code wiki / _hdl_build / nx_game_depth.nx
nx_game_depth.nx
buildroot/runtime/_hdl_build/nx_game_depth.nx
about
nx_game_depth.nx -- the MECHANICAL-DEPTH critic. nx_game_critic measures whether a frame LOOKS rich;
this measures whether the GAME PLAYS deep -- because a beautiful frame on a trivial click-loop is still
a toy. Depth is not asserted from the rules; it is MEASURED from OUTCOMES: run K distinct strategies and
see whether they DIVERGE. A shallow game funnels every strategy to the same win (spread ~0, nothing is
at stake). A deep game rewards good play and punishes bad -- strategies spread out, and a careless one
can LOSE. That divergence is depth you can measure, and it is exactly what the current click-to-sell
loop LACKS (you always win). Verdict SHALLOW/BASIC/DEEP. LIB. license_tier: ORIGINAL
dependencies 1 imports · 4 importers
imports: nx_syscalls.nx
imported by: nx_breeder_arena_gate.nxnx_drift_engine.nxnx_game_depth_gate.nxnx_game_redteam_gate.nx
structs
| none |
consts
| 10 | const GD_N: i64 = 4 |
| 11 | const GD_M_SPREAD: i64 = 0 // outcome spread across strategies (permil of the best) |
| 12 | const GD_M_STAKES: i64 = 1 // are BOTH win and loss reachable? (real consequence) |
| 13 | const GD_M_DIVERSE: i64 = 2 // distinct outcome buckets / K (strategies genuinely differ) |
| 14 | const GD_M_DET: i64 = 3 // determinism: fraction of strategies that REPLAY identically. Skill games |
| 17 | const GD_SHALLOW: i64 = 300 |
| 18 | const GD_BASIC: i64 = 550 |
functions
| 23 | func gd_score(scores1: *i64, scores2: *i64, results: *i64, k: i64, out: *i64) -> i64 |
| 66 | func gd_cap(v: i64, target: i64, weight: i64) -> i64 called by 1: gd_quality |
| 76 | func gd_quality(out: *i64) -> i64 |
| 87 | func gd_verdict(score: i64) -> *u8 |