WIFI should always delay 1 frame before it conducts. But, actually, it don't, it's random. Sometimes it delays and sometimes it don't. Wait until there's no spark anywher, press A several times, and sometimes it delays and sometimes it don't.
jacksonmj now made it able to delay 2 frames... Oops
if ((r&0xFF)==PT_SPRK && parts[r>>8].ctype!=PT_NSCN && ((parts[r>>8].life==3 && i < r>>8) || (parts[R>>8].life==4 && i > r>>8)))
I think this is better. When the conductor's particle order is after WIFI its life should be 4 (after all it will be 3), if it is before WIFI its life should be 3 to conduct (after all it will be 3)