Worlds of Design: The Revenge of the Kludge

Meet the Kludge: a tacked-on solution to a particular problem

In our previous discussion I explained my perspective on harmony in game design. Now it’s time for the opposite, but sometimes complementary, game design technique, the Kludge. It’s something of an ugly word itself, and that’s not an accident.

hammer-5318123_1280.jpg

Picture courtesy of Pixabay.

What’s the Opposite of Harmony?

I borrow the term “Kludge” from software (“kludgy” is the adjective that's used). A kKudge is a tacked-on solution to a particular problem, or a solution that works, but isn’t consistent with the rest of the program. In software (though not in games) a kludge can also be difficult to understand and modify.

The Kludge is hard to define in game design because one man's Kludge is another man's “nothing wrong with that.” How do you notice the Kludges if the game is a model of something, as RPGs always are? The Kludge will usually be inconsistent with the rest of the model, and may have nothing at all to do with what's being modeled. It may be there to fix some design flaw.

When I play games, I sometimes ask: "Why am I doing this particular thing?" If the only answer I can come up with is “because it fixes a design flaw,” or “because the designer liked it,” or “I have no clue why it's here,” then it is probably a Kludge.

Kludges in Abstract Games​

A Kludge is less obvious in abstract games because the game doesn’t represent anything (other than “a game”). Abstract games are collections of mechanics where there’s no attempt to model anything. This is different from a model where the context should help people play the game and the mechanics are expected to represent something that happens in a real or fictional world. Nonetheless, in abstract games you can have a mechanic that doesn't fit with the rest, that doesn't seem to have a useful function, clearly should've been replaced with something else, or simply should have been removed from the game.

You’ll often see this in early games that developed over time. Dungeons & Dragons is a prime example: role-playing mechanics were, at best, loose. There were hints here and there of how to role-play, but there weren’t a lot of rules to that effect. D&D originated from Chainmail after all, and didn’t even have a skill system at first. You could role-play all you wanted, and many groups did, but any rules that game masters came up with were a Kludge.

Advanced Dungeons & Dragons includes rules for training characters before they can level up. I’ve discussed at length why these rules make no sense to me. I think they were added to provide a way to syphon off the otherwise vast treasures characters collected. But the end result was that training turned adventurers into money-grubbers. Which is probably why those rules were deprecated in later editions.

There are those who feel today that D&D still doesn’t support other styles of play and any rules to that effect are Kludges; but clearly, given the success of Critical Role and other streamers telling stories using D&D rules and variants thereof, it’s possible. Whether or not it feels right to a game group likely determines how kludgy those rules feel.

The Land of Kludges​

Kludges are often added to games to solve a problem that appeared in testing. Most of the time it's added to fix a demonstrated flaw, but at other times, it's in the game because the designer liked it, even though it doesn't fit what he ended up with. (Remember, games often end up some “distance” from where the designer originally intended.) It could be the original idea itself, yet the game has developed in another direction. At that point, the designer should remove the original rule, get it out of there, but it's emotionally hard for a designer to do so.

For a board game example, there’s Catan. Catan is very popular, although I believe it has its own Kludges: the Robber card. Keep in mind there’s not a lot of interaction in Catan between the players except for the trading, and there's little you can do to actually hinder another player after the initial setup. I think the designer saw the lack of hindrance, and decided to add the Robber, which has nothing to do with the rest of the game (it restricts/prevents some resource acquisition). It feels to me like it was added to provide the potential to hinder other players without really integrating it into the game itself. If it represented mere bandits, a player’s soldiers should be able to do something about it.

Consider the online video games World of Tanks and World of Warships. In World of Tanks the entire idea of 15 versus 15 randomly assigned teams is a Kludge, in the sense that it has nothing to do with real warfare, but it's necessary to make the online game practical for a very large audience. In World of Warships the overall Kludge is to play in a small area, usually amongst lots of islands, places where real world battleships and aircraft carriers virtually never went. In both games we have the bizarre mix of nationalities of equipment: German and French and English and Russian tanks or ships on the same side, and possibly 15 individually different tanks or 12 different ships on a team, where military reality is uniformity. It's also a necessary Kludge but has nothing to do with reality. Both games break down as models of reality, and the Kludges are obvious.

It's Not a Kludge!​

When is a Kludge no longer a Kludge? When almost everyone accepts it as necessary. In big video games, both Harmony and the Kludge become obscured. Kludges viewed as downgrades to curb player power are often called “nerfs” to represent a sword that looks like a real sword but is just a foam weapon that does less damage.

It's easier to find things you think are Kludges in a game you don't like. Also, we have the limitation that some designers of puzzle-like games, whether they’re single player video games or solo tabletop games or cooperative games, tend to add things to make the puzzle solution more difficult.

Kludges are tricky to define. It doesn’t matter what specific mechanics you use, whether already very popular or brand new (the latter are increasingly rare). What matters is how they work together as a whole. I come in heavily on the side of this motto: “A designer knows he has achieved perfection not when there is nothing left to add, but when there is nothing left to take away.”

By far the best way to find a Kludge is to see a game through someone else’s eyes. Playtesting is key in this regard. Kludge is best found by people who are objective, or frankly don’t even like your game. It’s tough for designers to hear, but critics will find Kludges you can’t see, and it’s worth taking any criticism at face value and playtesting to find if your rules don’t match the game you intended. Designers need to recognize the inharmonious, and excise it!

Your Turn: We all know kludges in games we love to play. What’s yours?
 

log in or register to remove this ad

Lewis Pulsipher

Lewis Pulsipher

Dragon, White Dwarf, Fiend Folio

Dausuul

Legend
In chess, the ability for pawns to move two squares on the first move was a kludge introduced to speed up games. En passent was added as a kludge on top of the kludge, to preserve some elements of the older style of play.

In 5E, two-weapon fighting comes to mind. The use of a bonus action, the extra attack not getting your stat bonus to damage, and then the inclusion of dedicated feats and fighting styles to mitigate these rules... it's a mess. And 2024 managed to make it worse. Sigh.
 

payn

I don't believe in the no-win scenario
Seafall ended up being a high potential game that was kludged to death for me. I think any legacy game that has rules that are going to change and adapt is vulnerable to kludge destruction. A game thats intended for 12-24 sessions with major changes is going to be extremely difficult to play test. So, its a very intriguing idea these legacy games, but im not sure they can avoid the worst of kludge.
 

GMMichael

Guide of Modos
Modos RPG kludge: your attribute bonus increases or decreases following a pair of formulae, based on attribute score.

It's just clunky. The attribute score determines how much damage a character can sustain of that type, so the score can't equal the bonus or it would be too low. If the attribute equals the attribute bonus, a PC could add 10 to that to find the damage threshhold, but that's another calculation/kludge. The formulae make the attribute score half as influential on outcome rolls as skill points, which suggests that natural talent isn't as useful as experience or training. To remove the halving-process, the game could double all the numbers involved, but then the maths get larger, moving from kindergarten -level addition to 1st or 2nd grade addition, and the randomness of the d20 necessarily diminishes as the fixed numbers added to the result get larger . . .
 

aramis erak

Legend
By far the best way to find a Kludge is to see a game through someone else’s eyes. Playtesting is key in this regard. Kludge is best found by people who are objective, or frankly don’t even like your game. It’s tough for designers to hear, but critics will find Kludges you can’t see, and it’s worth taking any criticism at face value and playtesting to find if your rules don’t match the game you intended. Designers need to recognize the inharmonious, and excise it!

Your Turn: We all know kludges in games we love to play. What’s yours?

The drug rules in FFG Star Wars.

stake to the heart in BTVS... it is there to encourage action like the show... but it breaks the combat flow because you have to first hit, then check and see if the modifiers are enough to kill, if not, reduce it back to normal damage...

Initiative in WEG Star Wars 2nd edition... a desperately needed kludge... but close enough that few were upset by it.

D&D OE, HB, and BX: all hits do 1d6. (Hell, OE felt like a kludge to me overall.)

D&D 5e Inspiration and advantage/disadvantage.
 

EzekielRaiden

Follower of the Way
I've always felt that kludge has a more narrow (albeit not necessarily more specific) meaning.

It's a shoehorned solution. Something that clumsily, and generally forcefully, pushes play in a specific direction when the rules would naturally push in the opposite direction.

To give a pair of concrete examples from D&D 5e (which I think has a lot of kludgy/inelegant design), two things that both relate to the saving throw rules: Legendary Resistance and PCs getting proficiency in resistances. The first is more useful as an example case of a place where a kludge is knowingly and intentionally applied because an actual solution would require reworking a massive and poorly-structured subsystem (the spells), while the latter is more useful to show how a kludge can be the result of needing a "patch" for a problem discovered late in development.

In the former case, something the designers were aware of from a very early point in the "D&D Next" playtest was that big stonking boss monsters were incredibly weak in specific ways. In particular, even the much better-balanced 4e had shown that "solo" monsters were extremely sensitive to crowd control effects. Locking down an enemy early on could mean parties could have a reliable way to turn intended steep challenges into cakewalks. Knowing they had to do something, they went with Legendary Resistance: so many times, "boss" type monsters like beholders and dragons and such can just...decide they don't feel like failing a saving throw, so they succeed instead. It's a small set of "get out of jail free" cards for crowd control or ongoing damage effects. These accomplish their goal most effectively (given I just fought a 5.0 Beholder in my Monday game, I can assure you they do their job), but....well it feels pretty bad. Very bad, actually. It just feels like your smart plays are getting totally no-saled for no reason, so you know you should just not bother. That's not very fun. Instead of removing the no-fun aspect by doing something fun instead, we just swap one kind of no-fun (challenging bosses get flattened by saving throws too often) for another kind (don't bother even thinking about saving throws unless you can overwhelm Legendary Resistances.) This is an excellent example of a kludgy solution in action. It does what it was designed to do! But it does so in a clumsy, weak, un-fun way.

Kludge arising from needing to "patch" existing rules inefficiently can be seen in how 5e characters got their two and only two saving throw proficiencies. Originally, nobody was proficient in any saving throws. Instead, it was meant that your individual ability scores would do all the heavy lifting, thus being a reward for choosing to be more diverse rather than hyper focused on the "important" stats. The problem is...that just absolutely did not work with the monster abilities as designed. In a very real sense it was the mirror image of the problem above. Some folks had made minor mention of this midway through the playtest, but WotC either ignored them or didn't think it was an issue until an event I (and seemingly only I) called the "ghoul surprise."

The "ghoul surprise" occurred during an official, recorded test game meant to show off the system to fans, put it through its paces, get players excited for its potential. As part of that run, the party of four or five PCs (I don't recall the exact number) had to deal with some ghouls. The party had a small numbers advantage, and per the CR and encounter building rules, this should have been a perfectly reasonable encounter, not even Deadly. Instead, it was a TPK on camera, one that shocked the developers and kinda ruined their attempted "look how awesome this is" moment because it went so catastrophically wrong when it wasn't intended to. (I know some folks find this exciting, but their intended audience definitely did not.)

The real problem was that PCs failed saves too often, and the ghouls got multiple attempts to paralyze their targets every round (once per attack), meaning the PCs were essentially guaranteed to get stunlocked and thus die, without any ability to prevent it or fight back. This is, as I think most folks would agree, kind of a problem (even those who like the "stuff can go badly wrong" generally would prefer that a prepared party could do something, but that wasn't the case here.) Now, the designers nerfed the ghoul hard to address that end of things (it now only gets one claw attack per round and the paralysis DC is only 10), but the "ghoul surprise" proved that they needed to make PC saving throws better in order to mitigate this problem more generally. So...they gave PCs saving throw proficiencies.

Except this....didn't really fix the problem, because of the "all six stats are saving throws" design. Yes, it patched up the immediate problem, but it leaves most PCs abysmal at most saves. As in, even a typical PC is only going to succeed on that DC 10 Con save from the ghoul about 60% of the time. Compare this to something like feeblemind, which is an incredibly nasty effect, targets a save few characters will be good at (Int), and has a DC of 17, meaning most characters will have at best a 25% chance at success!

This is a crude patch over the problem, done crudely not because the designers specifically decided that that was the only option, but because there was very little development time left and they couldn't afford to re-tool saving throws again so late in the process. Had they done more and better testing, they almost certainly would have seen this coming and been able to get a more holistic or at least more robust change to the mechanics so a kludge wouldn't be necessary.
 

I think if you want to define this idea better, I'd go with the term ludomechanical dissonance, as defined by In this blog. A lot of whats discussed in the OP though is also just your standard ludonarrative discourse.

I think the whole thing about ludomechanical design choices, different from ludonarrative ones, is that they aren't intrinsically disconnected. A design choice could be satisfactory to both simultaneously, and that that I think tends to be a very desirable design to achieve.

That I'd argue is the purer definition of what I've described in the past as mechanics that make metagaming and roleplaying indistinguishable from each other. The optimal way to play follows from the ludic as much as it does the narrative and vice versa.

In video game land, Doom Eternal is exemplary of that kind of marriage, achieving it largely because it wasn't afraid to be a game to do it. You're the Doom Guy, and the best way forward is to rip and tear, and thats the case in the function of play as it is in the function of the narrative the game provides for, and it works despite the fact that much of the solutions are so blatantly gamey as opposed to diegetic.

But even then, that doesn't mean Doom Eternal couldn't be done in a way that makes that gameyness invisible. Plenty of ways to, say, give an authentic reason for any random demon being a loot pinata.
 

EzekielRaiden

Follower of the Way
I think if you want to define this idea better, I'd go with the term ludomechanical dissonance, as defined by In this blog. A lot of whats discussed in the OP though is also just your standard ludonarrative discourse.

I think the whole thing about ludomechanical design choices, different from ludonarrative ones, is that they aren't intrinsically disconnected. A design choice could be satisfactory to both simultaneously, and that that I think tends to be a very desirable design to achieve.

That I'd argue is the purer definition of what I've described in the past as mechanics that make metagaming and roleplaying indistinguishable from each other. The optimal way to play follows from the ludic as much as it does the narrative and vice versa.

In video game land, Doom Eternal is exemplary of that kind of marriage, achieving it largely because it wasn't afraid to be a game to do it. You're the Doom Guy, and the best way forward is to rip and tear, and thats the case in the function of play as it is in the function of the narrative the game provides for, and it works despite the fact that much of the solutions are so blatantly gamey as opposed to diegetic.

But even then, that doesn't mean Doom Eternal couldn't be done in a way that makes that gameyness invisible. Plenty of ways to, say, give an authentic reason for any random demon being a loot pinata.
The new Doom games definitely are an example of something being mechanically sound (and surprisingly enjoyable, at least for me) while not being especially invested in whether it makes total sense or not, though I would say some slight lore tweaks would be all you'd need to justify the loot pinata effect.

As an example, perhaps the souls of beings transformed into demons are rich in inherently violent psionic energy. The Praetor suit's design could tap into that violent energy if it's close enough and used in the right way, allowing it to transmute that violent energy into useful materials. But you have to be close enough, as psionic energy dissipates extremely quickly. This would also explain why the chainsaw is always capable of producing the loot pinata effect. It's pretty much by definition going to be the most gory, violent way a demon can die.
 

GMMichael

Guide of Modos
After some game design contemplation, I thought I'd come back and thank @lewpuls for this Kludge post. Examining my kludgy rule (posted above and below) has resulted in no less than three revised rules that could make my game more harmonious. Thanks doc!

For the sake of discussion, I'll repost the kludge here, along with three new alternatives, and see if I/we can write a mod (or just a rule replacement) that makes Modos RPG a bit more harmonious. The rule in question is Rule 104, Attribute Bonus:

  • Kludgy 104: an attribute provides a bonus to contests that depend on it, equal to the (attribute) score minus ten then divided by two if the score is even. If odd, the bonus is equal to the score minus eleven then divided by two.
  • 104v2: an attribute provides a bonus to contests that depend on it, equal to +2 if the attribute score is the PC's highest attribute, or -2 if the attribute is the PC's lowest attribute score.
  • 104v3: an attribute bonus applies to a contest when a PC's d20 roll is below the applicable attribute score. If the d20 result is below the attribute score, the PC adds the PC's character level to the contest result.
  • 104v4 (attribute IS the bonus amount): a PC adds the attribute score to all contests depending on that attribute. (This version involves rewriting the Max Damage rule, 313, and the Attribute rule, 103.)
  • 104v5: an attribute provides a bonus to contests that depend on it, equal to the attribute score less ten.
The idea behind the attribute bonus is that having a strong attribute (for example, Mental) makes a character inherently better at mental tasks, but training (skill) has a direct and greater impact on specific tasks (like Chemist skill used to create acids or ballistic foam). Version 5 is included above as a more harmonious example than the kludgy rule, but it would be difficult to implement because its effects would cascade into several other rules and intentions like the value of skill points vs. attribute points, the awards for level-ups, the relationship of die result compared to bonuses, and the manageability of maths at higher character levels. In other words, I don't like it 🤓
 

EzekielRaiden

Follower of the Way
After some game design contemplation, I thought I'd come back and thank @lewpuls for this Kludge post. Examining my kludgy rule (posted above and below) has resulted in no less than three revised rules that could make my game more harmonious. Thanks doc!

For the sake of discussion, I'll repost the kludge here, along with three new alternatives, and see if I/we can write a mod (or just a rule replacement) that makes Modos RPG a bit more harmonious. The rule in question is Rule 104, Attribute Bonus:

  • Kludgy 104: an attribute provides a bonus to contests that depend on it, equal to the (attribute) score minus ten then divided by two if the score is even. If odd, the bonus is equal to the score minus eleven then divided by two.
  • 104v2: an attribute provides a bonus to contests that depend on it, equal to +2 if the attribute score is the PC's highest attribute, or -2 if the attribute is the PC's lowest attribute score.
  • 104v3: an attribute bonus applies to a contest when a PC's d20 roll is below the applicable attribute score. If the d20 result is below the attribute score, the PC adds the PC's character level to the contest result.
  • 104v4 (attribute IS the bonus amount): a PC adds the attribute score to all contests depending on that attribute. (This version involves rewriting the Max Damage rule, 313, and the Attribute rule, 103.)
  • 104v5: an attribute provides a bonus to contests that depend on it, equal to the attribute score less ten.
The idea behind the attribute bonus is that having a strong attribute (for example, Mental) makes a character inherently better at mental tasks, but training (skill) has a direct and greater impact on specific tasks (like Chemist skill used to create acids or ballistic foam). Version 5 is included above as a more harmonious example than the kludgy rule, but it would be difficult to implement because its effects would cascade into several other rules and intentions like the value of skill points vs. attribute points, the awards for level-ups, the relationship of die result compared to bonuses, and the manageability of maths at higher character levels. In other words, I don't like it 🤓
An option, if you wish to preserve the traditional approach of ATT=floor(Attribute-10)/2, could be to let odd attributes have value in some corner cases. Some potential examples:
  • Ties favor the person with the higher raw score, rather than the person with the higher modifier
  • "Passive" checks increase if your stat is odd, perhaps even being equal to your raw score rather than being ATT+10
  • Making odd stat values unlock access to things (e.g. multiclassing at 13+, certain strong feats at 15+, skill "masteries" or the like at 17+, etc.)
  • Giving each odd stat value a clear, but small, scaling perk of some kind, e.g. 11 Str lets you jump +5 feet forward, 13 Str lets you jump +5 feet upward, 15 Str you ignore the weight of armor/weapons you're actually wearing (for carrying capacity), 17 Str you gain either a climb or swim speed (not both) of 10 feet or increase your existing one by +5 feet, 19 you have advantage on saves to break free of physical restraints.
More or less, if you want to keep the existing system, you'll need to expand to make odd scores actually valuable. Otherwise, you're stuck with the unenviable task of rewriting large portions of the rules to address the more elegant, but either chunkier (modifiers don't change, raw stats are eliminated) or larger (mods are eliminated, raw stats are used directly) numbers.
 

Related Articles

Remove ads

Remove ads

Top