Curious what part of the EBL documentation you are referring to, so that it can be corrected in the future. Anyway, I did a deep dive on what you posted and this is what came up, which makes sense and appears the forum is correct. Have a look.The key technical explanation was captured in the search snippet. Here's a clear breakdown of what's actually correct:

The Documentation is Wrong (or at best, misleading) — The Forum Consensus is Right

Your intuition is correct, and so is the forum thread. Here's the precise technical explanation:

How the Lag Filter Actually Works

The ECM reads the TPS sensor and lag-filters it every 12.5 milliseconds (80 times per second). The filter coefficient controls how quickly the filtered output chases the raw input. Third Generation F-Body Message Boards

The math that matters: Delta TPS = Current TPS − Lagged (filtered) TPS

This delta value is what drives AE. So:

• Smaller coefficient → slower filter → output lags further behind input → LARGER delta → more AE, sooner and longer
• Larger coefficient → faster filter → output catches up quickly → SMALLER delta → less AE

A smaller filter coefficient causes the output of the filter to be slower to catch up to the input. This increases the duration of the AE. With a slower filter (smaller coefficient), the delta will also be greater, providing a larger volume of AE. Third Generation F-Body Message Boards

So What's Wrong With the Documentation?

The documentation statement — "A smaller value will be a greater lag" — is technically correct on its own. A smaller coefficient does produce greater lag. The problem is the phrasing implies this is a bad thing or means AE comes in slower, when in fact greater lag = larger delta TPS = AE kicks in harder and faster from the tune's perspective.

The forum saying "lower the value to make AE come in sooner" is correct in practice — lowering the coefficient increases the lag between raw and filtered TPS, which inflates the delta, which triggers a stronger AE response.

For Your XDF Description

Here's how I'd word it accurately:
AE TPS Filter Coefficient (vs. Coolant Temp): Controls the lag rate of the TPS filter used to calculate delta TPS for AE. A smaller value = more lag = larger delta TPS = stronger and longer AE. A larger value = less lag = smaller delta = weaker, shorter AE. Affects both the magnitude and duration of TPS-based AE.
Smaller filter values result in larger delta TPS values, impacting the AE threshold, volume, and duration. Scribd

Your intuition that "higher number = more filtering with greater lag" is actually backwards — it's a counterintuitive parameter. Higher number = faster response (less lag), lower number = slower filter (more lag = more AE). The forum thread has it right

 

There seems to be some confusion in this thread because the documentation and the forum comments are describing two very different parts of the AE chain. When you separate those, everything lines up cleanly. The documentation is referring to how the ECM uses the filtered TPS to calculate the magnitude and duration of TPS‑based AE. A smaller coefficient creates more lag, more lag means the filtered TPS moves slower, a slower filtered TPS produces a smaller delta TPS, and a smaller delta TPS results in weaker AE and shorter AE duration. That part of the documentation is absolutely correct. When tuners say “lower the filter to make AE come in sooner,” they’re talking about AE enable timing, not AE magnitude. AE enable is influenced by the difference between the raw TPS (which moves instantly) and the filtered TPS (which lags depending on the coefficient). When the filter is slower, the raw TPS jumps, the filtered TPS lags, the difference between them grows, and that difference crosses the AE enable threshold earlier. So AE can start sooner, even though the AE pulse itself is smaller and shorter. Sounds a tad contradictory yes, but only if you assume the same TPS value is used for both AE enable and AE magnitude, but GM doesn’t do that, and the EBL is fundamentally GM code with patches layered on top. AE enable is influenced by the difference between raw and filtered TPS, while AE magnitude and duration are based on the filtered TPS delta...

- Rob