Factorio Forums

i have a 125hours game, but at around 45 hours my replay start to bug and does not play my game as I played.

There is absolutly no error message during my replay, and no error in my gamesave. my gamesave can still be played, but my replay is obviously broken.

here's a video where it start to look abnormal.
https://www.youtube.com/watch?v=6z0sIVe ... e=youtu.be

at 6m04s the laser turret seems to fire a bit too slow in my opinion.
at 6m08s, the first obvious thing that replay is broken : my research "degats cartouches 6" (i play the french version, it's the damage of shotgun) is changed without being fully completed. it look like the replay changed it just a few tick earlier than my real game. (when a research is completed, my game is auto-paused and i select a new research) so it look like a tick is missing somewhere (I manually put ressources into my labs, so i think it's not a problem of ressource getting stuck in my replay that have big consequence later but rather a tick problem)
at 6mn32s, my character seems to run rather fast compared to few second earlier.
at 6mn35s, the alien artefact are not picked, but it look like i tried to pick it, which is very suspicious, i'm pretty sure i grabed it in my real gameplay.
at 6mn41s, it's the critical point, a tree is blocking my path and the replay is definitivly broken. (i even die at 9m37s) which should be impossible in 0.14.23

here's my game file in the attachment of this post(it's in the top of the post). (my replay is included in it i guess).

To check the replay, you need to play it at x64 speed, it will be automatically paused when i launch the rocket (around 41hours ingame at x1 speed). (so you can actually go afk).
After the rocket launch, you have to press continue and wait approximately just a bit more than 4minutes at x64 speed to see the replay starting to bug.

I have no clue what is caused it, but during my game I did these things :
-changed the steam version to "no beta participation"(it was 0.14.23) to "beta participation : 0.14.23. (i tried to reproduce it, but with no bug found, so i guess it's very unlikely this can cause bug)
-chrome or an update, forced the game to alt-tab. (i tried to reproduce it using a script with autoit, to force every 2second the focus of another windows to make factorio alt-tab. my result were mitigate. I could not reproduce a similar replay bug. but i managed to make factorio a bit unstable after a while. (i could not load a savegame and i had to relaunch factorio), so forced alt-tab might be a problem for factorio.

I also suspect that replay can't support big factory.

Thanks for report. Replays can support big factory because this is the way multiplayer works in factorio - players actually exchange their replays as long as they play, and every such replay problem reported is very valuable because sometimes it helps to nail down multiplayer desync. I suspect it could be something biter related (kinda you stumbled at one during replay while it was dead in another universe).

A question - was it singleplayer or not? because other player actions are sometimes more prone to cause a "replay desync".

it was a singleplayer. I never played multiplayer in factorio.

i am almost sure it's not biter related. because it's impossible to explain why in the replay the research is changed without being completed.
at 6mn08 the text you can see "a changé la recherche en cours de degats cartouche 6 a vitesse de tir des balles 5" could be translated by "has changed the current research from shotgun shell damage 6 to bullet shooting speed 5". and you don't hear the "sound" that notify the research has been completed, so it's clearly not completed.

And i doubt i stumbled at a biter during the first 45hours of my replay, because in the video, i sucessfully manage to walk into a dense forest (in the video from 4mn to 6mn), i think getting stumbled into a biter earlier would have made me impossible to walk correctly into this forest and build my turrets laser to kill these biters spawner.
I may have stumbled after, but it would not explain the research bug.

Also, between 10mn18 and 10mn25 there's an obvious thing that there is a problem of "speed". i know the replay is completely screwed at this point but the character walk very very very slow. it's not normal, even without exoskelet the character should run faster in think. and it make me suspect once again that my feeling of character running too fast at 6mn33s is true.

edit : oh maybe i understand why the character is slow between 10mn18 and 10mn25, maybe because i run on a belt on the real game, that's why my movement speed could be decreased in the replay. i don't know how the replay is coded. It still does not explain why the replay went crazy at around 6mn on the video.