Skills systems in Mobas part 2

I already wrote a bit about why I don’t think the skill progression systems in Mobas doesn’t really deliver on their goals. Here’s a bit on what I think are some helpful things things to consider when approaching this problem.

No false choice.

One solution is just to remove this aspect of the character leveling system. Give players all of their abilities at level one and have all abilities progress at the same rate automatically on level up. This may seem to lose all of the customisation options for the player but in reality, there was only very little customisation to begin with. I talked about this previously.

There’s probably only one or two best build orders for each character that only very rarely change. Since the player is likely to level their abilities in the exact same order every time they play a character then really it makes more sense just not to give players the option to do it wrong.

This is one of the decisions that Blizzard made with their entry into the genre. Since the order in which you level your abilities is a false choice more than 90% of the time. Just remove it and have all abilities level at the same rate.

This has the pleasant side-effect of making the early game more aggressive and dynamic (as everyone has all access to all of their skills).


I think the reason that these systems don’t doesn’t provide good customisation options is that they’re too easily solved. There aren’t a lot of interesting decisions in the skill order because players are very quick to figure out the “best” order and then they just do that. Remember that there are several million committed, highly engaged players pooling their resources to figure these things out.

Even if players can’t figure out the absolute best skill order and perfectly solve the system they’ll usually find one that’s good enough to use in every game. Typically by looking it up online. A system that provides so little engagement for the majority of the player base just isn’t pulling it’s weight. It’s not adding enough depth to justify it’s complexity.

Skilled players quickly develop heuristics that let them determine whether 20 more damage on ability A is better than 30 more damage and a reduced cooldown on ability B. Much as they do for knowing whether +X damage is more valuable than +Y health. One way to circumvent this is by preventing players from using these heuristics by providing upgrade options that defy numerical analysis.


The best way to do this that I’m aware of is to use “incomparables”. By providing options that cannot be directly compared to each other it becomes very difficult to find the best choice. It’s much harder to determine whether it’s better to have a brief damage shield or a few seconds of invisibility. Or whether it’s better to have a short teleport or be momentarily intangible. Or whether it’s better to have your damaging ability apply a brief debuff or chain to other nearby targets. Or whether it’s better to have a brief activated movement speed buff or a short duration stun. There are real, interesting decisions to be made between options when customisation includes differences in kind rather than differences in scale.

Much of these decisions will depend on the situation, the composition of teams and on a player’s unique play style. Factors that can be different in every match.

A robust customisation system would have to ask players not only which abilities to upgrade but also how to upgrade them.

Incomparables in action (Heroes of the Storm)

To replace the ability point leveling system. Blizzard allow customisation by way of a system of talents. At first level and (roughly) every third level thereafter (1, 4, 7, 10, 13, 16, 20) players are given an option to select one of between two and five talents. The talent system makes heavy use of incomparables. When given a new talent choice players will typically have a choice of one new activated ability, one passive trait and one or more upgrades to their standard abilities. These talents rarely offer a strict numeric increase. Instead they offer a variety of utility effects such as increased range, reduced cooldown or a secondary effect.

Often these talents will differ in the gameplay challenges and play style to which they cater.

E.g. The Ice Mage Jaina has a passive ability which causes anyone hit by her abilities to be temporarily chilled (applying a slow). One of her talents reduces the cooldown of her straight line, skillshot nuke by two seconds in the event that it hits a target that is already chilled. This allows players that have good execution skills to play a technical ‘kiting’ game with this ability. Using the reduced cooldown to continually reapply the slow. The talent is stronger in the hands of an experienced player and it changes the way in which the character is played. The decision involved in taking this talent is more about how a player wants to play than which choice is numerically best.

Looking at data on a website like we can see that there is much more diversity in the builds that players actually use in Heroes’ as compared to DotA2 or League of Legends.


While Heroes’ talent system does a lot to foster diversity in game it does have some of the same problems. Many talents are “traps” that are too weak to be worth picking and there are many others that are clear ‘best in slot’ talents that players will always pick at a certain level.

There’re two main reasons for this. One is the poor/inconsistent use of incomparables. While many talents do a good job of avoiding strictly numerical bonuses to focus on playstyle differences and utility. Some others do provide a clear numerical bonus to a skill or a close analogue.

E.g. Jaina has a talent that causes chilled targets to take an additional 15% ability damage. Since all of jaina’s abilities apply chill this is tantamount to a flat 15% increase in all damage for Jaina. It doesn’t require a change in playstyle or any decision making to activate. It’s a simple numerical increase that is always active. It’s the opposite of the kind of interesting customisation choice that makes talents a worth-while feature. It is overwhelmingly the most picked talent at its level because it has a very clear value to players and a very low ratio of ‘effort in’ to ‘power out’.

When players are looking at two or more talents that add damage to an already damage focused character it’s very easy for them to figure out which talent will add more damage. Even if the way in which the damage is dealt is quite different. These aren’t really incomparables in any helpful sense.

Incomparables and ‘Role’

The second reason is that many talents are the only talent in their category that support the character’s “primary play style”. You can think of characters as having one of several possible roles: Sustained damage, burst damage, Area of Effect Damage, Healing (sustained, burst and AoE), crowd control, damage tanking, etc. Talents and abilities can be grouped into similar categories.

In any case where a player has a choice of options of equal power they will always choose the option that most closely matches the role of their character. Correctly, since that option is usually the most powerful.

E.g. Nova is a burst damage assassin. Focused on attacking characters on low health and killing them. At first level she has four talent options. One for mana regen (taken close to 0% of the time), one for better crowd control (also rarely taken), one for reduced cooldown on her primary damage ability (this gives the most damage overall – in theory) and one that gives a situational damage bonus to her primary damage ability when leaving stealth (80% pick rate). In practice nova players almost always favour the ability that deals the most burst damage. The reason is obvious, this ability lets Nova do what she was already going to be doing even better.

Nova has another talent that reduces her attack speed by to one third but increases damage by 300%. This ability gives no net increase in damage but is still her priority pick in more than 50% of games despite competing with another very powerful damage increasing talent. Nova players will trade less damage (general role) for more burst damage (specific role). This pattern holds across all heroes. Healers will trade other viable options for the option that gives the most healing. Tanks will trade interesting utility abilities for more durability. Players rightly determine that they have a specific job to do for their team and the best options the ones that allow them to do that job better.

There’s an exception to this. If a character is already sufficiently good at their role that they don’t need to do it better then they can diversify. A burst damage character who can already deal 100% of an enemy’s health in damage can diversify into utility or escape abilities. A tank who can already survive extreme damage can think about customising for more crowd control or attack power.

A solution to this problem is now clear. Ensure that all characters are able to perform their primary role adequately without talents or upgrades. Then remove any options that benefit that primary role. A burst damage dealer should have enough inherent burst damage that they don’t feel the need to customise for more. Freeing the player to choose between a number of secondary roles that provide escape abilities, durability, mobility, sustained damage, crowd control or healing. Similarly a healer should not have the option to customise for even more healing (or that becomes their only “real” option) instead they should start with all the healing they need and have the option to customise for more survivability, utility or damage.

Take Aways

So hopefully what I’ve discussed here has shown that the League of Legends / Dota2 Hero progression model doesn’t provide customisation or depth for the player. At best it is a knowledge barrier that requires players to research the correct build online before they play a game with a new character.

Hopefully also I’ve pointed the way towards a possible improvement by using incomparables to provide meaningful character customisation decisions that resist solution by using differences in kind not differences in scale.

Also, I pointed out some potential pitfalls to avoid and a guideline of avoiding options that apply long term stat bonus or are universally useful. Instead focusing on situational upgrades that encourage versatility and player choice. And a note to avoid creating upgrades that feed directly into a character’s main focus or “primary role” as these will always strictly dominate other options of equal power.


Leave a Reply

Please log in using one of these methods to post your comment: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s