Save Games; how do they work?

[caesarbear]

It’s just math guys.

[Zylon]

@Otagan

This demonstrates the fidelity of the engine…

The word is: Deterministic.

[Scuzz]

Is there any reason why some games will let you name a save and others wont? I hadn’t thought of the screensave idea but that is always a nice reminder of where you were when you saved, but it would take extra memory.

[Chris_Gwinn]

@Fugitive

Just plain old sloppiness in some cases. I was shocked to discover that The Witcher was taking up nearly 7 gigs in save games, of up to 20 megs each (yes, I am a hateable savebaby). Looking inside the save file, there’s a lot of text of things like item descriptions and notes, in multiple languages. WTF is that doing in the save file?!? It looks like they simply iterate over every item you have and dump its state to the file, but the item state includes everything about that item including its descriptions, and they don’t differentiate between mutable and immutable data, so it writes out a lot of redundant data.

That’s why I stopped playing The Witcher. It apparently decided to parse the save games when displaying the available list, and getting the list of games available for loading would take over a minute on my computer.

[WarrenM]

@Zylon

The word is: Deterministic.

[Chris_Gwinn]

@Scuzz

Is there any reason why some games will let you name a save and others wont? I hadn’t thought of the screensave idea but that is always a nice reminder of where you were when you saved, but it would take extra memory.

The default state of any feature is “Not Implemented”.

[Blips]

@Chris Gwinn

That’s why I stopped playing The Witcher. It apparently decided to parse the save games when displaying the available list, and getting the list of games available for loading would take over a minute on my computer.

So… delete some saves? You can manually go into the save game folder and clear stuff out you know.

[cliffski]

@Chris Gwinn

That’s why I stopped playing The Witcher. It apparently decided to parse the save games when displaying the available list, and getting the list of games available for loading would take over a minute on my computer.

That is really bad and sloppy coding, hopefully even a crappy indie like me wouldn’t be responsible for that. Too many coders now are sloppy, treating memory and CPU like an endless bucket they can never run out of. Even if something runs like lightning on your dev pc, you should always be testing worst cases on minimum spec.

Having said that… Fitting GSB on an iPad has been a nightmare :)

[Chris_Gwinn]

@Blips

So… delete some saves? You can manually go into the save game folder and clear stuff out you know.

I tried to. I don’t remember what made that not be a good solution. I think it also kept track of the game files in separate file or something.

[Chris_Gwinn]

@cliffski

That is really bad and sloppy coding, hopefully even a crappy indie like me wouldn’t be responsible for that. Too many coders now are sloppy, treating memory and CPU like an endless bucket they can never run out of. Even if something runs like lightning on your dev pc, you should always be testing worst cases on minimum spec.

Having said that… Fitting GSB on an iPad has been a nightmare :)

I assume it was IO issues - probably did something funky pulling the screenshot out of every save file in my directory.

[Otagan]

@caesarbear

It’s just math guys.

Undoubtedly true, but when I look at replays for other games and see how large those files can get for 20 minute long games with two players and vastly smaller unit counts, you have to have an appreciation for the fact that someone out there can distill a game down to its absolute basics. I’ve spent a lot more time than most gamers watching replays, so I tend to notice these things more often than many other people would.

It also highlights the fact that there is absolutely zero random chance involved in the game, which is something I can respect.

[skyride]

I wonder if LOL Replay (a client-side hack for League of Legends) works like the Supreme Commander example. Since it’s a multiplayer game, it would have to capture the input of all online players but I guess that information should be available on the client since the units have to be drawn on every client.

[Teiman]

A replay for a multiplayer game can be just the network data stream saved. So the game be set in “play mode” and use the data as is coming from a real server.

[Zylon]

@Otagan

It also highlights the fact that there is absolutely zero random chance involved in the game, which is something I can respect.

Or, again, it highlights that pseudorandom number generators are inherently deterministic. You just have to store the seed in the savegame file.

[WarrenM]

@Zylon

Or, again, it highlights that pseudorandom number generators are inherently deterministic. You just have to store the seed in the savegame file.

[Fugitive]

One of the downsides of that technique is that it becomes sensitive to engine changes. If you change something in the engine that would affect the gameplay, either old replays wind up going out of sync, or you have to try and keep the engine compatible with the behaviour in all previous versions, which could be hairy depending on how significant the differences are.

[Zylon]

@WarrenM

You’re an idiot, Warren.

[WarrenM]

U mad?

[TheTrunkDr]

@Otagan

While this may not be about save games in particular, I still think it’s somewhat relevant to the spirit of the topic.

One thing I always found interesting is how Supreme Commander saves its game replays. You can have a gigantic 8 player slog that takes hours to resolve and come out of it with a replay that is under 2 MB in size. Considering that you’re throwing around thousands of individual units in a game of that length, it seems very impressive.

What the game does is save the starting positions and a list of all player inputs over the course of the entire match. When you watch the replay the game just resimulates the entire thing from scratch by issuing the same orders as the players did at the exact same time. This demonstrates the fidelity of the engine, which goes so far as to track the trajectories of individual tank shells among other things, in that the same set of orders will produce the same results time after time. It doesn’t matter if my MA12 assault tank had 220 hit points at 11 minutes, 13.84 seconds. The game will figure that out based on the fact that it was shot three times by Mantis assault bots a minute prior so storing those variables is useless.

It also has the bizarre quirk of enabling you to tell how much micromanagement is in a game based on the file size. More orders leads to a larger replay.

I just want to point out that this isn’t all that new. Lots of games handle replays in this manner and have for ages, going all the way back to the NES and likely sooner.

[JoshV]

@Fugitive

One of the downsides of that technique is that it becomes sensitive to engine changes. If you change something in the engine that would affect the gameplay, either old replays wind up going out of sync, or you have to try and keep the engine compatible with the behaviour in all previous versions, which could be hairy depending on how significant the differences are.

Some do it by having different codepaths for different versions. Most just invalidate replays between patches.