With my recent (successful) attempt at a world record for Space Tactics on M.A.M.E., I have learned some very interesting things about INP's and I would like to share that with you all and voice my concerns.
During my practice sessions I was able to get over 300,000 points, but I found that playback of the INP file was erratic and not at all a representation of the actual game play. I tried several attempts, each with several versions of MAME. I tried WolfMAME: 0.106, 0.160, 0.166. I tried a much older versions of MAME as well (MAME 0.32) but that did not work either.
When playing back my recorded INP, the game seemed to go off on a tangent, doing it's own thing. My inputs were useless, and the was over before getting any where near my score. In fact, the highest score I was able to create via playback was just over 20,000 points.
I did some research and asked a few of the veteran players. What I found was that this particular ROM (spactics) has a known problem known as "de-syncing". It is not clear what causes the playback to de-sync, but it is evident in
all versions of MAME. Even INP's that are posted on MARP will not playback.
Does this matter for my score? Not really, as I provided a full video of me playing the game. I included the INP so that the keeners could open it up and check that the dip switch settings were accurate, and that an acceptable version of the ROM was being used, etc.
Why am I bringing all this up? I am bringing this up because I care about the integrity of the scoreboard and I believe that the score I beat out of first place is invalid and should never have been entered into the system.
https://www.twingalaxies.com/attachme...8&d=1447466939
There is no way that a TG referee was able to playback the INP from that submission, using WolfMAME 0.106. I do not believe that performance was reviewed via INP.