Limitations & Trust Boundaries

Limitations & Trust Boundaries

Every tool has a perimeter — a line past which what it shows you stops being reliable in the way you expect. This page maps that perimeter for Axiom BB Lite. These are not obscure edge cases. They are situations you will encounter during normal use if you configure the tool broadly enough, and some of them are situations you will encounter even with default settings if you do not know what to watch for. The goal here is not to discourage you from using the tool. It is to help you know when to lean on it and when to check it.

The blended band is not a signal

This is the most important limitation to internalize because it looks like the least important one. The blended band draws a smooth composite envelope on your chart. It has its own color, its own basis line, its own fill. It looks like a standalone analytical object.

It is not. The blend is a weighted average of whatever slots happen to be enabled with non-zero weights. It has no logic of its own. It does not discover anything the individual slots did not already contain. It compresses three readings into one, and that compression always hides more than it reveals.

What changes the blend:

• Enabling or disabling a slot

• Changing a slot's weight

• Changing a slot's timeframe, MA type, or length

• Hiding a slot (does NOT change the blend — this is the point)

• A slot resolving to na before it is zero-filled into the blend

The script also does not clamp weights to positive values. Negative weights are allowed, which means the blend can become subtractive or even collapse to zero when the active weights no longer sum above zero. That is real code behavior, not a documented "advanced mode," so if you are not intentionally experimenting with subtractive blends, stay non-negative.

Any decision you make based on the blended band alone should survive a check against the individual slots. If it does not, the blend is masking a disagreement that matters more than the summary.

The specific danger: A trader who watches only the blended band and treats it as an independent indicator will build habits around it — how far price typically extends past the blended upper, how the basis tends to hold during pullbacks, what a narrow blend means for the next move. Those habits are anchored to a specific blend configuration: specific slots, specific weights, specific timeframes. When the configuration changes — a slot disabled by accident, a weight adjusted and forgotten, a hidden slot quietly feeding different values — the habits no longer apply. But the blend still draws. It still looks like the same object. The colors are the same. The fill is the same. And the trader may not notice that the thing they are reading has shifted underneath them until the habits produce a decision that does not match reality.

Cross-ticker scaling is approximate

When a slot uses an Optional Ticker, the indicator adjusts the foreign symbol's BB values so they sit in your chart's price range. This adjustment uses a dynamic ratio based on the closing prices of both symbols at the slot's configured timeframe.

Where this works well:

• Instruments that tend to move together (e.g., SPY and QQQ, or two stocks in the same sector)

• Instruments with similar price magnitudes

• Short to medium timeframes where the price relationship is relatively stable

Where this breaks down:

• Uncorrelated instruments. If the two instruments move independently, the scaling ratio fluctuates in ways that reflect the changing price relationship rather than meaningful volatility structure. The overlay will still draw — it just will not mean what it appears to mean. Bands appearing to converge or diverge on your chart may be artifacts of the ratio shifting, not of actual volatility changes.

• Very different price magnitudes. Scaling a $50,000 crypto's BB onto a $50 stock's chart amplifies minor movements in the ratio. A 1% move in the foreign ticker produces scaled band shifts that look proportionally large on your chart. The overlay may appear volatile even when the foreign instrument is relatively calm. This is not distortion exactly — the math is correct — but the visual effect can mislead your intuition about what is happening on the foreign instrument.

• Different trading sessions. If the chart symbol and the foreign ticker have different trading hours, there will be periods where one has live data and the other does not. During those gaps, the scaling ratio may be stale or unavailable, and the overlay's behavior becomes unreliable. A US stock slot on a crypto chart will have dead periods every weekday evening and all weekend. See the data-gap section below for what this does to the blend.

How to check whether your overlay is useful: Compare the cross-ticker slot's bands to a standalone BB applied directly to the foreign symbol's chart at the same timeframe and settings. The absolute values will differ (they are in different price spaces), but the behavior should be similar — expanding and contracting at similar times, basis moving in the same direction. If the behavior matches, the scaling is doing its job. If the behavior diverges — the standalone BB is steady while the scaled overlay is jumping, or the directions of movement disagree — the scaling is introducing distortion that will mislead your read. Remove the cross-ticker slot or treat it as noise.

Blend behavior when a slot resolves to na

Before slot values enter the blend, the script wraps them in nz(). So if a slot resolves to na, that slot's contribution drops to zero while its weight is still counted in the denominator of the weighted average.

What this looks like on the chart: The blended band contracts or shifts toward the remaining slots' values, but more than you would expect from simply removing one slot. The shift is larger because the missing slot's weight is still dragging the average toward zero rather than being excluded.

Example scenario:

• Three slots, equal weight (33.3 each). Slot 3 uses SPY as its foreign ticker.

• Slot 3 resolves to na for some reason — invalid symbol, unavailable feed, or another unresolved state.

• The blend now averages Slot 1's values, Slot 2's values, and zero — all with equal weight. The result is pulled about one-third of the way toward zero compared to where it would be with only two slots.

What to do about it:

• If a cross-ticker slot is likely to go unresolved in your workflow, consider setting its weight to zero during those periods, or disabling it entirely. A weight of zero removes the slot from the blend calculation completely.

• If you notice the blended band contracting unexpectedly, check whether all enabled slots appear to have valid data at the current time.

This code path is verifiable: unresolved slot values get zero-filled into the blend. What this file does not prove on its own is every real-world condition that leads to that unresolved state, so treat any specific market-hours example as a plausible failure mode, not as a guaranteed one.

Same-timeframe stacking is not MTF stacking

Setting all three slots to the same timeframe with different MA types or lengths is a valid use of the indicator, but it is not a multi-timeframe stack. It is a multi-parameter comparison on a single timeframe.

The blend in this configuration averages three variations of the same timeframe's data. The result may be useful as a consensus center across MA types, but it does not carry the cross-timeframe meaning that the indicator is designed to provide. Band expansion on one slot and contraction on another in this setup means the MA types are responding differently to the same data, not that different timeframes are showing different volatility states.

If you want to compare MA types at a single timeframe, this works. Just do not read the blend as if it represents cross-timeframe consensus.

On Bar Close applies to everything

The On Bar Close switch is global. It affects:

• All three slots' BB computations

• The cross-ticker scaling ratio (if any slot uses a foreign ticker)

• Consequently, the blended band (since it is derived from the slots)

There is no way to make one slot use confirmed values while another uses building-bar values. If you want confirmed data on your longest-timeframe slot but responsive data on your shortest-timeframe slot, this indicator cannot do both simultaneously. You choose one mode for the whole stack.

For a full treatment of this tradeoff, see MTF & Repainting.

What the indicator does not tell you

Direction. The indicator shows where price sits relative to the bands and basis. It does not tell you where price is going. Price above the basis does not mean "bullish." It means "above the smoothed center for this timeframe." In a strong downtrend, price can cross above the basis repeatedly on short pullbacks without the larger move reversing. In a strong uptrend, price can stay above the basis for hundreds of bars without that state predicting anything about the next bar. The basis position is a description of the present, not a forecast.

Significance of band touches. Price touching the upper or lower band is a description of where price is relative to recent volatility. It is not a reversal signal. The textbook interpretation ("price at the upper band is overbought") is a simplification that markets violate constantly — in trending conditions, price can ride the upper band for extended stretches, touching it bar after bar without reversing. If you treat every band touch as actionable, you will get a lot of actions and not many of them will be useful. The bands tell you the statistical range. Whether a touch at the edge of that range matters depends on everything else you know about the market, not on the touch itself.

Whether the blend configuration is good for your context. The indicator has no opinion about whether your choice of timeframes, weights, or MA types makes sense. It computes whatever you configure. If the configuration is poorly chosen — timeframes too close together, weights skewed for no reason, a cross-ticker overlay on an uncorrelated instrument — the indicator will still draw, and it will still look authoritative. The chart will not warn you that your setup is producing noise. It will present that noise with the same colors, the same fill, and the same visual authority as a well-chosen configuration. The quality of the output depends entirely on the quality of the setup.

Whether you should act. This bears repeating because it is the most common overreach: the indicator is a context tool. It shows you volatility structure across timeframes. It does not generate entries, exits, or positions. It does not tell you what to do. If you find yourself treating the blend or the alignment alerts as trade triggers without additional confirmation from your own process, you have crossed the line from using the tool for what it offers into depending on it for something it was not built to provide. The indicator gives you a better read of the room. You still have to decide what to do in the room.

Trust checklist

Before relying on a read from this indicator, run through these:

• Is On Bar Close on? If not, am I looking at live data or historical data? (Only live data is meaningful with it off.)

• Do I know which slots are enabled, and are any of them hidden?

• Do I know what the blend weights are?

• If a cross-ticker slot is active, is that symbol's market currently open and providing data?

• Am I reading the blend as a summary, or am I treating it as an independent signal?

• If the blend disagrees with the individual slots, do I know which slot is causing the disagreement and why?