Ghost should not spawn at the same spot over multiple rounds when sv_neo_ghost_spawn_bias is enabled

[kassibuss]

Build Info

20260201_bc59026

Description

When using sv_neo_ghost_spawn_bias is enabled, then the ghost can spawn at the same position multiple times in a row. This is the opposite of the intended behavior of the command (and the plugin that it is based on).

Might be caused because of the removal of the weighted ghost spawns

To Reproduce

1. Enable Ghost Bias
2. Observe the ghost spawns

Expected behavior

Ghost should switch its spawn position after two rounds of spawning in the same position

Actual behavior

Ghost can spawn at the same position over a few rounds. For example, testing the spawn bias on Oilstain, the ghost spawned at the same position for 4 rounds in a row.

Operating System

• Windows
• Linux

Version/Distro

10

Machine's CPU

No response

Machine's GPU

No response

GPU's driver

No response

Build's compiler

No response

Additional context and Screenshots

No response

[Rainyan]

Do I understand correctly that you wish for the bias behaviour to iterate through all the available ghost spawns in a set sequence from first to last, and have the teams play sides for each such spawn, and finally wrap around and start again from the beginning of the sequence?

Currently, sv_neo_ghost_spawn_bias ensures that for an even round number m_iRoundNumber, we use the same ghost spawn as was used on the previous round, so for example round 2 will use the same spot as round 1. But since round number 3 is odd, it will pick a new completely random spawn location to be used for the next pair of rounds, and that random spawn could end up being a repeat.

Tested with the following, and monitoring the Spawned ghost at coords: messages in console, and I can verify the bias does work "correctly" for what I described:

sv_neo_ghost_spawn_bias 1;
neo_bot_quota 2;
sv_neo_preround_freeze_time 1;
mp_chattime 1;
neo_score_limit 99;
neo_ctg_round_timelimit 0.1;

map ntre_oilstain_ctg;

But whether that is plugin-parity, or if we should instead go in a sequence, I don't know.

Also worth noting that for NTRE, spamming neo_restart_this 1 can't be used for the bias testing, because it resets m_iRoundNumber which is used for the odd/even logic.

[Agiel]

Soft as Hell's plugin that we've used in the past randomizes the list of spawns when the map is loaded and then cycles through them. I think that's ideal for comp because you're guaranteed to play all the spawns, but if we want it to remain completely random it should at least not stay on the same spawn more than twice in a row when bias is enabled (even if that technically makes it less random).

[kassibuss]

Ideally, it should switch spawns after two rounds no matter what (or at the very least reduce the chance of it spawning at that spawn again? I am not a code wizard so I dont know if that would be worthwhile.), even if it does make it techinically less random like agiel said.

On the other hand, it also might just be the issue of weighted ghost spawns on certain maps. For example, we played a scrim on Grid and the ghost spawned at either inside the central tower at mid or at the upper portion of it for majority of the match.