Retrospective Design: Nomad

Tremble and Nomad, as I like to think, were the bastard children of our design team. They came at a crucial period of transition–we were moving from a sporadic, “when it’s done” schedule for hero production to a regimented, “one hero per two weeks” schedule. Even though Tremble was in development for a month, a significant portion of this time was wasted due to haphazard patching, and design leadership being away from the office during the first vital week of live testing. This led to Tremble taking an exceptional amount of time to complete, both in concept and for art, as we were rushed and pressed for time in the face of a new hero-building process. As such, Nomad was a very short period of time to mature in live testing; roughly five days, the shortest of any hero we’ve released.

I’ve always felt like Nomad was one of my best designs, save for the passive, which was something I’d never felt satisfied with. It exemplified the playstyle I most hated; one-shotting. I highlighted most of my disappointment in a previous blog post, but the skill continues to haunt me due to its absurdity; it’s not imbalanced by any means, being on a melee hero that largely can’t harass with his autoattack, but it made a hero that was intended to exemplify mind games and trickery to instead be reliant on building pure damage and trying to snipe kills. Players who dislike Nomad tend to point out his ability to dish out massive damage in an instant, damage that’s entirely the result of the passive adding too much damage to the ability. It’s a problem we’ve wrestled with on Deadwood as well. The hero isn’t necessarily imbalanced–a common comparison is Pebbles, who can output as much or more damage at a faster rate, yet Deadwood draws much more criticism. Rather, the problem lies in the distribution of damage on the hero–that is, even though the overall damage output may be similar, or even less, than a hero that uses two or more abilities to deal damage, the perception that the majority of damage results from a single ability is highly frustrating to players, as it induces feelings of being cheated, or being beaten by a hero that requires less skill than their own.

What would I do now?

I don’t know. I don’t know what changes I would make to Nomad to fix this, as his spike damage output, like Deadwood’s, has been so ingrained into identity that any changes to the ability would effectively create a completely new hero. I still firmly believe that one-shotting isn’t conducive to a well-designed, well-planned game. Regardless, there are still many concepts from Nomad that I enjoy.

Posted in Uncategorized | 31 Comments

Retrospective Design: Tremble

Hindsight is 20/20. Deadlines aren’t! Let’s see what I did wrong…

Tremble was the first hero of mine to be implemented, and was rife with problems. We selected him for the interactivity of the Mounds and teleportation, but the rest of the design was largely unfinished, as he was my newest and least thought-out design (coming after Midas, Sylph (unreleased), Cerberus (scrapped), Astaroth (scrapped), and Nomad). His development occurred during a painful transition process, as he was the first hero to fall into the new “one hero per two weeks” schedule. The original intent of the design was to create a proxy for DotA’s Broodmother, but incorporate some unique mechanics, primarily in the addition of teleportation between webs. I’d long had a fascination with semi-permanent summons, especially structure-like summons, which create a very interesting dynamic of dependence between the character and his summons, making positioning and planning extra important. I’d always perceived Broodmother’s webs to be problematic due to their lack of direct counter, but this perception led to a particularly fatal design decision with Tremble’s Terror Mounds–they were given health. This sank Mounds into a balance quandary, as destructible summons must be more powerful than indestructible summons simply due to an availability of a counter. Terror Mounds needed to be useful early, but not completely phase out later when teams roamed together and were able to quickly destroy them. During the laning phase, the presence of the powerful destructible Mounds created a psychological calling that invited players to constantly try to kill them. The problem induced here was largely in the mind of players, as they would spend so much time trying to kill it that they ignored harassment, creep kills, or the more direct counter: Wards of Revelation. I learned an important lesson here about summons:

  • I: Invisible and destructible: Summons in this category are the most annoying to the player, as a clear, hard counter exists (detection). This typically means the summon has to be fairly powerful to accommodate for players countering it a portion of the time.
  • II: Invisible and indestructible: This is slightly less annoying than the above, as they are only partially countered by detection. However, the fact that they have two layers of strength makes this combination the strongest.
  • III: Visible and destructible: Visible and destructible is the weakest combination of properties, and generally these summons must have a strong passive, lingering effect. Tremble’s Mounds fall ionto this category, and they were hence given plenty of health.
  • IV: Visible and indestructible: On the other hand, a summon that is visible and indestructible is perhaps the least annoying for an opponent. They see it and understand that they can do nothing about it, so therefore accommodate their play around it.
  • In terms of player annoyance, from least to greatest: IV, II, I, III
  • In terms of post-balance power, from least to greatest: II, I, IV, III

Not only are Tremble’s Terror Mounds the most annoying to players, they must also be very powerful to accommodate for their visibility and destructibility. In retrospect, Tremble’s Mounds should have been visible and indestructible to alleviate frustration, then balanced from there. While the ability to counter his teleportation is important to his current form, the amount of time required to track down his Mounds ends up reducing overall action in the game. It’s about taking the lesser of two evils in order to foster a well-designed game.

Here’s what I would do now:

  • Terror Mounds: now invulnerable and unselectable, movement bonus reduced slightly
  • Terror Port:  reveals you for the duration of the channel, tunneling effect plays at the destination Terror Mound during the channel
Posted in Balance, Design, Heroes of Newerth | 12 Comments

Reddit AMAA Friday

MsPudding and I will be conducting a Reddit AMAA (Ask Me Almost Anything) this Friday from 8-9PM Eastern US Time (UTC -5). I’ll link to it in this post when it’s up. Just a heads up!

It’s up!

Posted in Uncategorized | 8 Comments

Hero Development Work-in-Progress: Druid #3

This is the third and final piece detailing the continued development, and eventually death, of the Druid concept in HoN.

We left off three months ago with the Druid, then named Artemis, in a very peculiar spot. At the time of publication, the publicized design was already a week-old, and as many savvy commenters pointed out, was rife with problems. What happened the month-and-a-half post-publication was a frantic search for any possible way to salvage the design, and unfortunately I came to the conclusion that under my current design directives for the future of Heroes of Newerth, she was simply not feasible. Allow me to explicate.

A lot of players have picked up on a concerning trend among hero designs for quite a while now, primarily that newer heroes are equipped with a generalist kit capable of tackling too many diverse situations. This complaint can be simplified into the generalization that new designs had “everything”, with everything defined as both crowd-control and escapes. Secondarily, they point out that newer abilities were too complex, with too many effects. A common argument here was that a hero like Witch Slayer could be effective with only one or two effects (i.e., damage and stun) on each spell, whereas newer heroes had the tendency to be overloaded with three or more.
There’s certainly an element of truth to both arguments, but it’s extremely important to understand what contexts they are valid in. Most players tended to propose that generalist kits are poor design because they make a hero too strong–this isn’t necessarily true, as numbers are what ultimately makes a hero strong. A simple retort based on empirical data would be Chronos, who has two stuns, a slow, one of the highest base damages in the game, physical and spell “evasion”, and an escape. On paper, he’d be the best hero in the game, but in practice, he’s quite lacking in his current state primarily because of numbers. I subscribe heavily to the idea that everything can be balanced; however, it is poor design when extremities must be taken into account to accommodate for insane concepts. A related complaint that has a hint of truth is that there are too many heroes that can stun, and too many heroes with escapes. This is a faulty argument primarily because the number of heroes picked per game has always been 10, and will be in the foreseeable future. The availability of more options does not preclude the viability of existing options unless the new options are simply superior to existing options. In this case, the core problem points back to numbers and balance, not the core design. What I do find to be an acceptable argument is that of prevalence and creativity. When players see that we release too many heroes with what they perceive to be similar, generalist kits, that to me is a problem as a designer, as it impedes on the perception of variety, which in turn affects fun. And having unfun designs is unacceptable! So I created two guidelines for hero design to follow.

  • Avoid escape mechanisms.
  • Avoid stuns.

This leads us back to Artemis. In designing her, I discovered a quandary that would become effectively unsolvable without breaking one of the above rules. Artemis was, at her core, an incredibly flexible hero. General design principles would dictate that you needed a certain level of flexibility to justify the multiple forms, or you would be better off leaving the extraneous mechanic out–Fayde was a great example of forms being more detrimental than positive to the overall gameplay. Though I wanted to stay away from generalist kits, the on-the-fly form-changing aspect was different enough from existing designs that I felt it was justified. However, I understood that public perception may not be the same; Master of Arms was just released, and we had another form-shifting hero in the works that would’ve been released in very close proximity to Artemis. This in itself was not a strong deterrent, but the implications of a generalist hero with two vastly different forms made it very difficult to design for.

  • Artemis was intended to have two forms: a ranged caster form and a melee damage form. I tried multiple ways to differentiate the two while keeping both useful; I assigned CC and utility effects to the caster form, while the melee form was given damage. Doing so however forced players to engage using set patterns rather than adapting on the fly; this effectively meant that the form mechanic served more as a “scroll bar” to the next sequential set of abilities than a shift of mentality and playstyle. I tried giving the ranged form active abilities and the melee form passive damage boosts; this approach was much more sound, but it also made the hero seem incredibly bland.
  • I ran into a particular problem designing for the melee form. As mentioned earlier, I wanted to avoid pseudo-blinks and stuns; however, such skills would be the perfect bridge between range and melee form. If there were ever a panacea to the creative block, this would be it. I was not, however, willing to compromise my directives for my own design, as I felt it would set a poor precedent not only for myself, but for the rest of the design team.
  • At some point, [S2]Buro pointed out to me that there were certain elements of the design that were unique enough to warrant being the core mechanic of a standalone hero. I realized this was true; throughout the growth of Artemis, and in my attempt to make her a unique design, I’d added so many bells and whistles that she was reaching a level of complexity wholly undesired by me, and likely the community.

The combination of these factors was enough for me to call it quits. Artemis was in a position where the fatigue of solving an apparently unsolvable problem had begun to erode my excitement for continuing her design, so I pulled the plug on her. Three months of hard work down the drain, yet not all was lost; I took Buro’s advice, and relocated the two most redeeming abilities from Artemis (according to testers) and stuck them on an older design, the “Sylph”. I like to think that Sylph is my crowning achievement in HoN design for her simplicity, uniqueness, and synergy… she will also be the final hero I design for HoN. I hope you guys will enjoy her!

Posted in Design, Hero, Heroes of Newerth | 22 Comments

(Blog) Changes

Cue Tupac song.

The blog address has obviously been updated. You can also access the blog at

Also, WHERE THE POSTS BE NOME? They be in my heads, but I’ll get some juicy stuff out later this week regarding Artemis.

Posted in Uncategorized | 1 Comment

The Designer’s Folly

Design is not linear; there is not a simple scale by which you can rate it, and it cannot be dichotomized into simple good and bad. Rather, the elusive perfect design rests in the valley of an inverse normal distribution; that is, the theoretically perfect design tends to be the most difficult to achieve, while deviations from the norm become more and more common as designers adhere more to or less from standard design principles.

Good design principles don't always result in pleasing design.

This brings us to the designer’s folly–overdesign by way of overimplementing “good” design principles. One of the results this leads to is players being forced into a gameplay pattern based on the designer’s mindset and specifications, not based on the design’s potential. What makes this particularly dangerous is that it befalls primarily the most conscious designers who tend to overthink and analyze, then plan and respond to, each and every possible outcome and consequence of a design decision, thereby shutting down the possibility for creative, unintended use of mechanics.

The diction there is extremely important, as the distinction between proper versus intrusive response is what creates the most lasting impression on the end user. A hypothetical hero designed as a ganker may find a more appropriate gameplay niche as a powerful, perhaps imbalanced pusher. Part of the continued journey towards being a better designer involves tempering your own expectations for what players will achieve from your design, accepting that the novel path is in fact a sign of a healthy design and not an indignant rebuke of your own intentions. A poor designer would not plan for this, nor would he accommodate for it post-release–a lack of response. A good designer would plan for this, and either prevent the problem or fix it post-release after evaluating the live situation–an overresponse. An excellent designer would plan for it, then adjust the design accordingly to accommodate for all modes of play–letting the player decide where he wants to take the design. It’s not unlike watching a child mature; though you may be tempted to lend a guiding hand, sometimes it is best to go laissez-faire.

Posted in Design, Heroes of Newerth | 22 Comments

Hero Development: Midas

Although he’s being released this Friday, Midas was actually the first hero I designed and scripted at S2. Interestingly enough, he’s perhaps one of the least thematically different from his original concept, the “Alchemist”, which can still be found here on the suggestions forums.

Click to see the Spotlight video!

I never actually scripted the version of the Alchemist found at the above link; it became immediately clear to me that mana manipulation was not the way to go for a cohesive, fun experience. Instead, I set out to take the original concept of spell combinations, then go from there. At the time, I’d had no programming, scripting, or otherwise technical knowledge, so the Alchemist provided a unique challenge with his state detection and complex mechanics. To tackle the concept of “hitting two spells at the same time”, I created the Reagent mechanic, which was simply a lingering state that was applied after being impacted by one of the Alchemist’s spells. Each of his spells applied a different Reagent, and when one Reagent detected the presence of a separate Reagent, they would combine to create the golden Transmuted state.

The first working iteration of the Alchemist featured a different skillset than what’s featured in the spotlight, as well as what’s found in the Suggestions thread. His abilities were as follows:

  • Upheaval: The Alchemist throws a flask, raising a line of flesh-eating Mandrakes and creating an impassible wall. This was effectively a vector wall that would attack nearby enemies.
  • Restoration Wave: A wave of volatile energy springs forward, healing allies before combusting violently. This spell is almost identical to the final “Lion’s Pride”.
  • Mandala: This alchemical circle is imbued with harmful seals that punish the aggressive. Enemies standing on the mandala will receive an attack and movement speed slow.
  • Transmute: Enemies affected by more than one of the Alchemist’s spells at once are temporarily cast into soft, malleable gold. Transmute also provides a bevy of passive bonuses, including cooldown reduction and a cast speed increase.

The Alchemist sat in the hero pool for quite a while, and ultimately his skillset was made obsolete when Empath was given a vector wall, perhaps for the better. In terms of feel, the design lacked the distinct snappiness that Midas now provides–that is, both Upheaval and Mandala featured lingering gadgets that were unbecoming for applying the Reagents. This posed a conceptual, mental problem in that it was difficult to convey to the end user when exactly the Reagents applied–did they apply on initial spell impact, or were they continuously applied throughout the span of contact with the gadgets? I initially chose the latter, but it became abundantly clear to me that neither option, given the skillset, would be intuitive.

Roughly a year ago, we decided to do the Alchemist as one of the upcoming heroes (this plan was obviously deferred for quite some time), and so we needed to record his announcer sound. We do them in batches, and usually in advance. After compiling a list of potential hero names, we send them off to Don Morrow to do his epic announcer thing. That’s why we don’t have announcer sounds for some new heroes sometimes–they weren’t in the previous batch. I wanted to just go with “Alchemist”, since that’s quite the perfect description of his concept, but obviously the DotA hero of the same name provided opposition to the thought. DivA came up with “Midas” on the fly to resolve the problem and we ran with it. I took this opportunity to redesign the Alchemist based on the Midas theme.


Readers may remember this picture, which I posted a while ago. This was the working model for Midas.

His new skillset was as follows:

  • Solvent Salvo: Midas throws five flasks in a line, coating enemies in an armor-eating solvent. Compared to the final “Golden Salvo”, this ability had an additional armor-reduction effect.
  • Unstable Remedy: Effectively identical to the previous “Restoration Wave”, save for minor differences. Largely identical to the final “Lion’s Pride”.
  • Black Powder Blast: This alchemical powder is imbued with harmful seals that create an intense flash of fire. This ability was the basis for Amun-Ra’s Ignite, except it lacked a delay.
  • Transmute: Identical to the previous transmute, but without the cast speed increase.

With this redesign, Midas became a snappy, instant-feedback hero. The previous design issues were resolved. Before I was able to show off this concept to the other designers, I was asked to complete another hero concept–the Gravekeeper, now known as Amun-Ra. During Amun-Ra’s development, I felt that Midas’s Black Powder Blast suited Amun-Ra more, as it synergized with his tendency to be in the middle of a battle. So again, Midas was put on hold as Black Powder Blast was converted to Ignite for Amun-Ra. When time came to resume progress on Midas, he was given the new ability “Elemental Warp” to reduce redundancy between heroes. The delayed blink would facilitate the positioning for his other two skillshots, so it made perfect sense on him. Finally, he was complete.

Midas will arrive in Newerth this Friday after a very long wait–I hope you guys will enjoy him!

Posted in Design, Hero, Heroes of Newerth | 37 Comments