Competitve Gaming — A.K.A. The Fewest Random Factors

I once read a post on a particular forum stating that “any game or mod can become competitive, given time.” I then started actually thinking about just WHAT makes a game capable of being competitive, and why eventually, given a non-distinct amount of time, any game could possibly become a candidate for “competitive play”.

What I guess is that with time, people become accustomed to the random factors that may reside in a game, and thus players with the right skill set have adapted enough to the point where the random factors no longer strongly affect a skilled player.

Let’s take an example; id Software’s Doom. Still actively played today (though not in professional, sponsored tournaments), Doom has been ongoing with competitive gameplay for many years. I estimate that around the early turn of the millennium is when the combination of modernized, useable modding tools, source ports, and player base all merged enough to form a competitive field for the game. Take this with a grain of salt, as my knowledge of the full-blown, global, internet-based competitive scene is a bit limited although my strong background with the Doom communities has given some reasonable foresight into this (CSDoom’s source emerged around early 2001, which is the time ZDaemon and SkullTag emerged as the go-to multiplayer source ports).

Doom’s fast-paced action made it a game of accuracy and reflexes…for the most part. Players moved at insane speeds and weapons did a lot of damage most of the time. Capture The Flag (CTF) and Duel modes are the most common competitive modes, and while Duel focused greatly on the combat aspect, CTF focuses on a team aspect. Doom does not have many illogical random factors in it, but there is one that stands out: the RNG used to determine weapon damage.

Doom’s weapon damage formula for guns (any weapon that shoots hitscans and puffs or bulletpuffs) was


This doesn’t have a huge impact on the main weapon most players used–the Super Shotgun–but for some weapons that don’t kill in a single hit in close range, like the Shotgun or the Chaingun at any range, there can be some luck involved in the outcome. While I don’t intend to spark any outcry from any Doom players who read this, it is a matter of truth that there is a tiny bit of luck involved in how much damage the weapon deals. I’ll give it that about 95% of most scenarios will not matter, as generally one guy is skillful enough to make the kill no matter the random factor, but in some instances it can have an effect. From a professional, completely objective standpoint, any sort of complete luck involved with dealing enough damage to kill your opponent generally makes a game unable to be taken seriously at a very highly competitive level.

So what is my point with all of this? Well, what I’m trying to get at is that for most games to be entirely competitive, the best choice of action is to eliminate as many random factors as humanly possible (within the means of game balance, of course). I’m going to list a few pointers I’ve come up with that may keep things in line (along with some balance tips as well).

–Never, ever have random damage
This sort of goes without saying, from my example mentioned above. Skilled players will instinctively know how many hits their opponent can take before dying, and as such, randomness played into that makes it impossible to actually determine that. This means no random damage for EVERYTHING; guns, melee, etc. However, introducing a weapon with randomness as part of it’s play style or as a feature of the particular weapon may prove interesting and can be acceptable, though it will still likely not be allowed in a competitive setting.

–Stick with a general system for your weapons, and make unique ones differ from it
For instance, say you decide to make all your guns deal reduced damage the further away the target is. This basically means that all your weapons require close-quarters fighting to be effective. This makes combat really boring. So, what you can do is introduce a weapon (say, a long-range rifle) that actually deals normal (and by normal, I mean above-average) damage at all ranges, but actually deals LESS damage up close. This changes the player’s play style to that of a long-range role, and mixes up the gameplay considerably.

–Random spread is OKAY for competitive play…within limits
Let’s take the previously mentioned system–all weapons dealing less damage the further away the target is–and also add random spread on top of that. This would make the required close-range combat extremely frustrating! Sometimes you’d land a shot you aimed correctly, sometimes you don’t. It factors in luck and luck is not competitive because it interferes with skill. Removing random spread in this situation is generally a good idea, because then one’s aim at close range actually allows their skill to be fully realized. This is likely the primary reason Team Fortress 2’s competitive play disables random spread, as it uses the exact damage model I have described. ON THE OTHER HAND, if you have a damage model that uses the same damage per hitscan at any range, then random spread is acceptable. This is because then, assuming they had a reasonable amount of thought put into them, the weapons will be designed to have random spread that is equivalent to their assumed effectiveness of range. For instance, a shotgun may have extreme (random) spread, but at really close range, which is where it would do the most damage and thus be the most effective, the random spread is almost a non-factor. However, randomly spreading at longer ranges, thus more likely missing your target and being generally unreliable for damage designates that longer ranges are outside of it’s specialty, and as such random spread can only help balance such a weapon.

–Sniper Rifles should never, ever be used like Shotguns
So you decided to add a semi-automatic sniper rifle that kills in 2 shots to anywhere on the body (or in 1 shot if hit in the head). You figure it needs some balance, so you make it zoom in a lot, giving the player a narrow field of view, and you make it spread immensely when you aren’t zoomed in. You give it a 10 round magazine, because you want the player to have enough shots that spamming it is okay, but generally not encouraged. To top it off, you also give it very high recoil after each shot when zoomed in. Then you play test it with people and find that they are using them like a shotgun; running in close to their target and spamming the bullets off as fast as possible until the guy drops, because your unzoomed random spread isn’t enough or the damage is just too high. What do you do? You can’t give it some kind of fixed spread pattern because people would eventually figure it out, and increasing the spread further doesn’t do much, and may cause for extremely unrealistic shots (bullets don’t come out of the gun at 90-degrees…). Your options appear to come down to this: either reduce the overall damage, reduce the rate of fire, or give all players a means of instantly killing you if you get too close (which is what Call of Duty went with).

The best option is actually D. You create a revised damage model for that particular weapon (and/or any that end up like it), where it actually does less damage up close, but does full damage past a certain distance (a distance that would deemed suitable for a sniper rifle, whether that is medium or long range is up to you). The idea behind this example is to get creative with balancing your weapons so they actually fill a particular role. This will enhance the sense of variety for your weapons, and then players don’t feel as if all your weapons are too similar. It also helps prevent one weapon from feeling like it’s the best weapon above all (unless it is intended, in which case different balancing rules go into that).

–Playtest, Playtest, Playtest!
If my experience in making the WhoDunIt mod has taught me anything, it’s that even the slightest change to a weapon system can break balance completely. The two biggest examples from the mod are the Shovel and Pool Cue. The Shovel, prior to some changes I made (read: actually fixed) to the stun system (in the mod, if you were struck by a melee weapon it would reduce your move speed a bit for a time, the time could be stacked with multiple hits to a certain limit), was perfectly balanced and was the only weapon that hadn’t undergone any changes since it was introduced in an ancient, very old beta release. However, after fixing the system to work as intended, it proved to be WAY too powerful, able to not only stun lock the murderer or other innocents, but also dish out 40 damage a swing against the murderer! People quickly exploited this and the shovel became one of the strongest weapons in the game, when my intention was to balance them all as best as possible.

The other example is the Pool Cue. This weapon, from the start, was meant to be a supportive stun-based weapon, possessing the highest stun per hit in the mod, but with low damage output. Back before I fixed the stun system, it was the worst weapon in the game. After I fixed it and gave it some buffs (not realizing just how well I had ended up fixing the stun), such as slightly more damage, increased attack speed, and longer reach, it ended up being the best weapon in the entire mod, sometimes even outdoing the guns! Even sub-par players could solo the murderer in a 1v1 fight, despite the fact the murderer should win almost all 1v1 fights as long as A) He has decent aim with the knife, and B) the innocent has a melee weapon. The murderer’s knife does a maximum of 40 damage in a hit, attacks relatively fast, has decent reach, can backstab for instant kills, and it increases the murderer’s run speed while drawn. The only thing it doesn’t have is any form of stun capability, which is purely for balance reasons (along with making noise while drawn and being visually obvious, clearly indicating the guy wielding it is the murderer). However, the pool cue, doing a maximum of roughly 30 damage a hit, having the longest reach of any melee weapon in the game, and attacking about as fast as the lead pipe (the fastest melee weapon in the mod), on top of inflicting up to 5 seconds of stun time (which, in the code, was a 30% reduction in move speed; significant enough to prevent people from fleeing) made it entirely overpowered. And in reality, with the previously broken stun system that didn’t work correctly, all other changes would have been perfectly fine. But because I had fixed a feature that actually allowed it to work properly, it became a monster.

The point is: Playtest everything, because even the slightest change or fix to something could potentially cause a complete imbalance unintentionally. If I had managed to playtest the mod a lot more before releasing it, this situation may not have happened.


And there you have it, a few tips or ideas about balancing a game and potentially making it competitive. There is certainly a lot more that could be discussed here, but I’ll leave it at this for now since I don’t intend for this to be an end-all guide to designing a competitive game. :X

Leave a Reply