Hi
@antwerks
Can you run through what your mods are intended to do? I haven't really looked through it with a lot of intensity, but I'd be curious to know what of the original is working for you and where you want to make changes. As always, changes are good and I encourage people to run with kernels of ideas and make them work in a given situation or style!
Cheers,
mashume
This version is I think is a step up from the original wick script because it stops acting like a simple candle-shape detector and starts acting like a
filtered operational context engine. The original version mostly asked one narrow question: which wick was bigger, upper or lower, and did that exceed a threshold? That was fine for visual context, but it treated every wick event too similarly. This new script adds a lot more structure around
whether the wick actually matters. First, it keeps the core wick-imbalance idea, but then it
weights the imbalance by ATR-relative candle size, which is important because a dramatic rejection bar should count more if it is large relative to the instrument’s normal movement, not just large inside its own candle. Second, it overlays
8/20 EMA regime state and also adds an optional
EMA slope filter, so a bullish lower-wick event inside a rising bull-trend regime gets treated differently from the exact same candle inside a weak or deteriorating tape. Third, it adds a
location filter, which makes the signal more honest by asking whether price is near the fast or slow EMA rather than in the middle of nowhere. Fourth, it adds an optional
volume confirmation filter, which is a good upgrade because a wick on weak participation often means much less than a wick formed on sponsored volume. Finally, the biggest conceptual difference is the
persistence engine. Instead of only reacting to the current bar, it counts how many bullish or bearish wick events have occurred over a recent window and can require a minimum threshold before qualifying the current signal. That turns the script from “one bar did something interesting” into more of a
sponsorship / absorption / repeated-defense detector, which is much closer to how a swing trader should think about building evidence.
Operationally, I’d treat this script as a
decision-support layer, not a standalone trigger. The lower histogram gives you two related but different reads. The raw imbalance shows the immediate wick character of the current candle, while the weighted imbalance tells you how meaningful that wick was after adjusting for ATR-relative size. The smoothed trend line helps you see whether the wick-flow pressure is building in one direction over several bars rather than just flashing for one candle. The bull and bear pressure plots add another useful layer by showing whether repeated bullish or bearish wick events are clustering over your chosen persistence window. The cyan and magenta signal points are the most selective output, because they only print when wick structure, trend regime, location, volume, and persistence all align, depending on which toggles you leave on. In practice, for a swing-style workflow, I’d use it like this: if you see
positive weighted imbalance, rising smooth trend, stronger bull pressure than bear pressure, and a bullish absorption signal while price is in
bull trend or
repair, that is telling you sellers were absorbed and buyers are defending dips in a context where the tape is at least somewhat supportive. On the short side, if weighted imbalance is negative, bear pressure is dominating, the trend line is rolling over, and you get a bearish rejection signal in
bear trend or
bear cross, that is telling you rallies are being rejected rather than sponsored.
For actual decision-making, I would read it in layers. First ask:
what is the regime? Bull trend and repair are the only contexts where bullish absorption should really matter, and bear trend and bear cross are the only contexts where bearish rejection deserves real respect. Second ask:
is the imbalance real? That means looking at weighted imbalance, not just raw imbalance. Third ask:
is it persistent? If bull pressure has been building for several bars, that matters more than one isolated lower wick. Fourth ask:
is it sponsored? Volume confirmation helps here. Then ask the final trader question:
where is price? This script uses EMA proximity as location context, which is better than nothing, but you should still layer in your own higher-value references like prior-day levels, VWAP, premarket high/low, or round-number strike areas. In other words, this script is best used to tell you
how the market is behaving around a level, not whether the level itself is important. That’s the second-order distinction. The candle-wick engine can tell you buyers are absorbing or sellers are rejecting, but you still need to decide whether that behavior is happening somewhere meaningful.
My practical usage instructions would be simple. Put it in the
lower panel on the timeframe you care about, probably 15m if you’re using it tactically and higher if you’re using it for swing context. Start with all filters on so you understand its stricter form first. Watch how the weighted histogram behaves around pivots and trend pullbacks. Use the smooth trend line to judge whether wick pressure is building or fading. Treat the pressure labels as a short-term scoreboard, not a signal by themselves. Only pay serious attention to the bullish or bearish signal dots when they align with your broader structure read. If price is reclaiming an important support area, the script is showing bullish absorption, and the tape is in repair or bull trend, that’s the kind of confluence worth acting on. If price is bouncing into resistance, the script is showing bearish rejection, and the regime is bear cross or bear trend, that’s the kind of fade or avoidance context that matters. The main thing is not to use this as “dot means buy” or “dot means short.” The upgrade from the original is that this script is trying to measure
quality of wick behavior, not just existence of wick behavior, and that makes it far more useful as a swing-trader context tool.
