# Hidden content

Unlockables, debug features, button codes, cut and unused content. Every find says how sure it
is:

- *verified in game* - seen in an emulator (with the date, and the build or code used to get there)
- *from the code* - read in the disassembly, not seen running
- *from the data* - in the game's files or tables, whether or not the code uses it
- *from a guide* / *from fans* - a published source, not yet matched with the code

Addresses: `$bank:addr` CPU addresses (LoROM; file offset = bank*0x8000 + addr - 0x8000), RAM as `$addr`.
Optional patches that switch some of this on are in [PATCHES.md](PATCHES.md).

**Prior art (2026-11-03).** Two web searches (Japanese and English) found nothing about codes,
debug features or cut content for this game; The Cutting Room Floor, Data Crystal and GameFAQs were
not checked by hand. Everything below comes from the code.

## Contents
- [Multitap error screen](#multitap-error-screen) · [Searched and not found](#searched-and-not-found) · [How to see it yourself](#how-to-see-it-yourself)

## Multitap error screen
*from the code*; shown running with the detection forced (2026-11-03, headless emulator; not with a
real multitap in port 1)

At start-up the game reads both controller ports (`$01:F72E`) and tests each for a Super Multitap.
Bit 6 of `$0B03` means a multitap in port 2 (the intended setup); bit 7 means one in **port 1**. In that
case `$00:8CAA` calls `$0B:9E00`, which loads a text screen and stops in an endless loop:
"2Pコネクター側にマルチタップをつなぎ直してからRESETボタンを押して下さい。" (Reconnect the
Multitap to the 2P connector side, then press the RESET button.) Tiles `$3C:D800` (4bpp, 128), map
`$0B:9E86` (32x28 bytes), palette `$0B:A206`. Normal play never shows it. A release patch can't trigger
it either; it needs the hardware condition. The English build redraws the text (tools/errscreen.py, translation/text/multitap.csv); seen with the detection forced: `screenshots/hidden-multitap-error.png` (original), `screenshots/hidden-multitap-error-english.png`.

The cartridge also carries Nintendo's "SHVC MULTI5" multitap BIOS (ASCII credits at `0xd0b`, `0xf6d6`,
`0xf7a4`, `0xf84f`), which is what the port check comes from.

## Searched and not found
- **Button codes:** only three places read the pads (`$01:F5B2` in matches, `$03:A398` in menus, both
  four-port). The masks tested are Start (`$1000`), A (`$0080`), B/A/Start (`$9080`) and the d-pad.
  No Select, L/R or X/Y mask and no combination was found. (The scan sees AND/CMP masks only; a code
  read by bit shifts would escape it.)
- **Name entry / passwords:** there is no name entry.
- **Jump tables:** the per-wrestler table at `$01:87B9` (index `$0B23` & `$1E`, 16 slots) has 10 distinct
  handlers for the 10 wrestlers and the same default in the other six; `$00:E477` and `$00:EA70` have 10
  entries each. No spare roster slot. The HUD name table (`$08:8100`) holds exactly the 10 names.
- **Debug strings:** ASCII scan found only the header title, the MULTI5 BIOS credits and the HUD names.
- **Flags:** the walk's flag list (private notes) shows nothing set but unread that matters.
- **Dead code:** the walker leaves about 20 KB of unreached bytes in banks 0-3; most are indirect-jump
  code and data. Not read through.

## How to see it yourself
| Find | How |
|---|---|
| Multitap error screen | Real hardware: a Super Multitap in controller port 1. Emulator: patch `$01:F76F` and `$01:F775` to `EA EA` (NOP both branches) and boot. |