAH and the structurally typical pullback to the last higher low.

Why: The relationship is nonlinear, and the table in the lab shows all six combinations at a glance: built structurally and degressively, the tipping point sits at 25.3 points — far outside the noise. Built tight and rising, it sits at 12.5 points, in the middle of the band. The same three buys, the same market, and one pyramid survives an ordinary pullback, the other does not. In passing, the table marks which combination meets the 50 % rule from chapter 35A: exactly one ratio.

Exercise: Set "150 % (rising)" and "tight to the high" and read the traffic light. Then find in the table the only row with the note "✓ 50 %". The distance between these two rows — 12.5 against 25.3 points — is the price of greed, in points.

4 · The stop trail — the invisible part of the pyramid

What: The stop is yours: drag the red line with the mouse, with the arrow keys or the buttons beneath it. After every add the trail recommendation from condition 2 appears — stop below the last higher low. The balance computes the worst case at your stop live.

Why: After the first add the worst case at the original stop stands at −2.2 R. One click on "trail", and the same line shows +0.2 R. That is half the core of the risk-free-add definition: it is not the add that keeps the risk small, it is the trailed stop. And the tool also knows the opposite error: whoever pushes the stop too tight below the price out of fear gets a warning before the next candle — and may let himself be stopped out in the noise, textbook-correct with a profit and wrong all the same.

Exercise: Provoke both errors once on purpose. First take an add and skip the trail — watch only the worst-case line. Then drag the stop up to just below the price and let it run on. The two error messages you see are the two most common ways to lose a pyramid.

5 · The tipping-point guess — intuition against the average

What: At the high of the leg the tool stops and asks a single question: how many points of pullback can your position take before it is underwater? You set your guess on the slider. Then the tool computes: high minus break-even, to the decimal place.

Why: Almost everyone guesses wrong, and the direction of the error is telling. Whoever guesses too low is reading the cost basis of the last tranche — the most expensive, the most present. Whoever guesses too high is reading the first entry and forgetting that the average has long since moved up. Both errors have the same effect in the market: you manage a position whose geometry you do not know.

Exercise: Give your guess before you read on, and note the deviation. Then build a different pyramid (different ratio, different locations) and guess again. If your second deviation is smaller, the eye has begun to see weighted averages.

6 · The side by side — the same leg, two truths

What: Two charts, the same market, three buys each. On the left the rule pyramid: adds at structure, at most 50 %, stop trailed. On the right the stacking trap: full tranches near the high, stop stays at the starting point. A shared time slider runs both through the advance and the pullback that follows — 12.5 points from the high, normal noise, no structural break. Below it the Marco file with the numbers from chapter 35A.

Why: The moment that carries this mechanic sits at the high: there both sides show almost the same number — +5.44 R open against +5.25 R. The trap does not look wrong, it even feels larger, a hundred contracts against forty-three. Then comes the pullback, and the truths part: the operator stands at the low at +2.75 R open, his trailed stop secures +1.79 R, whatever comes next. The stacking side stands at −1.0 R — and on its stop, left in place, hangs −14.75 R. The planned risk of this trade was one R. The trap almost multiplied it fifteenfold, without a single buy looking like a mistake.

Exercise: Drag the time slider to the high and read both sides twice: once the "open" line, once the "worst case" line. The first line differs by 0.2 R, the second by 16.5 R. Ask yourself honestly which of the two lines you have your eye on at the high of a real trade.

7 · My scale — the tipping point in your own currency

What: You set the account, risk percent and instrument (E-mini or Micro) at the top; the fixed-fractional formula from Book II translates everything into your contracts and dollars. The setting stays saved on your device.

Why: The R mathematics is scaleless, the buildability of a pyramid is not. With 100,000 dollars and one percent risk the E-mini carries exactly two contracts — no degressive three-step pyramid can be built from that, half of two is one, half of one is zero. The same numbers in the Micro give 25 contracts and full granularity. And whoever, with a small account and 0.25 % risk, is handed zero contracts reads the most honest line of the tool: then no tranche can be built — that too is a result, not an error.

Exercise: Set your real account and your real risk and check whether a degressive three-step pyramid can be built at all in your instrument. If not, that is no invitation to enlarge the tranches — it is the argument for the Micro.

Limits & honesty

This tool computes exactly — on a model. The limits stand here, not in the fine print.

The trend leg is constructed. Every candle is anchored by hand so the teaching geometry works out exactly: add stations at clean structure, a pullback that ends precisely in the noise band. Real trends are messier, real higher lows more negotiable. The model proves the arithmetic of the pyramid, not the course of your next trade.

Condition 4 is named, not checked. The risk-free-add definition requires an orderflow verification at the add level — absorption or stacked imbalances. A model without tick data cannot perform this check; in the live trade it is the part that distinguishes a structural level from a reason to buy.

Slippage, commissions and gaps are missing. The worst case at the stop assumes execution at the stop price. A gap through the stop, thin liquidity, a fast market — all of it makes the math worse in practice, never better. The aggregate R values of the tranche distributions from chapter 35A are also model assumptions that the book itself puts out for your own verification per instrument.

The noise band is a convention. Eight to fourteen points of normal pullback hold for this model leg with its one R of eight points. In your market "normal" means something else — the method of setting the tipping point against the typical pullback depth carries over; the numbers do not.

And the fundamental point: The navigator produces no signals, no forecasts and no recommendation to pyramid at all. It makes a calculation visible. All R and dollar examples are simulations in the sense of our risk notice.

Case vocabulary

The two anchor terms of this post, verbatim from the WVPO glossary (translated from the canonical German):

Risk-free add — "A secondary tranche that never lifts the trade's aggregate risk above the initial 1 R level. Four conditions, all mechanically met: primary tranche ≥ +1 R in profit, add stop below the last M5/M15 higher low, add sizing ≤ 50 % of the primary tranche, filter-level-4 verification at the add level — if one is missing, the add is discarded." (Book III, ch. 35A)

Anti-martingale — "The dividing line of scaling: you add only to a winning position (primary tranche ≥ +1 R). The add to a losing position (averaging down) is martingale — permitted in no topology." (Book III, ch. 35A)

If you want to practice the prior stage: the position-size calculator determines the primary tranche with which every pyramid begins. And if you want to know what an edge does over a thousand runs with drawdowns before you leverage it: Monte Carlo — 1,000 futures of the same edge is the next step after this tool.