And that's the last shovel pun I'll make here. The game is great. Critics agree. I put about 20 hours into the game and nabbed all the feats (achievements). As I greatly anticipate the extra content Yatch Club Games promised from their kickstarter, my overall feelings about Shovel Knight come with a hint of reservation. From the looks of things, the game will earn one of my Games Of The Year awards. However, I don't anticipate using it as a model of excellence. I think this feeling comes from how every layer of design in Shovel Knight (mechaincs, enemies, bosses, level design, difficulty, etc) is a bit under tuned and over designed. To be clear, 95% of Shovel Knight is solidly designed. Don't be discouraged by the rest of the 5% that I'll focus on here.
Many call Shovel Knight a platformer. I'm a bit of a stickler when it comes to language, gaming terms, and genre. Yes, you jump around to progress through this game. Yes, there are a lot of platforming challenges. But there's also a large focus on combat. I lean toward the genre "action side-scroller" because there are many Shovel Knight encounters where overcoming the enemies is more than a matter of jumpping over them as shovel strikes and projectile relics are common.
As far as mechanics go, there's MOVE, JUMP, STRIKE, POGO, and RELIC. Being heavily platformer based, JUMP is primary for maneuvering. In the air, players have a high degree of air control with little to no momentum like MegaMan. On the ground changing direction from left and right after moving at top speed causes a slight skidding animation like Mario.
I found holding down in mid air to activate the POGO mechanic a bit irksome for two reasons; 1) POGO is too easy to activate on accident especially with the Wii U gamepads sub par D-pad. I didn't expect this from a Nintendo controller, but unfortunately, holding the right arm of the D-pad and putting downward pressure on it will cause the input to be down+right. I've had the same problem with all of the D-pad directions. What this controller design affects in actual gameplay is the accidental activation of the POGO mechanic when intending to simply move left or right.
Traditionally, gameplay mechanics (actions/verbs) are isolated to individual buttons. If the D-pad only handled moving (like in Mario or MegaMan) the player would only move no matter how they manipulated the D-pad. And because the up and down directions typically don't do anything in these old school games (unless there are ladders to climb or a DUCK mechanic) players can freely swing their thumb around the D-pad smoothly transitioning between different vectors/variations of the same MOVE mechanic. It's not easy to transition left to right without pressing the up or down direction. The thumb swivels in a ball socket; it tends to move in arcs. With Shovel Knight, pressing down on the ground activates a mini duck (which I think is more animation than mechanic) and down in the air activates POGO. I'm not sure why all MOVE mechanics weren't kept to the D-pad and all shovel based mechanics were activated with the same "shovel" button. Pressing down + shovel button is the kind of intuitive Smash Bros like scheme that would be a good alternative. At least in Shovel Knight, pressing down once in the air activates the POGO state so players can then concentrate on horizontal aiming with MOVE. Not having to hold down to maintain the POGO took a lot of stress off of my hand.
Strike the earth. Shovel strikes are a primary way of interacting in the game. You unearth dirt, strike enemies, and destroy walls. While standing, press the attack button for a simple, short ranged shovel swing. Press while in the air for an air swing with the same range, timing, and power. It's here where our main characters most defining mechanic falls short. The level design (platforming paths and obstacles) are very much in the style of MegaMan. Many of the enemies are designed in the style of Zelda II enemies and Castlevania in that the take multiple hits, have multiple attacks, randomized AI, pathfinding, and they also block off your limited paths requiring combat action to progress safely.
The problem is MegaMan's M.buster is a projectile, a mechanic that engages player skill with various types of timing challenges with JUMP+SHOOT and attempting to hit moving targets from across the screen. Castlevania NES features WHIP as its primary combat mechanic. Though this whip has start up frames for the wind up, it's also very long and stuns or freezes enemies briefly. Compared to the core combat mechanics of these games Shovel Knight's STRIKE is an odd fit. It's slow. You must get very close to enemies due to the short range. Many enemies don't get stunned or knocked back much. And then there's the self knockback or recoil that happens when hitting some enemies. I like how STRIKE, a melee attack, has a sense of weight (recoil) and physicality (start up frames), but its design doesn't match the fidelity and style of the retro design of Shovel Knight. Compared to Castlevania, MegaMan, and even Mario, the enemies you fight in Shovel Knight are more complex, the environments are more dangerous, and successful attacks have more drawbacks.
STRIKE can't be spammed either like the M.buster making basic combat require more thought and planning. While I appreciate the slower pace to the combat, the timing gaps between swings leave the player vulnerable which can be frustrating to deal with when fighting quick and random/complex enemy behaviors. STRIKE is a pretty straightforward and somewhat dull mechanics. Imaging playing a brawler or a fighter with just a hook punch. There's not a lot of close-quarter-combat variety with just one move. Many old school games did a lot with just one move. But these moves were easier to see, use, and more effective on enemies.
Furthermore, NES games didn't do much with complex melee attack interactions because of the difficulty in calculating and conveying such actions. For NES games, touching enemies resulted in damage because these are simple interactions with clear feedback that the system can calculate without much trouble. You won't find many complex/variable interactions like reflections or knockback trajectories in Mario or MegaMan. Shovel Knight's STRIKE is attempts to take a relatively complex type of mechanic (melee) and make it work with level design designed around simpler, clearer, ranged interactions. The result is STRIKE creates awkward moments after hitting stronger enemies where their automatic contact based hitbox is turned off temporarily, the enemies continue moving, and then the hitbox turns back on. Sometimes the enemies continue to move into you despite being hit. Sometimes they are knocked backwards enough for you to recover from your STRIKE ending animation frames.
Because of the relatively slow STRIKE animation cycle before and after the swing, the recoil, the knockback to some enemies, STRIKE is one of the riskiest, least interesting mechanics in the game. For this reason POGO is much preferred as it provides protection from landing on most enemies and obstacles, it bounces upward at a large height equal to a full JUMP, the player can maintain horizontal air control, and the player can move horizontally while attacking. The ground STRIKE is not only slower and riskier, but the player cannot move and attack. While I generally prefer stop-and-pop style mechanics, it's strange when the STRIKE start up animation can be canceled by a jump resulting in no attack coming out. After all, the air STRIKE and ground STRIKE are identical. Also, if you land on the ground from an AIR strike the animation is not canceled. For such simple mechanics, issues like this are a bit irksome. But the most disappointment comes from when pressing the JUMP button doesn't cause Shovel Knight to jump. 1-out-of-100 JUMPS or so produces this error. Aside from the charge STRIKE mechanic, every grounded non-hurt state can be JUMP canceled out of.
RELIC rounds out Shovel Knight's core abilities. Some are designed for offense (war horn). Some defence (phase locket). Other's movement (gear). Most have multiple functions. It's great that there's an option in the controls menu to change the activation of RELIC from up+action to its own individual button. The benefits of quality controls/mechanics design is no mystery. When one button always jumps and another button always activates relic and the final button always STRIKES, the potential for error from unintentional execution decreases. For simple games like Shovel Knight, mechanic/controller individuality is key.
The enemies in Shovel Knight are a mixed bag. Creativity and visual design wise, the enemies range from decent to great. It's a bit of a bummer to see so many rats, knights, slimes, and wizards of various colors. Fortunately, every enemy is designed with different patterns, effects, and AI. I enjoyed the familiar enemy types like the Medusa head like flying sword/dagger/missile things on Tinker Knights stage.
On a whole, the enemies in Shovel Knight are more complex and dynamic than classic NES games. They react to player position, move around the stage dynamically, and often have different choices they can make. For every simple beetle, cannon ball, dripping flame, there's a more complex slime (random jumps), propeller rat (tracks player position), and dark bird (drops attacks from above). Then there are the even more complex enemies like the colored knights who advance, retreat, jump, and block with an AI of their own.
In comparison, Mario and MegMan enemies are mostly set in place in a level. In the original Super Mario Bros, you couldn't even scroll the screen left so once an enemy is off screen it's gone forever. There were also pits or pipes that removed enemies or trapped them in small spaces respectively. MegaMan enemies tend to do little chasing. They either are stationary, patrol, or are design to fall off the screen in some way. Exceptions include the Lakitu, Bullet Bills, Angry Sun, Shadow MegaMan (MM9) and the big jumper-crushing robots.
The extra dynamism (mobility and reactive AI) for Shovel Knights enemies can create additional challenging scenarios. Often times enemies can attack in groups, crowd together, and follow you to later parts of a level. At the same time this dynamism can easily allow for too many enemies on screen that overlap creating challenges that are more difficult to read (feedback and clear conveyance) and difficult to overcome. At its worst, the enemy encounters are more cluttered than they are compelling. In these moments I'm reminded of the difference between the solid level design of Mario games versus the less tuned more extreme level mods create in level editors.
The same trend of more complex enemies applies for Shovel Knights bosses. Mario bosses are the simplest, with 2 patterns and very little tracking or reactive AI (Bowser from SMB, Boom-Boom from SMB3). MegaMan bosses tend to have a few moves that it will cycle between or randomly pick from. This design is coupled with fairly predictable/simple movement patterns (because moving is also attacking in these games). With so few projectiles and relatively slow movement, these bosses provide a challenge without being so overwhelming as to limit the ability for the player to learn in the heat of combat and make adjustments. At the same time, the boss attacks are easy to understand with counters that are easy to see. The spaces and opportunities to dodge are large pockets and lanes. These bosses are designed to be solvable which makes the mastery process smoother as there are less cheap/unfair encounters. It also helps in Mega Man that you can do damage to these bosses from long range while you scout information.
The Shovel Knight bosses are designed more like fighting game characters/AI than the retro games explained above. Shovel Knight bosses have more options, faster movement, more layered attacks, and more phases. Their complexity is even beyond the Mega Man 10 hard boss design which gives each boss one additional move compared to the normal difficulty. Some have praised the Shovel Knight bosses for their lack of patterns. In reality, many of the Shovel Knight bosses do have patterns just like MM bosses. The difference is they react to the player's position and other conditions and have so many different moves that their pattern based movement and attacks are less obvious. This is not to mention how the bosses get additional attacks when they drop below half health.
Just like with the enemy design, these new complexities clutter up the boss battle gameplay challenges. Some battle encounters feel like fighting Street Fighter AI opponents due to their randomness and move options from a given state. While fighting games are designed around a guessing game of sorts due to the prevalence double blind interplay encounters, action games like Mario and MegaMan shine due because of the opposite approach. Bosses in MegaMan 10 especially are designed so that with sufficient skills (Dexterity - Knowledge - Adaptation - Reflex - Timing) players can master all the Robot Master bosses without taking damage, without making guesses, and without using any weapon besides the default M.Buster (see my videos here). This design makes the boss battles very clear, focused, with no tricks. In other words, you won't need a special move that you might not have to escape the bosses traps. Just think on your toes, aim well, and play.
From my experience the Shovel Knight bosses simply are not designed with this level of tuning. And the result can be very frustrating. For example, chasing down the teleporting Plague Knight when all you have is a short ranged normal attack is like a bad Street Fighter match up. Even when you close in, you can't really combo Plague Knight or do a lot of damage. Trying to rely on the reflecting his small arced projectiles back at him is not only too difficult (trying to hit one very tiny fast moving ball out of the air with a slow, small ranged STRIKE while dodging the attacks) but even if you reflect the projectile it is likely that it won't hit the boss as the boss is frequently moving. If Plague Knight hasn't moved, it could be already launching another attack at you, not waiting around to give each of its attack phases any breathing room. Factor in how the terrain is constantly changing (being built and broken down) and the whole boss battle can quickly become a chaotic, slightly unfair, clutter of interactions.
With Polar Knight, after he turns the ground into spikes, his triple snow ball attack can be nearly impossible to see coming and impossible to avoid depending on the spikes around you and the boss proximity. Polar Knight can switch so quickly between defending his head from your POGO strickes, to walking forward, to throwing triple snow balls, to crashing iciles that you may have no time to maneuver and no time to properly defend yourself. Sometimes these attacks are on the screen at the same time eliminating what few safe spots you might have had.
The two cannon balls that create pits on Propeller Knight's boss fight have no unique tell. Even looking into the background to see four incoming cannon balls doesn't inform you that two will be crashing down on the stage leaving pits in their wake. If you stand in the wrong spots, you'll fall into the pit. The cannonballs drop so fast, that walking out of the way may be impossible to react to without foreknowledge. It's a similar story with Black Knight's meteor attack later in the game. I know learning new things from direct experiences carries a degree of trial-and-error, but these boss attacks are more cheap than necessary. These tuning issues (balancing attack patterns, movement speed, STRIKE properties, etc.) affect beginning players and advance players as they attempt to gain skill and master combat.
There are a ton of great level challenges/setpieces throughout every main level. The optional challenge levels are more straightforward (almost puzzle like) in that they force the player to use specific relics in specific ways. For the most part, the entire game does a great job of exposing the player to new dangerous/interactive elements in a relatively safe way before ramping up the difficulty and putting these elements together into more layered challenges.
I didn't particularly like the darkness sections in Spectre Knight's stage or the overlaid obscuring elements in Propeller Knight's stage. For Spectre Knight's stage, with so many Pits around, waiting for the lighting strikes to illuminate the safe paths was a bit too obscured. There didn't seem to be much of an option other than just playing it safe. There's something about the waiting that comes from trying to understand the how the platforms will move that I greatly dislike. If you can only see the platforms intermittingly because of the lightening strikes, and the platforms you must jump to are moving on their own timers, putting these together creates an even more complex timer that players have to wait even longer to fully understand. The "rail track" visual design indicating where the moving platforms will go helps a little, but the platforms continue to move in mid air past where the track ends. So you can't really use the tracks to indicate how far these platforms will move. Furthermore, each platform is different and must be learned individually. See video here.
For other gripes, some enemies have odd placement in levels so that they activate and start attacking when off screen, or they're placed near the screen edge. Some enemy attacks like the purple knight enemy, are thrown off the top of the screen before falling down toward the player. Some of the encounters are in very cramped hallway where it's difficult to jump on top of these enemies. The purple knight on Treasure Knight's stage comes to mind for two of these case (see video here). I'm all for challenges that force players to think a bit harder and be a bit more precise by limiting commonly used options. But, again, due to the retro way the entire game is designed focusing on close quarters melee combat also focuses in on the system's shortcomings. In the video linked above, the player takes damage trading with the purple knight after attempting to POGO of its raised shield. The player also attempts to hit the shield night in the head with a JUMP + STRIKE. IT doesn't work either as it registered a shield hit. These are the awkward and finicky interactions that are typically avoided by using Relics and being able to JUMP freely and use POGO.
There are secrets stashed everywhere. Breakable walls, secret areas off the top of the screen World 1-2 style, and non-solid walls are the main ways secrets are hidden in Shovel Knight. I appreciate that most can be found with a careful eye and some fairly mindless checking by STRIKING at walls. It's even better that most secrets aren't freebies, but additional challenge rooms.
For a final note on level design, I feel like the levels in the game are a bit too long. I'd like to take some time later to map out the level's pacing by looking at the obstacles and themes introduced and how the difficulty and compleixty builds. On my first two playthroughs, I felt that there were so many game ideas conveyed that the overall progression of the level was a bit of a grab bag.
Is Shovel Knight a hard game? The truth is, I can't tell. Discussing game design and difficulty is tricky. To get much value out of this pursuit, it's important to be able to talk about challenges objectively. It's one thing to think a game is hard or easy. It's another to step outside of your expeirence and see how the game might be for other players or to consider how your experience would be different if you explored other options. Getting to the bottom of Shovel Knight's difficulty will bring us closer to understanding why I think the Shovel Knight as a whole is over designed.
I first played through Shovel Knight on normal without cashing in any health upgrades or buying them. I bought all the armor, shovel upgrades, and relics I could. But I didn't use any ichors. This made the game a lot harder than it was probably intended to be. The game was hard because I could die so easily, not because the level and boss challenges were particularly insurmountable. In other words, I had less room to make mistakes in my learning process. The game was fun and quite doable especially after I took off the Magic armor that gave me double damage.
Afterwards, I jumped right into new game plus where I bought the rest of the health upgrades. So this time the biggest difference in my play experience was that there were fewer checkpoints in the game and less health pickups in the levels. When I was low on health I would use one of my two red ichors that I replenished regularly because refils are free. So aside from the added frustration of dying and playing more of the level than I thought, the game was about the same in difficulty until I got to the boss rush near the end. Perhaps much of it was easier since I knew the levels already, I had the relics from the beginning, and I knew how to abuse the phase locket (more on this later).
After new game plus I beat the game from a new file while earning 7+ of the game's hardest feats at once:
The hardest part about this final run through the game was beating it under an hour and a half. I was a goof and goofed around collecting gold in the first 3 levels of the game even though I knew I wasn't going to spend any money. I love doing a minimalist run through a game. M.buster only is one of my favorite ways to play MegaMan. But STRIKE as a melee attack isn't as engaging as the MegaMan projectile. Though I liked not even having the option to cheese my way through challenges with the Phase Locket, I felt that many of the encounters were a bit dull. Relics really do spice up the experience. I didn't realize how much they added to the movement in the game.
I still don't feel like I hit a difficulty sweet spot with Shovel Knight. Part of the reason why I feel this way is because there are so many variable difficulty elements. In fact, I argue that there are too many. Between the optional paths in the levels, using relics, buying health upgrades (which can be grinded for), buying armor, filling the chalices, magic upgrades, playing on new game plus, and going for feats there is a lot of player elected difficulty variation to experiment with. Ultimately, as is the nature of emergent, open, or non-linear games, it becomes harder to tune so many options to provide a focused gameplay experience.
While some degree of player elected variable difficulty is a great thing. Too many options can overwhelm the player. I heard people saying online that their playthrough of Shovel Knight as too easy because they had upgraded so much. So without the ability to make an informed decision based on my own preferences, I opted to play the game without getting health upgrades. All the while, I never knew if the game was unnecessarily (or uninterestingly) hard because I had low heath, because the game itself was hard for everyone, or because I wasn't using some other option/strategy. One of the best benefits of more linear, straightforward game design (less options) is that the player is more likely to be highly informed of the challenges at hand and better able to make decisions. Because of all the upgrading in Shovel Knight, like so many RPGs (read more about this tuning concept here), finding the sweet spot balance of difficulty is ever elusive for players.
The other drawback of providing so many options in a game like Shovel Knight is that the game becomes harder to balance so that all options provide rich, interesting experiences. Let's face it, aside from falling into pits, the entire game can be cheesed by abusing the Phase Locket. The cost of this relic is super cheap. Don't want to get hit off the platform? Phase it. Mistimed a jump? Phase over and over to hover in mid air! Don't want to battle this large enemy. Phase through it. Don't want to deal with Polar Knight and his crazy spike traps? Phase and STRIKE him over and over.
In general, I like the PHASE RELIC. I think the ability to avoid all damage is actually the only thing that can counter some of the cheap boss attack combinations giving the player an easy way to escape and learn the boss's moves. Don't know what the boss is about to do? Dont' feel like taking damage from its random, difficult to predict movement. Phase and learn. Perhaps the Phase Locket would be less abusable if it cost more magic points or if it didn't cancel into itself with reuse. It would be interesting if players had to let the phase timer completely run out, become vulnerable again, and then attempt to phase back.
You can buy your way into better weapons, armor, health, and more. You can grind to get more money. Good players tend to have a lot of money because they die less and therefore lose less from failed attempts to retrieve dropped money bags. Struggling players will tend to have less money to spend to make the game harder for them. Regardless, the best way to not lose money for either type of player is to spend it while you have it. For this reason, the infinite check point along with the money bag Dark Souls like retrieval system hinders more than it helps.
Though many consider having lives and gameovers a thing of the past, these design features are still highly effective. This feature is a double edged sword though. If a game is highly polished and well tuned with fair challenges and a progression of game idea presented through gameplay challenges, then giving players limited lives to make it through works the best as players can use their failures as clear feedback in the learning process and a measuring stick. If you make too many mistakes in MegaMan you have to go back, rethink your approach, and run through parts of the level again to refresh on concepts. Often times you'll develop skills need to overcome later level challenges by repeating prior challenges. In games like MegaMan and Mario, if you lose a powerup or use up some of your Robot Master weapon ammo, the consequences ripple forward onto your next lives. These suspended (carried over) elements give the player's efforts across lives a progression where repeated attempts have slightly different conditions.
In Shovel Knight, like most games with infinite and frequent check points, players are more prone to continually attempt the same troubled sections until they make it through. This is generally how games with infinite checkpoints are. Two other side effects of this design are an increase to the difficult of sections and and increase in length. Though the level challenges are fairly manageable, each stage is a bit long. And the bosses are a bit too hard in an unfair, cluttered way. Of course, once you master the game you can go through it all without dying. In such a case checkpoints are clearly not necessary. However, for the first time players, how frustrating would Shovel Knight be if they had to restart the whole level after a few deaths? With these drawbacks in mind, I wonder if the difficulty design works against less skilled and beginning players. Sure they can persevere their way through the game. But because of the inifite checkpoints, they're more likely to lose money and learn lessons the hard way while doing it.
I don't want you to leave thinking Shovel Knight isn't a great game. I still sing its praises to the tune of the main theme. And I still await for all the DLC extra goodies coming my way. Like I said above, all the issues in this article make up about 5% of the game's shortcomings. Many of the points are subtle, nitpicky things. But others point out design decisions in precise places in the game's design that cause a variety of different odd affects on the user's end experience. The more I try to dig into the gameplay and systems of Shovel Knight, the more these little issues come into focus. Overall, I step back and look at all the stuff in the game. So many upgrades. So many roaming challenges and bosses. So many secrets. So many nooks and crannies. So many options. And so many cheat codes; over 300 codes! I feel that these codes reflect the kind of design sensibility that's at the core of Shovel Knight; a lot of cool, a bit of excess, and the burden to organize and dig through it all lands on the player.