Thursday, May 10, 2012

The failures of Rust - Developers notes: part 2


Through the numerous play testing that Rust has gone through since the contest began, certain parts of the map always seemed off. I was able to get these parts to a reasonably well playing state during alpha, but these were “small” tests. Even with the reassurance that it was “a memorable layout”, I wasn't too happy with how Rust was playing. I decided that I needed to take what I had, and move on to detailing, since time was becoming short, hoping that I would have enough time to test and fix it, but due to an overwhelming semester at school, I was cramming at the last minute. I had to enter what I had, and hope for the best. The testing for the contest became more focused (at least, so far) and the short-comings of the map became much more clear and evident. I feel like it takes a semi-decent person to admit that something they did is bad, so here is post about what I think is wrong with Rust, and what I think I need to do to fix it.

Issue 1 - Control Point 2 (CP2):

This is probably the weakest point/yard of the whole map (third most complained about part of the map too). It was originally designed to be indoors, but detailing such an area proved to be difficult, so I set it to be an outdoor yard. Gameplay wise, this did not change too much except that Demomen and Soldiers could have more mobility. For this area, lack of sentry spots and “hard to hold” (these two go hand in hand technically) is the most numerous negative feedback I receive for this area. But one of the largest issues that I see for this area is not necessarily something in the area, it's something that isn't. A majority of the major battles for CP2 take place just outside the area, at the main gate, rather than inside, on the yard. First, lets take a look at the area around CP2:

Fig 1: Red dots are sentry spots, yellow lines are common sniper sightlines.

Looking at the above image, we can see that there are more sentry spots at the entrance of CP2, rather than covering the yard of CP2. Through watching demo's and simply playing the map during testing, I can attest to this. The most common sentry spots are four spots around the shack. But why is this? Why is Red building so far forward?

Analyzing some replays of the map, I noticed that Red has a slight respawn time advantage (this was confirmed when I looked at the source file). I don't know the exact value, but it is probably close to 1.5-2 seconds. This is BIG. This means that Red can position themselves farther up inbetween Blu's respawn waves; thus we see the more forward line (and, when the line is broken, we see Red building past CP2, into the next yard. But this is something different, and acceptable).

Solution:
Before writing up this post, sitting down and REALLY looking at the problem, breaking down what was going on and why certain things were happening, I was afraid I was going to have to rebuild the whole area. But I realize now I won't need to. The first part will be to increase reds respawn time by about 1 second (to start). This will give Blu a better chance to push INTO the yard. The second part of the solution is to make the current sentry spots more attractive and a bit stronger. This will be a little bit more tougher, but some tweaking to the paths, and some selective cover to narrow down the side routes should solve the problem.


Issue 2: The Final Hill

The final area of Rust received the second most negative feedback from testers. In my previous blog post I explained my design theory behind it, using “tiers” as a way to describe the paths and increasing difficulty Blu would have pushing up the hill. While it worked out in theory, it didn't work out too well in practice. Playtests showed that while Red could defend, the final area, there was no real coordination behind it, no dedicated sentry nests and that any long-term defense could be easily countered by Blu. This has lead Blu to having a relatively high winning percentage. Lets take a look at the final area again, where I outline (in their respective colors) the routes Red and Blu can take.

fig 2: The routes between the tiers; note the accessibility of blu into red's final tier. 

You can see motions between tiers, and the decreasing amount of routes each team has. But this isn't good enough for Red's defense. There is just too much motion for Blu around final. This problem is analogous to trying to hold back a large wave with a wall of sand, you need a LOT of sand to be effective.

Another issue regarding final that is commented on a fair bit, is the sightline that blue has through a window at the lower part of the final yard (in tier 1). This sightline was added as a way to counter any obvious defense that was set up on the cart, but this window has proven to be too powerful.

Solution:
In short, Red needs a stronger defensive advantage, and Blu needs to be limited further in the possible routes they can take. The first thing I will do is remove the stairs in the upper part of the previous image and add a one-way gate for the lower route. This will reduce the routes Blu has to one, main route. But, I also want to create a stronger position for Blu in the lower building at the start of the yard. For this I will move the capture point back farther, closer to the building. This will mean that Blu will capture the zone sooner and receive the forward-spawn quicker. In turn, they will (should) be able to set up a stronger presence in the structure, and hopefully be harder to move out. I will not adjust Reds respawn time right off, I'll be interested to see how rust plays without doing so. If it looks like the route changes are working, but Red just needs a little bit more of an oomph, I will lower it by no more than 1 second.

fig 3: What I will modify routes to. The brown line denote cover. The lower line will be a one-way route down for red.

Issue 3 – The MOST complained about issue.

Okay guys, I get it. Clipping sucks. I ran out of time to do a better job. Avoid getting stuck on things.

Conclusion:
Whilst these changes might not be the best fixes, they are the strongest changes I could make that isn't a complete area overhaul (which no one likes to do). I will be holding off making these changes until the judging for the contest has concluded (out of courtesy to the other applicants. Updating maps during judging can lead to confusion as to what version is the contest version, which can skew grades). Though, you can expect a new version out very soon after the contest winners are announced.

You can see Rusts Development Thread (along with a map download and more screenshots) here
And you can see all the feedback given for Rust here

Friday, May 4, 2012

Design comments on the dynamic elements of Rust - Developers notes: Part 1.


This is the first of a few posts I'll be making about pl_rust and the design thought process I had, whilst making the map. Enjoy.

Rust was a map originally designed for TF2Maps.net Major Mapping competition #7: Payload - Dynamic Element. The contest was a twist on an older Payload only contest, but this time, mappers had to include dynamic elements as part of their design. For Rust, I decided to play it safe, and not be overly ambitious with what elements I included, as to help promote synergy of the dynamic elements into gameplay. The final dynamic elements ended up being a "timed element" in the form of a lift bridge and roll back hills.

The first dynamic element encountered is a "short" roll back hill which concludes at the first capture point. Originally this position was a vertical elevator that players would need to ride on to push the cart to the next level. The area around this though was a bit of a choke point, and I felt that a vertical element would cause too much of spam into the area and it would be too hard for attackers to push. I changed the area before the point to a simple roll back, so that there was still forward movement, moving any fire away from a single point on the track. (I also was having trouble properly getting the elevator to work on the technical side, but the given reason is a major factor in its removal).

fig 1. the "short" roll-back hill

The second dynamic element is the timed lift. Timed elements in payload are relatively new, the first being implemented in Thundermountain (Three Stage, Attack/Defend Payload) and then again used in Barnblitz (Single stage, Attack/Defend Payload). I wanted to give red time to set up a defense around control point 2, but did not want to extend the distance between CP 1's area, and CP2's area. The timed element fit into this area well, because not only did it allow Defenders to set up fortifications, but it also adjusted the routes in the area, so vary gameplay.

fig 2. the timed element, allowing red to set up defenses.

The final dynamic element is a "long" roll back hill. The third capture point is at the base of the hill and the final point at the top. I wanted to create a feeling of completion and reward for players who were attacking if they could successfully push the cart up both legs of the hill. This hill does consists of a flat "break" halfway up. This break made it so that a single push wasn't needed for the whole hill and that attackers didn't need to be extremely coordinated to successfully win.

fig 3. the final hill, marked "the hill of doom"

Hills in themselves are natural advantages to those who sit at the top; in this case, the defenders. Balancing the attackers power versus defenders power was tricky, but not impossible (something similar to this was done in plr_pipeline stage 3). To visualize my thought process for the final, think of each of levels as a different tier. Tier 1 is the base, Tier 2 the break, Tier 3 the finale. Red has automatic control of Tier 3; this is where they spawn, and this is where they have the best visibility on the battle below them. Tier 1 should automatically be Attackers. 

fig 4. overview of the final hill and visualization of tiers of the finale area.

Since they control the bottom point, they control that area (which is a building). Defenders have 3 routes from their safe structure, directly into Tier 1, Attackers also have 3. For attacking Tier 2, Defenders have 2 safe routes (and the main hill), and red still has 3. For attacking Tier 3, Defenders have 1 safe route and, again, the main hill. Attackers have 2 routes (includes the main hill) and a 1-way safe route. As you can see, progressing up the hill, Attackers get less and less room for error, while defenders have the same amount of routes for pushing. But, as attackers push and capture the tiers, defensive positions for sentries and sticky traps lower, so that by the final tier, all defenses are focused on one strong position. It wasn't easy to balance, but when I broke the area down into "tiers" and looked at combat on a tier-by-tier basis, the problem became easily solvable.

Rust has been entered in TF2Maps.net Major Mapping Contest #7, and is waiting judging results. For full documentation on development or to download the current version, you can view the Rust development thread at TF2Maps.net here.

Monday, April 9, 2012

It's the small things

Okay, well. The new detailing on rust is coming along, so that is good. But I was also thinking about the little things I could do this summer. Stuff that won't take me forever to do, but still are useful. Anyways, here's the list.

  1. Finish Portfolio - release to the interwebz
  2. Play more Call of Duty 4 and then begin learning Radiant.
  3. Write a tutorial/article on why Smaller maps doesn't mean better - add to portfolio.
  4. Create an updated Youtube series on how to use Hammer and make TF2 Maps.
  5. Turn this blog into less of a news, and more into resource about level design.
For 1, I need to finish Rust first, then do 3.
For 2, I'm working with another guy and we're going to learn it together in our spare time.
For 3, I need to just find the time and some examples.
For 4, That should be too hard, but I need to just outline what I want each tutorial to cover.
For 5, I need to stop doing what I'm doing now. I will be doing some cool stuff AFTER finals and AFTER the submission date for rust. I'm going to make a developer video, kind of providing commentary on what certain area's are set up a certain way. This, of course, will be attached to my portfolio.

Seriously, the next blog will actually have some design theory in it, at least the total 1 person reading this will mind.


Sunday, April 1, 2012

Rust and summer

Been a little while, I got busy with school work.

Anywho, a couple of big things!

1) I decided I didn't like the 90% of rust that I had detailed. So I started over. I don't want to release something into a contest that is going to have a lot of the public playing if I don't like the map. Visually, it is now akin to Barn Blitz style of detailing. I hope to have that done by the deadline. Don't know if I'll have much time to do anything else besides just detailing and optimization; that all depends on how quickly I work and school work I have.

2) Pseudomagnetism is a go! I'll be doing level design for my friend, ezekel, mod. Together we're coming up with a few more mechanics, plot, artstyle (going to be an old USSR Cold War research facility) . That'll be my summer project after rust is officially out. If it takes longer than the summer, so be it. I want to to make sure we do the best we can. I will be using this and my Youtube channel as a way to bring updates as I can. More info to come in the coming months. (As more idea's get cemented, I'll blog about them and why we're choosing to do what we are doing)

So thats it. Nothing too exciting.

Ciao.

Monday, March 19, 2012

42, is the magic number

There is about 42 days left to finish detailing Rust. That means it's time to start a new project! I mentioned it in my last post that I was going to probably do a 5CP map for comp, but I decided to revive cp_frozen and finish that. I'm basically rebuilding stage 2 and 3 while fixing up stage 1. This won't interfere with whatever my summer project is, I plan to do both of these together.

I still don't know what I'll be doing for this summer. Hopefully it will be a collaboration project with a programmer friend. Though, he might have a job, so who knows. We're looking at doing something with the Source engine, but nothing specific has been decided. If something doesn't happen, I'll be doing much more work with Radiant, as a way to push usage of that engine (Using Call of Duty 4: Modern Warfare 1).

Interesting thing I saw was that Borderlands 1 might have modding tools somewhere, so I may look into that for UDK work. I haven't found anything yet, but I'll keep looking.

The portfolio is coming along, I'll be finishing it up when rust gets finished and released. Don't have a domain name or host up yet. I may also include a "How to use Hammer" tutorial series on the portfolio.

Anyways, tis' be all for now.

Saturday, March 10, 2012

pl_rust update and internships

Well, spring break was fast. I was able to get about another third of rust detailing (first artpass) and more framework on the portfolio done (coming soon, I swear!). I would have had more done on rust, but the wonderful artist block popped up (and my birthday). I still intend to have it art'd and tested by the 30th of April.

Another oppurtunity popped up, I had a friend shoot me an application for Raven Software level designer internship. I'm jumping on this as a way to learn Radiant, though I do not intend to submit an application. If I do, I doubt I'll get an interview, let alone accepted and add that to financial reason (cursed College student finances! *Shakes fist*) I probably wouldn't be able to do it this summer. But I'm still going to learn Radiant, and look forward to possibilities later on.

I believe my next project will be a 5CP map, as my "final" tf2 design project. This will be designed for highlander and 6v6 competitive play. I've decided to do a competitive map because I feel that it is a very good design challenge. Other possibilities (if I don't do 5CP) is 3 stage A/D (not comp) and a steel-type map (comp).

For now, here is an image from pl_rust. This particular area was nasty to detail because it was originally designed as an interior area.

Saturday, March 3, 2012

Spring Break Agenda

Spring break is now upon us here in Happy Valley. I'll be traveling home, but I'll be bringing my computer with me so I get as much as I can done with pl_rust and my portfolio. For rust I'm basically going into super-detail-mode and for the portfolio I'll be populating the map pages and re-writing my tutorials and adding some guides.

For the map pages, I'm going to do something a little different than other designers on their portfolio. I'll provide plenty of images from the map, laid out in a nice gallery, but I'm also going to outline the though process on why I did what I did. Basically going to analyze my own map from a designer stand-point.

In the way of guides and tutorials, I have a couple of ones I'll be getting together. One is on lighting and using lighting to help guide players and help them recognize important areas, the other is from TF2 and how "smaller" maps do not mean faster or better gameplay. That article will cover mainly payload and payload race and probably general control point gametypes.