The feedback provided here is essentially a summary of the collective feedback that TL Strategy had on the maps. Before you come asking questions about your map specifically, make sure you've read and thought about whether these things applied to your map or now.
Firstly, let's talk about some general things that everyone should be doing in the contest.
90 Degree Overviews
Yes, slanted overviews look cooler. But for the purposes of judging distances properly we really need 90 degree overviews. I know that if I put this in the ruleset that people will still submit slanted overviews (because there were multiple people submitting 2p maps this time around...) so I'm not going to make it a hard rule. But it does greatly assist the judging team in being able to quickly assess your map and gives it the best possible chance to be considered. It's worth pointing out that by submitting slanted overviews you're not penalised (e.g. Uvantak's submissions both were slightly slanted) but it may have meant that some of our initial assessment was misguided -- don't let that happen to your map!
Including Spawn Positions
Often it's really easy to tell where the spawn locations are on a map, but if you're doing something nonstandard it doesn't hurt to include the position of the spawns. The reason is simple and I'll explain it through example. Take Providence by IeZaeL, there's no way the judges would have known where the bottom left and upper right spawns except for the fact that he included the location of the spawns. Another example, Into the Wilds has not totally clear spawn positions -- sure we assume that the spawns are in the corners, but that information is 100% clear. tldr; to give you map the best shot include spawn positions.
Why did my map get cut early?
Two main reasons. First, the proportions are off. By this I mean that your allocation of space is not at the standard we expect for the competition. Vast open blank areas, ridiculously inaccessible naturals/thirds, huge distances between bases etc. are all proportion issues which are things to be worked on for future seasons. This doesn't mean we hate innovation, its just one of those things that amateur map makers are highly inexperienced in and need more practice. Many top mappers started off making those mistakes too, over time you'll get a better grasp of appropriate distances.
The second main reason is that the map is obviously imbalanced. This was less of an issue this time around but is still one of the reasons maps get cut early. If your going to introduce an obvious imbalance into your map (i.e. double entrance naturals, multiple entrance mains, thirds a long distance away) then you need to have really thought through the implications of that concept and made adjustments to your map to reflect that. The maps that are cut early clearly didn't make those adjustments and hence were cut.
Rule of thumb: if you're still an inexperienced mapper adhere to the rules and try to understand what maps standard maps standard. Understand all the rules first, then you can create maps which break them. Breaking the rules first is a recipe for failure. Uvantak is a great example of this, he's someone who has progressively improved over many seasons of TLMC and he's done this through executing standard-ish maps at an increasingly high level. Do the same and maybe one day you'll get a double placing in a TLMC!
The following issues relate more to the maps that were up for finalist consideration but did not make it for one reason or another.
By now it should be well established that less than 40s for a nat2nat distance isn't going to work. I appreciate that with 4p maps that this is difficult (close spawns often end up breaking this rule), but you should be designing around this restriction. The reason why 40s is the minimum we'll accept (and even then, ideally its 43s+) is because the rush distance becomes too short making 2 base play stronger and making it more difficult to secure later bases. Remember: most attacks are going to travel from the natural to the natural or third to the natural so that is the effective attack distance -- the main2main distance is mostly for scouting purposes (unless you did a in-base natural, of course).
Some maps which failed to do this include Fragments of Time (close spawn), Nature Spirit (horizontal spawn) and Into the Wilds (all spawn). You'll know if your map fell into this trap since you needed to submit that information in your application.
One question everyone should be asking upon creating a map is "can Protoss FFE here?". If the answer is yes, the the natural design is reasonable. While not solely a Protoss issue (all races like standard naturals for a variety of reasons) this question is sufficient to resolve all difficulties races might encounter. This is a more common mistake in maps cut early, but also persisted in some of the maps which were trying to make the concept work. Unfortunately, adding collapsible rocks is not sufficient to make a two sided natural work (as there's no real way for a Protoss to collapse one side, thus failing the FFE test).
Maps which tried to make this concept work include Dominion Rose, Cactus Valley and Dragon's Nest. We feel that if you're going to make this concept work, more considerations of the difficulties that the double entrance nat to be reflected in the design of the map.
Yeonsu proved that Blink Stalkers are mighty powerful. Future maps need to learn the lessons that we learned on Yeonsu and not repeat those mistakes. If a main (or main/nat) is given too much exposure to Blink Stalkers then bad things are going to happen. As such you should be limiting the Blink Stalker accessible ground in each main to limit their strength. Additionally main/nat need to be positioned so that the defenders travel time (to defend blink stalkers) is shorter than the attackers, else the attack gains too much of an edge.
Maps which fell into this trap include Hunting Grounds and Coriolis and Providence (top left, bottom right)
1FF Ramp Mains
PvP and ZvZ rely on ramps having one forcefield width -- in ZvZ this means that the ramp can be blocked with 2 Queens which means that some of the strength of ling/bane is mitigated. Additionally, same height main-naturals can also cause many problems if they are not treated with extreme care. While this kind of feature alone would not disqualify a map (since ramps can easily be changed) it's sad that this basic element of gameplay is still overlooked by mappers.
Maps which fell into this trap include Shrine of the Fallen, Arcadia and (to some extent) Providence.
Low ground spawns don't work, ever.
Not so much of an issue this season. But was a reason why some maps got cut early. The concept just doesn't work. This concept was extensively tested when we were considering the original version of Keru. We found that it gave aggressive strategies far too much power and once a player lost control of the ramp the game was over. In addition warp in was incredibly strong in addition to bunker contains.
High ground can be very powerful, careful where you put it
High ground is often undervalued as a strategic point on the map. I assume this is largely due to people thinking that the lack of a high ground mechanic (aside from vision) isn't actually that big of a deal. Well, unfortunately it is. High ground can be a very powerful strategic position -- the old IPL map Darkness Falls hinted at the strength of high ground when the fourth bases could be taken by a Terran and with PF support secure a strong position against Zergs. This would offer protection for their rally point and provide a strong staging ground for future attacks.
Two maps which had high ground that was too strong were Nimbus (which suffered from the issues described above when Terrans spawned counter clockwise to a Zerg) and Nature Spirit. Let's talk a little more about Nature Spirit since that's something that we discovered during testing that we hadn't encountered before.
The high ground at 12 and 6 proved to be much to powerful in testing. While we knew that high ground over looking a natural such as in Lost Temple was broken, it turns out the same thing is true with third bases (at least in this instance). This is unintuitive, since you'd expect Zergs (or Protoss) to be able to defend against Terran drops controlling the high ground by the time the game has progressed to a third base. But as Medivacs are such a staple of Terran armies it means that Terrans aren't deviating from their normal strategies in a significant way while Zergs and Protoss need to do something quite radical (like 2 base mutalisks or rushing spire after three base leaving them open to a host of other Terran attacks). Ultimately because the medivac is so core to the Terran army and that around 8 supply of units can shut down a full mining base with little counter play this feature used around thirds is broken.
Accessibility of thirds
This is already somewhat covered by map proportions, but this is still used at a high level for no good reason. At this point in SC2's life we're at a point where the game revolves around third bases in the Zerg matchups. Often this means Zerg players defending off two base plays from Protoss and defending from Terran harass while doing what they can to limit Terran expanding. Moreover if Zergs are allowed free thirds on these maps, then the pendulum swings the other way and makes it incredibly difficult for Protoss/Terran to take thirds; which in turn just encourages players to play off of two base. If you have a reason to break this metagame then you need to have an exceptionally good reason and to be able to execute this idea at the highest level. Extreme care needs to be taken to ensure that Zerg players are not unfairly disadvantaged, which in most cases is just ignored.
I appreciate that the mapping community wants to try an innovate, but your innovations should be focusing on the map between the two players and not the first three bases. Play with fourths, play with the center, see what can and can't work there... but only extremely well executed maps which have made calculated decisions to compensate can break the standard three base mold.
Maps which fell into this trap include Exodus and Coriolis (which did at least try to mitigate some of the issues by using a gold base, but not to an extent sufficient enough for the strategy team be sure it wasn't broken).
This more often than not relates to the previous issue in terms of making thirds accessible. Another issue which makes them inaccessible is a lack of appreciation for rotational spawn imbalances. But rotational imbalances extend beyond that to include differences in medivac vulnerability, fourth accessibility and blink surface areas. To some extent we're happy to accept some degree of rotational imbalance, so long as the advantage isn't overwhelming.
Maps like Nemesis, Samsara and Hunting Grounds exhibit problems like this.
The great chasm of nothing
This also ties in with the third base discussion, it feels like map makers have become so fixated on the layout of the first three bases that the rest of the map feels like it is thrown together as an afterthought. Large expanses of "empty" space don't really do anything good for the game. Alterzim at the moment has large expanses of open ground in the center of the map and I don't need to tell you what impact that has on the game as games go long. The point of this is that the best maps go beyond just designing for the first three bases and continue to design the map to ensure play on the map continues to be interesting as the game goes long. Best example of this is Cloud Kingdom which not only had an interesting three base setup, but after that the 4th/5th base encouraged interesting troop movement and continued to make the game interesting (well, as interesting as BL/infestor can be).
Maps which fell into this trap were Dominion Rose and Colonial Province.
2-in-1 maps trying too hard
We saw a lot of 2-in-1 maps do the following in this contest. One set of spawns would be very standard and normally very playable. The other set of spawns would be some ungodly mess doing things that break the game. If the other set of spawns were standardarised these maps would have easily been up for contention. You don't need to be doing crazy things with the first three bases to make a map interesting!! Stop invalidating all your hard work in crafting a map by slapping on two spawns which are unplayable.
Examples of this include Providence, Chimaera and Jungle Remedy (to some extent)
Don't make your map too chokey/restricted in movement
Chokepoints are map features which we're all accustomed to because of forcefield. The strength of forcefield in small chokes is such that the best maps must account for this. This typically means sparring use of small chokes to create strategic areas of the map to do battle, as opposed to putting them everywhere without too much thought. Yet still we see maps which over use small chokes and make forcefield way too strong amongst other issues (notably making the life of Zergs hell).
An example of this can be seen in Purifier.
Swarm hosts add a new dimension of difficulty to designing maps. After a certain point in ZvZ/ZvP (and ZvT against mech) swam hosts become one of the stronger strategies in the Zerg arsenal (that probably goes without saying). One of the strengths of swam hosts (as a siege unit) is their ability to control space. Thinking about where swarm hosts might position themselves once you're at a four/five base stage will give you insight as to the strength of the unit.
Take Colonial Province where in horizontal or vertical positions the power of swarm hosts is quite formidable. They're able to control key locations behind two chokes which basically locks out the opposition from doing any direct engages (unless they're very far ahead). Making mapping decisions which allow players counterplay against swarm hosts is something that more maps should be thinking about.