Campaign Maps and TF2 - a "working plan"

[Zed]

It can prompt you for any map you can buy stamps for I presume.

 

[EArkham]

I've only ever gotten prompts to buy stamps for ctf_turbine, despite putting tons of time last year on pl_borneo. Ages ago, well before the surplus of new maps, I played turbine quite a bit because it was popular with the community servers I frequented.

The stamp suggestion system might need some tweaking... seems like it's going by "total time played," and not "total time played past six months."

Have all the Valve maps, new and old, been compressed with the new algorithm thinger? If not, this really should be done no matter what's decided re: removing maps.

 

[Egan]

I'd be ok with the idea of low performing maps being taken off servers if they were still considered to be part of the official lineup and still could come back in featured competitive playlists occasionally like Idolon has mentioned:

where maps relatively quickly cycle in and out of a featured "playlist"

I'm sure a lot of community map authors would like to see their maps played competitively at least for a small time. And tests (of my 170+ gamedays) have shown that when we enforce teamwork (turning off all-talk) even regularly-boring CTF maps become fun.

I'd prefer though that this wasn't an issue all together.

--

If this is an issue about compressing the game's file size, then (as mentioned) they should 'bsp compress' all of the past maps - this would save at least 1.5 GB of storage space (my laptop has no custom maps, and I excluded already-compressed workshop maps).

 

[Pocket]

I've already said how I feel about culling maps, but to sum it up:

• Some of the old "community" maps ought to go (Fastlane, Junction, Watchtower), not because they're unpopular but because they suck
• Campaign maps should be rotated out like in CS:GO but if Valve happens to really like some, they should buy them out and make them permanently official just like in the old days
• Also get rid of ctf_sawmill because no, just no

As far as Half-Life 2 textures: Replace some of them (e.g. that ugly railing) with proper TF2 style ones with the same filename, move the acceptable looking ones (e.g. stairfronts, tool and dev textures) into the TF2 textures VPK, purge the rest. One upside to the death of dedicated servers is that any maps that would be broken by this don't get played anymore anyway. Except trade maps, but... well, I've already said how I feel about those too.

As for other space saving things: would it be possible for all the vpk files to get some form of compression, like a .bz2 or zip file (or like repacking the bsp for that matter)?

Pretty sure VPK is already a compressed archive format, plus there's probably not a lot you could do to textures that are already stored as compressed image formats.

 

[ics]

Perhaps Valve could, if doing this - or any system where certain maps are kind of, promoted but not added - extend the stamp system to them, and make it clear that stamp purchases on a map are an excellent way to recommend it be added to the game. The mappers get small amounts of money from stamps as well as the publicity, and maps that are chosen then get the usual amount. Being in an operation then gets you downloaded with the game, gets you the money from the operation, and gets you included in whatever fancy contracts system is in use for that operation.

Stamp purchases are really not the best way to say which maps are popular or should be added and what are not. Really small minority actually buys them and so the end result may not be what community wanted and there is also possibility for exploiting. Someone seeing his own map there could even buy 1000 stamps because he would get the money back anyways, just with tax but if the map would get into more wider spread like in mission rotation and earn pass revenue, it would be worth it. Even paying for friends to buy passes and return the money later would be possible.

Current way maps are chosen is interesting and actually quite good, since we have no idea what maps they will pick.

 

[CommodoreKong]

More to the point, as I interpreted it, 'low performing' maps would be in terms of playtime, popularity, balance. So, if we use Junction, a map which is unpopular and arguably unbalanced on larger servers, I would think that it would get the 'low performance' marker, because not too many people are excited to play the map. So I would think that 'low performing' would basically be a measure of how fun it is for players. I miight be over thinking it now, but you'd think that naturally, you would want to be diverting your players to the statistically 'most fun' maps so that the perceived quality of TF2 goes up, rather than being more hit-or-miss. You want to increase the chances of new players hitting the most fun maps, which means they'll have the best enjoyment from the game, and a better chance at a longer retention.

Based on what you posted though it sounds like Junction wouldn't actually be up for removal, just community maps added since the campaign system started. Is that correct?

 

[Vel0city]

Pretty sure VPK is already a compressed archive format, plus there's probably not a lot you could do to textures that are already stored as compressed image formats.

Nope. I did a quick calculation of all the VPK files with hl2_textures (in Team Fortress 2/hl2) which made it around 1.27GB. I then opened the dir file which showed, you guessed it, 1.27GB.

 

[Werewolf]

culling out the low performing maps?

This is just for clarification sake, but what exactly are we talking about here? Maps that don't get played on enough such as tc_hydro? Or maps that are popular (measured in terms of how many people connect to a server running that map) but are poorly optimized and thus cause players to leave or even crash out.

Surely popularity could be measured in terms of connects, disconnects and average playtime. I think you'd have to ignore connections during setup time and after the match has been won. My idea about this is that if players see a map they like is being played on a server, they join the server to play it and stay on that server till the map time runs out. Counter to this is when the nextmap is one players don't like, so they all disconnect.

 

[Fruity Snacks]

This is just for clarification sake, but what exactly are we talking about here? Maps that don't get played on enough such as tc_hydro? Or maps that are popular (measured in terms of how many people connect to a server running that map) but are poorly optimized and thus cause players to leave or even crash out.

Surely popularity could be measured in terms of connects, disconnects and average playtime. I think you'd have to ignore connections during setup time and after the match has been won. My idea about this is that if players see a map they like is being played on a server, they join the server to play it and stay on that server till the map time runs out. Counter to this is when the nextmap is one players don't like, so they all disconnect.

More to the point, as I interpreted it, 'low performing' maps would be in terms of playtime, popularity, balance. So, if we use Junction, a map which is unpopular and arguably unbalanced on larger servers, I would think that it would get the 'low performance' marker, because not too many people are excited to play the map. So I would think that 'low performing' would basically be a measure of how fun it is for players. I miight be over thinking it now, but you'd think that naturally, you would want to be diverting your players to the statistically 'most fun' maps so that the perceived quality of TF2 goes up, rather than being more hit-or-miss. You want to increase the chances of new players hitting the most fun maps, which means they'll have the best enjoyment from the game, and a better chance at a longer retention.

I explain that here.

To be honest though, I don't know and I didn't ask more about it.

 

[Blade x64]

To be honest, I wondered why they didn't do something like this sooner. We have a lot of maps that most players never see and for good reason. Removing excess bloat while still keeping the maps accessible is an ideal solution. The only difference is that you have to download these maps should you ever meet the edge case of a server running one.

 

[Vel0city]

To be honest, I wondered why they didn't do something like this sooner. We have a lot of maps that most players never see and for good reason. Removing excess bloat while still keeping the maps accessible is an ideal solution. The only difference is that you have to download these maps should you ever meet the edge case of a server running one.

That's the problem. Valve servers run these maps. So unless Valve is willing to remove maps like Junction or Fastlane from the map rotation you'll always need to download maps over and over again.

 

[Blade x64]

That's the problem. Valve servers run these maps. So unless Valve is willing to remove maps like Junction or Fastlane from the map rotation you'll always need to download maps over and over again.

I fail to see how it's a problem. If a map is unpopular, the chance that a player will join a server with it running is low. And once they've downloaded it, they have it as long as TF2 remains installed. So no, players will not be downloading maps over and over again.

This is no different from playing a new custom map on your favorite server.

 

[Fruity Snacks]

Did everyone miss the part about if maps do get culled, they get tied to a workshop link, so if a server owner puts them in rotation, you'll just download the workshop map and voila, it's there? No need to reload, because it's always there after you DL it once, and it'll always refer to that item?

 

[EArkham]

It's kinda funny how the general consensus went from "if they remove maps, people will freaking riot!!" around the time of Gun Mettle to "okay, yeah, that IS quite a lot of maps, isn't it, maybe we could stand to lose a couple" now after Tough Break.

I wonder what would happen if all community maps were removed and then rotated back in periodically as big batches. Kinda like CSGO, only with larger map sets, but not necessarily tied to a campaigns or updates.

Also, if a map is pulled and goes back to the workshop, does that mean the author moderates that workshop entry as far as updates and so on, or does it become a Valve workshop entry that they moderate?

 

[Fruity Snacks]

No idea.

 

[Blade x64]

As I see it, as long as it is a map run on official servers, Valve will be maintaining their own copy. The original mapper is welcome to make another entry for updates that will be reviewed to eventually make it into the game.

 

[UKCS-Alias]

I actualy think that in the current state it would be too early to drop out maps unless they are heavily underperforming. But in the current state i only see that happen with arena mode (its not in quickplay).
To me even with 8 maps in a certain gamemode rotation its just not enough. 3 maps obviously get played alot. But only having 5 extra maps isnt realy what i would consider as enough options to choose from. Hell, to me the point where culling would be needed is when a popular mode exceeds 20 maps, and for non popular modes it exceeds 5 maps (so far these modes dont even have 5 maps or are out of quickplay). Culling is just a lazy way to get rid of less popular maps without even attempting to save it for those that actualy do like it (junction still gets played, dont forget that)

Even if a map was originaly added and realy worthy to get culled, i still think its simply bad to do so.

Instead however, they should make it more reliant on the workshop where a map only gets downloaded when you want to play it, and culled when you havent played it for months. But still remain permanently in the rotation. Only the realy popular maps are downloaded in the update itself, and when in a vote a certain map would be selected it just should pre-download the map.

"But what if the workshop is down and the next map is one i dont have?":
It wouldnt be diffirent to how pub servers have to handle it then. But for that they could still have the server you play on supply a seperate download option just in case. That it requires more effort... well, im sure it would be a good system for games like csgo aswel.

 

[Vel0city]

I wonder what would happen if all community maps were removed and then rotated back in periodically as big batches. Kinda like CSGO, only with larger map sets, but not necessarily tied to a campaigns or updates.

It would be better if Valve did that first with their own maps, or the ones they bought like years and years ago. While maps like Fastlane and Junction and Frontier may have worked back in the 2008-2009 era, they certainly don't do well now because while the game itself has evolved with all the unlocks and weapon balances, the maps have stayed the same.

Again, we're talkin here like maps take up most of the space. In terms of single files, yes. But there aren't that many maps shipping with the stock game (97 to be exact) compared to the thousands of textures and models and sound files and other technical related things like all the dlls and whatnots necessary to run this game properly. Maybe we should look more into getting those things down in size and then come back to the topic of temporarily removing maps/making them Workshop downloads to save space. Heck, repack all maps first (because apparently that isn't done either), see how much space that saves.

 

[Pocket]

On a related note, I'd like to see some changes to how Quickplay works. Right now, the Valve servers come in two flavors: 24/7 [gamemode] and 24/7 [campaign]. There's no official servers that just run the whole rotation anymore, so no wonder arena maps and Hydro never get played.

A popular thing to do nowadays is to just have a bunch of dynamic servers whose IPs are obfuscated by the equivalent of quickplay or matchmaking, and repurpose them on demand based on what players are interested in playing. But that's mostly games that have no support for dedicated servers at all, and the ones that do support them I assume have a strict split between those and official servers — either you go into quickplay and get dumped into an official server, or you pull up the browser to find a third-party one. Anyone here play other online games? I'd like to see some of you weigh in here with your experience.

 

[MegapiemanPHD]

I can see this making sense for specific types of maps. As many have stated, Arena is unsupported so putting it somewhere that's accessible but not part of the initial download makes sense. What else makes sense, to me, would be holiday specific maps. Maps like Mann_manor, Helltower, and so on. These maps are only really played once a year and generally aren't needed any other time. Plus they already have non halloween versions in the base game outside of Moonshine and Millstone.

Some of the odd arena to koth conversions I would agree with as well but not all of them, like nucleus and sawmill. Although CTF sawmill is just awful.