Custom games have always been popular in the Dota community: they break the monotony of the original while allowing the players to hone their skills and practice every now and again. Few custom games are actually successful and there is only one that I consider truly legendary – Auto Chess that even inspired Valve to release a separate game called Underlords.
And even though the majority of players switched over to the official successor, this map’s online activity hasn’t dropped below 5k players for over a year – at the peak of its popularity exactly one year ago 200 000 people were playing it at the same time. Even stand-alone games don’t always reach that level of online activity.
And now it seems the Chinese geniuses have created yet another masterpiece. A map called Custom Hero Chaos was released in February this year and quickly became popular among streamers and regular users alike. The gist of the game is that eight players begin by choosing their characters out of three random heroes, then upgrade up to five skills of their choice which drop randomly in packs (Ability Draft style) in the course of the game. Depending on the particular hero and the acquired skills the players choose a defense strategy to survive 40 creep waves, the last one featuring Roshan as a raidboss.
Players start with three lives that can only be lost to the creeps. Ultimately, the game is really about creep defense, not winning duels. But because of how the economy operates those to things are closely related. In every round (wave) a couple of heroes entertains everyone else with a fair duel while the others can place bets on the outcome, and the calculation factors are pretty interesting – it’s possible to get hella cash by betting on an underdog and make an epic comeback or break away even further.
Various strategies and builds are available here, one could go for physical DPS with useful passives for creep defense or a pure mage/caster, create a badass invincible tank or combine all of that to achieve the best result.
The game also has some serious drawbacks. First of all, skill distribution is completely random, so a good half of the players usually quit during the first part of the game as soon as it becomes clear what to expect. The abilities should probably be split into groups (by passive type, creep def, targeted spells) and the packs should contain a skill from each group.
The biggest problem, in my opinion, is the existence of overpowered skills that make the game really imbalanced (Spectre’s Dispersion or Viper’s Nethertoxin, for example, or Wraith King’s Reincarnation can all single handedly win the game in certain situations).
On the other hand, the existence of worthless or ineffective abilities which either have nothing to do with this game type or just become useless by the 20th wave or so (Lina’s Dragon Slave of Dazzle’s heal, for example) is also pretty annoying. I think the players’ choices of abilities should be analyzed and the most useless ones should be excluded.
And to top things off, some improvement of items or an option to upgrade the existing ones by combining them or buying something in addition would be nice. All in all, this custom game turned out to pretty cool and interesting, there are usually 10-20k players online 24/7. Still, there is room for improvement in terms of balance to prevent half of the players from leaving at the very beginning and the remaining three or four players fighting for the rest of the game – a rather boring activity considering that the comeback mechanic doesn’t work very well in this scenario and those who quickly managed to get useful items and powerful skills leave everyone else far behind.