Development Blog Update - Committal and top-down level design in NS2
Flayra
Game Director, Unknown Worlds EntertainmentSan Francisco Join Date: 2002-01-22 Member: 3Super Administrators, NS2 Developer, Subnautica Developer
Game Director, Unknown Worlds EntertainmentSan Francisco Join Date: 2002-01-22 Member: 3Super Administrators, NS2 Developer, Subnautica Developer
Comments
Ahh... Ohhh...
What does this mean for the NS2TR? It sounds like your not planning to have the map creation assets done until the very end.
Is the NS2TR being regarded as a separate project, from the planning and development point of view?
Commitment is good, downloads are better <img src="style_emoticons/<#EMO_DIR#>/biggrin-fix.gif" style="vertical-align:middle" emoid=":D" border="0" alt="biggrin-fix.gif" />
That´s a good way to prepare the foundations for any kind of project and, in this case, a great way to prepare everyones creativity towards the same end and focusing the all the effort on meaningful abstractions.
<!--quoteo--><div class='quotetop'>QUOTE</div><div class='quotemain'><!--quotec-->Are there any areas in your life you should have a little more faith about and just commit to?<!--QuoteEnd--></div><!--QuoteEEnd-->
Regarding your question, I think we cannot go Pro 24/7 and think we are always making our decisions based on certainty. Sometimes the best decisions and, mainly, ideas come from intuition. That´s a motto I try to keep in mind and not go mad under pressure.
+1 That was a really cool perspective on the design. The parallels seem obvious in retrospect.
Do not move into the area of props and hi-rez textures.
But some basic lighting and some color should be informative enough
for a player to discern things about the level.
fullbright with the same texture everywhere makes it incredibly hard
to ascertain whether you are on to something good or not.
Not even full color textures ...orange box in simplicity but more NS style (dark gray, dark green, dark poop)
Build tons of levels and playtest them until the best shake out.
Paint with bold strokes.
Then refine the ones that shake out.
rinse and repeat.
To be honest the more levels you throw at your game.
the more likely you are to see balance issues ahead of release.
even if you don't keep them.
good luck guys keep plugging.
Do not move into the area of props and hi-rez textures.
But some basic lighting and some color should be informative enough
for a player to discern things about the level.
fullbright with the same texture everywhere makes it incredibly hard
to ascertain whether you are on to something good or not.
Not even full color textures ...orange box in simplicity but more NS style (dark gray, dark green, dark poop)
Build tons of levels and playtest them until the best shake out.
Paint with bold strokes.
Then refine the ones that shake out.
rinse and repeat.
To be honest the more levels you throw at your game.
the more likely you are to see balance issues ahead of release.
even if you don't keep them.
good luck guys keep plugging.<!--QuoteEnd--></div><!--QuoteEEnd-->
Seems like that'd be more of a manpower issue unless they get it out to community to make these kind of "beta" maps
no seriously though, the only thing that will stop you from wasting a lot of time and effort is PLANNING! I've managed to spit out a map in about 3 weeks, thanks planning, planning planning - i probably spent about a month here and there just brain storming ideas, i worked out every room I wanted to create (i actually worked out 15 rooms, but ended up scrapping 4) I worked out the purpose of the rooms - which then brought on visual ideas (of course i had rough theme in mind already and I tend to keep a portfolio of cool looking designs and structures which i can draw apon when i need an idea).
unless you are an extrememly experienced mapper there will always be technical issues to overcome.
and of course - if you want quality you have to be willing to $$$ for it - generally its worth it in the long run.
Why don't you guys do what CS:S did ? they had 2 maps based on the originals apon release - then they added more,s o why not start with eclipse or something like that - it wouldnt take long to make, and use that as your baseline. then make adjustments to it so that it fits the style of game you want - and then when you are happy, use that as a new baseline or template to create all new maps from - list the reasons why it works and why it doesnt work etc - just like you guys did with ns1 mapping resource page that thing was great
Walter Benjamin
Because they don´t have, with NS1, the massive player base CS 1.6 had, therefore thay must/should release the game with a handful of maps from the beggining, simply because there won´t be hordes of players willing to wait. Besides, I think players tend to get tired of playing the same map all the time, even more considering we are talking about a RTS/FPS, and not a simple FPS.
I'm sure playing the old maps on the new mechanics would be interesting anyway. However, the gameplay changes from NS to NS2 are most likely HUGE compared to CS --> CS:S. The requisites of a functional NS map are going to be a lot different.
Now I have doubts, and will not stay around for NS2 anymore. I've seen this fail nine too many times.
I really hope that was a bad attempt at sarcasm...
The iterative article actually sounds a lot like what my team is doing right now, except we're not making a game we're making computer chips. Funny how that all works. The parallels are actually quite striking.
I have some factory pictures including machines, fires, buildings if it is needed. But all my inspirations are coming from my family (my wife & son).
For those questions there are my some advices if it is helpful;
Choice some pro players at least from 4 pro clans (24-50 players) as beta testers and design a web including votes for everything for that beta
(maps, textures, weapons, tactics, game play, engine etc.). And let them discuss & vote those.
And choice best ones (max votes), focus on them, improve and release.
1. "How can we make big design choices until we've proven it works?"
- Don't make any choices, make beta as flexible as it is. e.g. marines with com & aliens with com, without coms etc. Test it with
as many as it possible using low graphics may be without textures <img src="style_emoticons/<#EMO_DIR#>/smile-fix.gif" style="vertical-align:middle" emoid=":)" border="0" alt="smile-fix.gif" />.
Many new weapons etc. And vote all of them in that beta testers group. maps, weapons, tactics etc. and
some new ideas.
- Focus on the best choices
- Improve visuals, textures, lights, smokes, fires bla. bla.
- Some mathematical measurements in game design should be available for all map designers. e.g. total distance of RT's to base and hives.
total arrive time from base to a RT as a rine or alien including doors, steps, elevators etc. Room availability for lerk, fade, onos.
2. "How do we know we're hitting our visual quality goals without completely finishing an area?"
- First the best maps are the most balanced maps without visuals. So same maps with some changes may be easy to adopt first maps.
Reverse X,Y,Z of map and add new textures, lights, new visuals and some additions like some mechanics (pistons, engines, electricity motors, reactors etc.)
- Release a design of top view of map, and some visuals and vote in beta tester group.
3. "How many props will we need, and what percentage of those will need to be unique landmark props versus generic props?"
- %20 Best, flexible and the fastest engine including physics, light dynamics, etc.
- %20 Ultimate High Graphic maps for future graphics with options for lower graphic cards also (e.g. 8600 would be base for a lower gfx card.)
- %20 Game play including single play, highly focused on to multi-play
- %10 New much more weapons (may be some weapons for a map) including fire guns, mobile sieges, laser guns etc. (Like Fear 1&2, BF-2142, Quake etc. )
- %10 Much more marine & alien models, folk groups may be.
- %10 Menus and useability specs., weapons selection, upgrades, tech tree, command console screen
- %5 Online connection abilities, balanced user input & output whatever his bandwidth is.
- %5 Include a graphic benchmark for a map for those heavy graphics. Benchmarks are good things to show your game & graphics and also name of the game in a benchmark lists.
(camera walks in the map and more than 50 aliens jumps to many marines , fires, explosives etc.)
4. "How will we know when we're done creating assets?"
At least (multi-play only) :
- High graphics based fast engine which works in many gfx cards without bugs ( include less physics )
- include 8-10 maps with high graphics
- include 2-5 new weapons
- include 2-3 new marine models , and alien 2-3 different alien model packs (dark models, light models, ...)
The Best (including single play) :
- High graphics based fast engine which works in many gfx cards without bugs ( include High physics, mechanisms, objects )
- include 8-20 maps with high graphics
- include 2-10 new weapons
- include 2-5 new marine models , and alien 2-3 different alien model packs (dark models, light models, ...)
- Single play with 5-10 levels
- Commander play only mode with same levels with AI bots
- Multi-play with AI bots
- Map design tool
5. "How many varied maps will we get out of each art set?"
- include 8-10 maps with high graphics whatever art set is.
6. "How do we do all this without wasting effort (and while having people work remotely)?"
Core things (if u want to sell) should be included in the release of the game. Other packs (like new engine, maps, models etc. ) should be upgradable.
Yilmaz [ABLE.ns *] from Turkey
<a href="http://www.yyoru.com" target="_blank">http://www.yyoru.com</a>
We did this once before, it didn't work out so great IIRC...
Not to say pro players should be excluded, but they shouldn't be the only testers either.
One more thing: Don't make more than 4 or 5 levels to start with. That worked for Team Fortress 2 and it will really work for you guys. <img src="style_emoticons/<#EMO_DIR#>/smile-fix.gif" style="vertical-align:middle" emoid=":)" border="0" alt="smile-fix.gif" />
Are there any areas in your life you should have a little more faith about and just commit to?"
Making a decision you believe in and sticking with it the whole way through is extremely important in most things in life. I'm the type of person that makes a decision and sticks to it the whole way through, it really makes everything a lot easier. You become a more reliable person to yourself and others. You have to try something before you can fail it and try it again, and indecisiveness is really a waste of time.
Best of Luck
One more thing: Don't make more than 4 or 5 levels to start with. That worked for Team Fortress 2 and it will really work for you guys. <img src="style_emoticons/<#EMO_DIR#>/smile-fix.gif" style="vertical-align:middle" emoid=":)" border="0" alt="smile-fix.gif" /><!--QuoteEnd--></div><!--QuoteEEnd-->
A lot of the maps that TF2 is releasing nowadays are really bad community maps that have huge flaws with them. Sure they get updated but very slowly. What NS2 should do is take community maps but actually test them and see if the gameplay is fun and enjoyable for the regular pubber but also very viable in competitive scene.
Also the few maps that they released at the start are the same maps that everyone is still playing. A lot of servers are on a 4-5 map rotation and that just gets stale quickly.
The thing about mixing pro and amateur players in a beta test, is that they're not all at a similar skill level in terms of video gaming. The 'pro' players will annihilate the noobs if they start out on a level playing field, just b/c they have more precise FPS skills.
For getting general gameplay/map feedback, I'd suggest pub games of whoever shows. For getting nuanced feedback, toss in people who know how to take advantage of anything that might even resemble being unbalanced and exploiting it to the max in order to win the game. That, and having people being able to aim and have situational awareness is important for balance purposes.
Look @ Blizzard. They're getting the best SC players and having them critique SC2. :x. They're not recruiting -bad- SC players, instead, they're recruiting the top. The top understand the game better than anyone, and understand the nuances that even the game developers don't have the time to grasp. NS isn't SC. It doesn't require the com to key 200+commands/min. to become really good. But it does have a steep learning curve (not as much as SC it would seem), but still very steep. You want the pro's testing the game when it really counts.
Marriage.
then the next 10 times you say " wow look how cool the physics are", and proceed to throw grenades at yourself and watch your body rag doll.
But what a good game comes down to is gameplay and map design. look at all the lasting games, notably CS. the graphics a crap now, however the gameplay keeps people coming back.
i wouldnt stress that much about the graphics imo.
but im there for the teamwork.
S!eep
I agree with starting with 4-5 dev made maps originally. Get these balanced and as flawless as possible before getting the game out the door, rather then 15 maps that are unbalanced. So basically saying we want quality, not quantity!
oorrhhh grandmaster Flayra asks deep philosophical question.
I should probably just commit a water management career path rather than just flirting wiht the idea.
I dont really know much about map design, ive mucked about in hammer editor a few times. I agree with the above state quality not quantity. If ns shipped with 5 decent maps i would be satisfied (but id expect more to appear from both devs and community over time).
How ever i do belive that the initial map pack will have to a certain extent sell the game. Atleast two of them are going to have to wow the players.
Of course you need to test it with people who don't share your gameplay proclivities or at least keep them in mind.
Once you do that everything else - graphics, sound, flow, balance - they can all be changed to fit.
But the question which keeps resonating in my head is: Does their Zen apply here?
First I just want to commend you guys for just sticking with this and keep pushing forward with it.
From what I've seen, you guys are doing an <b>amazing</b> job. The engine video I saw showed promising results!
Getting back on the topic, I think I know how you guys feel. Currently I'm making a marine in 3D with a concept artist and things are becoming frustrating. We get a base concept, I start modeling away. Somethings look good but somethings just look bad. Looks good on the concept but once it's built in 3D, doesn't look as "cool" in some areas. Go back and change all the meshes, and come up with something better. It's a constant repeat process for us right now and it's very frustrating.
But the question is, "How much time do you have to do this whole process?" This process of, okay that doesn't look 100% like what we wanted, let's go back and change. After the changes, I'm sure there will be OTHER changes that you would want to make.
So really, how much time do you have?
No one is perfect. Plus when money and time has a deadline, you really just have to do what you can to fix it up as best as you can with the time given and move on. Be proud of it afterwards since you did your best and call it a lesson learned.
You can take these lessons and make sure they don't happen on the next project or level.
Anyways I think I've blabbered enough. The project is showing strong results guys and keep pushing!
If you guys ever need help w/ 3D modeling w/ environments or whatever, give me a hollar. I really like where this project is going and would love to help you guys out (www.digiegg.com). I wouldn't mind taking that test you guys sent long time ago again either so just let me know.
All the best and I'll keep my ears open!
-Wooyang
<b>- Local balance
- Global balance
- Navigation</b> (including Recognition)
<b><!--sizeo:3--><span style="font-size:12pt;line-height:100%"><!--/sizeo-->Concentric Multiplayer Map Development<!--sizec--></span><!--/sizec--></b>
As I've mentioned in the NS Mapping Forums, my ideal way of mapping would be to create main areas of focus first and then interlink them. This is a three-stage process that begins first with concepting your key areas in general -definitely not concrete- terms, then building your key areas and testing them initially for balance, and then linking them together by corridors or more complex structures and testing the ensemble for overall balance. Like dropping a pebble into a pool of water, you will work from the centre outwards.
First, to be clear about which are the key areas, in NS these were the Hives, the MS and the res nodes, as well as any obvious relocation spots (but that would come later on in the process as the layout is decided on). The MS generally has 2-3 exits, the Hive generally has 3-4 and res nodes generally have 2-3.
<b><!--sizeo:3--><span style="font-size:12pt;line-height:100%"><!--/sizeo-->Concepting the map<!--sizec--></span><!--/sizec--></b>
Here you would probably go from a global concept and fill it out with more ideas for locales. For example if you pick a Refinery you already have a good idea of art style including colour scheme, soundscapes and prop types. But it's really important at this stage to come up with unique areas that will be instantly recognisable to the user, because 'recognition' aids massively in map <i>navigation</i>. When I say 'unique' I don't mean 'all-new', I mean that they are unique from eachother. E.g. <i>Ventilation</i> is not sufficiently different to <i>Atmospheric Processing</i>, they need to be different in look and their look needs to link back in with their function and their function links back in with their name.
It's acceptable in fact it is better to go with clichéd settings if the user is going to be able to recognise them from their artwork immediately. The sooner a new user links a location to its name, the sooner they link a route to that name and the sooner they learn the map and become able to navigate for themself without the need for a Commander. By consequence, the sooner they can know the map well enough to be able to sit in the Command chair and understand the needs of their force from zone to zone. I'll come back to this in the <i>Navigation</i> section because it's very, very valid. To sum up, you begin working on the quality of your map navigation from when you begin concepting it.
You can work out a provisional layout to give you an idea of how you'd like things to fit together, but don't look at your key areas and linking areas as jigsaw pieces, look at them as hexagons and squares. They don't have to fit together exactly as the first layout suggests, there is enough flexibilty in your process to change the layout and develop it according to overall balance needs.
<!--sizeo:3--><span style="font-size:12pt;line-height:100%"><!--/sizeo--><b>Local Balance</b><!--sizec--></span><!--/sizec-->
When you have your areas concepted, you can begin to think of how your 'landmark' art pieces will fit into an interesting room layout. As soon as you have a simple layout, you can test it alone as a separate module. When this is done for multiple modules you intend to be in the same area, you can then link them up via corridors, ducts, maintenance halls, vents, and so on. This gives you more control over the travel times between areas, since changing a corridor length is much easier and quicker to do than changing the layout of a feature room because a corridor does not require as much or as complex unique artwork. This will also help you arrive at your global balance very quickly.
<b>Unit/Module-Testing Multiplayer Locales</b>
But, before this, something I also mentioned was a method of testing the key areas of the map. If we look at NS we know that for every Hive there are only a few entrances either side can attack from. When a battle is pitched there, there is no letting up, it generally goes on until one side is sufficiently defeated whether by elimination or forced retreat. The battle does not switch location, teams don't generally get pushed back to the previous key area, they generally fight in the same spot using the same approaches until they win or lose.
So as soon as you have an idea for the layout, using placeholder art for the landmark aspects, you can test the functionality. Some of these tests can just be run with a few players and will give you very quick indications of whether the Hive/MS or Res Node is too susceptible to certain plays such as grenade spam or siege lockdowns.
<u>Functional Test Cases:</u>
<b>Locational and Situational Combat</b>
There are quite a few scenarios you can put together for this, but essentially the idea is to build boxrooms outside the key area's exits and put team A's spawns there and team B's spawns inside the key location. For certain scenarios, you may want to use Phase Gates (or DI?) instead of spawns to get a closer simulation. Some scenarios to test would be:
<i>Marines attacking</i>
- Hive 1 Shotgun rush
- Hive 1 campathon (just a couple of LMG Marines camping the exits and seeing how easy it is to keep the Aliens in their Hive)
- Hive 2 Jetpack rush
- Hive 1.5 Siege (Sieging the active or growing Hive with one Hive growing)
- Hive 2 Heavy train (with or without Siege)
- GL spam
- Ninja PG (the Marines are all standing in the Hive shooting it down and the Aliens need to use the Hive transport to arrive - this tests the balance of the respawn locations)
<i>Aliens attacking</i>
- Onos (make sure the Onos can navigate properly)
- Hive 1 'rush the phasegate'
- Hive 2 All out attack (make sure Fades, Lerks, Skulks can navigate okay and it isn't very easy to get caught on architectural fluff)
- AA/Proto assault (ensures that 1. the IPs and PG can be placed effectively and 2. the beacon respawn locations are balanced)
<i>Other</i>
- Relocation (can the Marines gain an unfair advantage -regardless of location- by building all their main structures here instead of at MS?)
You can take the results of these tests and continue development to address the outcomes. A Hive that is particularly susceptible to JPs for example doesn't necessarily mean the map is broken, it just means you have to be careful not to give the JP too much advantage in the rest of the map.
<u>Recognition</u>
A bit later on, but still relatively early in the process, you can test if your key location is recognisable and matches its name. Get a set of screenshots or concept art together (preferably early screenshots) and also get a list of the corresponding area names (the ones users will be using over the mic or in chat). Now ask someone independant to pair the two sets together. If you have any mismatches, you need to find out why they mismatched and fix the issue.
<b><!--sizeo:3--><span style="font-size:12pt;line-height:100%"><!--/sizeo-->General Balance<!--sizec--></span><!--/sizec--></b>
Once you add in the interconnecting corridors and get everything fitted together, you can do some more normal multiplayer testing to get a feel for the overall balance. By this point, though, you shouldn't have to make any drastic changes, especially not in terms of scrapping or producing big art pieces.
<b><!--sizeo:3--><span style="font-size:12pt;line-height:100%"><!--/sizeo-->Navigation<!--sizec--></span><!--/sizec--></b>
Navigation is massively important in differentiating new or less frequent players from more hardcore players. The stages of learning a map from my perspective are:
1. Main routes (these will be learnt in sections, for instance in your first game you may only learn the main route to 2 Hives because you aren't being asked to retrace steps to the final Hive because the endgame will probably be short.
2. Danger spots (anywhere you can get surprised or where you have a natural disadvantage)
3. Subsidiary routes (lesser-trodden routes to get around the map more quickly)
4. Deeper knowledge (situational intimacy with the geometry of a given location, allowing you to do things like jump backwards onto a rail or bounce grenades off corners and know where they'll land)
If you have a good, solid design that aids map navigation, phases 1 and 3 will be much shorter, allowing the focus to shift to phases 2 and 4 where better players wield their knowledge over you for a bigger advantage. Here's a detailed explanation:
<b>Map Navigation: Lucid versus Origin</b>
Compare for example the NS maps: <i>ns_lucid</i> and <i>ns_origin</i>. Origin has very strong thematic art schemes for its key areas.
- Furnace (Hive): Strong warm colours, pits of brightly roaring molten lead.
- Cargo Bay (Double): Cool blue-green colours, a big crate held from a loading arm, cargo lifts and cargo bay doors.
- Ventilation (Hive): Same colour scheme as Cargo, but they are linked intrinsically in the middle of the map. A large duct and fan whirr repetitively overhead.
- Biodome (Hive): A fairly neutral colour scheme inside, with blue sky visible through skylights and strange plant types growing from the floor
Even the res node locations have the same strong sense of theme: Computer Lab, Laser Drilling, Power Core. Ore Extraction has a strong theme (flatbed device for transporting ore) but it doesn't tie in very well with its name. However, like all other locations in the game, the names tie in with their zones (see below).
Furthermore, in Origin the colour scheme for every one of those 4 locations extends to all the surrounding corridors. The map is divided into zones by colour, much like the map of the <i><a href="http://en.wikipedia.org/wiki/The_Crystal_Maze" target="_blank">Crystal Maze</a></i>. These zones even have names: Control Room, Mining Operations, Industrial Sector, Science Annex, along with signposts in the corresponding colour positioned at strategic locations along the walls. This means every time you step out of a map zone, the colour scheme changes and you can tell subconsciously which zone you are heading to. The colour scheme will even change <i>within the same vent</i>.
In stark contrast, maps like Lucid do not have strong locational themes. Lucid takes longer to learn and is harder to navigate because it has more abstract architecture and art styles, and more obscure location names. The Hive and MS names in Lucid are: Marine Start (can't argue with that), Nemesis, Fear and Void. There is nothing about the art style of those locations that links them immediately to the name. Other place names are similarly forgettable: Rumour, Glare, Conflict, Fate, Stumble, Slumber, with 'Lost' summing it up quite nicely.
Regarding the lighting, if you look at Fear (Hive) there are patches of white, orange and green in the same room. Throughout the map most corridors are white, but they and the key locations have small blotches of colour indiscriminantly positioned in random directions according to the type of light source. This would be fine if particular types of light sources were used in different sections of the map, but this is not the case with Lucid, so lighting does not aid navigation either.
The short version is that Origin uses thematic design at all levels to aid map navigation while Lucid eschews all of this. They are both good solid maps once you've learnt them, but to begin with, with Lucid you'll probably have your map up most of the time instead of looking where you're going.
<b><!--sizeo:3--><span style="font-size:12pt;line-height:100%"><!--/sizeo-->Conclusion<!--sizec--></span><!--/sizec--></b>
<b>Positives</b>
- Once a key area passes gameplay testing it's basically finished and can be left to the artists to spruce up until towards the final builds. You can be confident it can be made to work well within the overall map layout, which is negotiable because it fits around these key locations.
- Each completed section essentially doubles as a testbed that can be used for testing your weapons, classes and abilities
- Next to no artwork wasted
- Should produce solid maps on release, so having only 5 maps on release shouldn't be an issue
<b>Negatives</b>
- Lots of test hours required
- You won't get to have a complete picture of the level until late in the development process
- Depending on how inventive you can be with the 'corridors', levels may feel a bit disjointed and formulaic
I think you should be testing maps as early as possible as each location is being made to give valid feedback. Because you are creating whole new gameplay, I would say that you should be looking at 2 months internal testing with at least 12 dedicated testers and a further 1-2 months after that for open or closed beta testing prior to launch. It would be best to involve the internal testers with the beta testing as well to see how a large experience gap affects the game and if you feel it's going to be a priority to address that post-release.