Card: controlled test — f16 MTP head is a wash (~0.83 both); mtpF16 redundant, use base

README.md

@@ -79,19 +79,19 @@ the **output head** (lm-head) and the **MTP draft head**:
 |---|---|---|---|---|
 | `…-STRIX-embQ8.gguf` | 21.8 GB | 4-bit | 4-bit (Qwopus donor) | **fastest** — the recommended default |
 | `…-STRIX-embQ8-headQ6.gguf` | 22.2 GB | **Q6_K** | 4-bit (Qwopus donor) | better output quality |
-| `…-STRIX-embQ8-mtpF16.gguf` | ~22.4 GB | 4-bit | **f16** (BF16 base-27B donor) | experimental high-precision head |
+| `…-STRIX-embQ8-mtpF16.gguf` | ~22.4 GB | 4-bit | **f16** (BF16 base-27B donor) | experiment — no acceptance gain vs base |
 
 - **Output head** (`output.weight`): 4-bit vs `Q6_K`. Q6_K is a touch more faithful at ~5–7% slower
   decode (trade measured on the [27B sibling](https://huggingface.co/plunderstruck/Qwen3.6-27B-MTP-ROCmFP4-GGUF)).
 - **MTP draft head** (`blk.96`): the base & headQ6 variants graft a **4-bit** head from a **Qwopus-27B**
   donor (a Claude-Opus distill); the **mtpF16** variant keeps the head at **f16** from a **BF16 base-27B**
   donor — mirroring [PiehSoft's high-precision-head choice](https://huggingface.co/PiehSoft/Qwen3.6-40B-Deckard-MTP-Q6_K).
-  **Measured MTP acceptance (mtpF16):** ~0.77 cold, but **~0.80–0.89 sustained (avg ~0.85)** across a warm
-  6-turn coding session — right in PiehSoft's reported BF16-head range (85–100% sustained). The
-  4-bit/Qwopus-donor head hit ~0.95 on a single short test, but there's **no warm multi-turn number for it,
-  so this is not a controlled comparison** — both heads are solid (~0.8–0.95). Pick **mtpF16** for the
-  PiehSoft-style high-precision head, or **base** for max speed/simplicity. (Earlier notes called mtpF16 an
-  underperformer based only on its *cold* number; the warm sustained acceptance corrects that.)
+  **Controlled warm comparison (identical 6-turn session, f16/f16 KV both):** 4-bit/Qwopus head **~0.82
+  sustained**, f16/base-27B head **~0.84 sustained** — a **wash** (the gap is within turn-to-turn noise; the
+  earlier 0.95-vs-0.77 split was just cold-vs-cold noise). **Conclusion: the f16 MTP head does not
+  meaningfully improve acceptance** — neither head precision nor donor moved it. So **`mtpF16` is effectively
+  redundant with `base`** (same speed, ~0.6 GB bigger). **Use `base` (or `headQ6` for output quality);
+  `mtpF16` is kept only as a documented negative-result experiment.**
 - **Q8 embeddings** (not f16): the source is Q8_0, so f16 would be fake-f16 bloat — Q8 matches the
   source precision exactly.