`sds scenarios --lances N` now writes a V2 scenario with a per-unit `force` path, and the host writes the same path into the result for assertions. An 8v8 two-lance match made separate `Lance 1` and `Lance 2` force decisions instead of one `all` bucket. Co-Authored-By: Claude Code <noreply@anthropic.com> Change-Id: I4148ae3b16cda446ab7baed78f350da58a44a0b2