Creature Feature: Earless Hoichi

Unlike previous Creature Features, Earless Hoichi is not actually a creature. In fact, he’s the unfortunate victim of a long-dead military clan. The story of Earless Hoichi (耳なし芳ー, literally “Hoichi without ears”) is a classic Japanese ghost story, and is required reading for anybody interested in understanding Japanese horror.

You can read the story of Earless Hoichi for free online, or you could buy a copy of Lafcadio Hearn’s fantastic Kwaidan. I’ll present a short summary of the tale here, but I highly recommend that you read the original.

Hoichi is a player of the biwa, a kind of Japanese lute. He is a fantastic player, one of the best in the country, but unknown and poor. Hoichi is also unfortunate enough to be blind. Having no home, he seeks refuge in a temple (Amidaji, which still exists today as Akama Shrine), where a friend of his resides as a priest. One night while sleeping at the temple, Hoichi is approached by a gruff samurai who insists that he play for the samurai’s master. Even though it is the middle of the night, Hoichi follows the samurai to a secluded area that seems to be filled with people talking in hushed voices. He is asked to play a rendition of the Tale of the Heike, a story at which he is particularly skilled. His performance is so astonishing that his audience is moved to tears. On the way back to the temple the samurai tells Hoichi not to speak of the event, as his master is traveling incognito.

Hoichi is visited by the samurai the next night, and the night after that. The priest of Amidaji notices Hoichi’s absence, and has some of his servants follow the lute player the following evening. What they see astounds them; Hoichi appears to stand up in the middle of the night and walk briskly out of the temple. His speed is unbelievable for a blind man, and eventually they find him playing his lute furiously in the middle of a deserted graveyard. When they drag him back to the temple, Hoichi relates the story of the samurai and his secret master to the priest.

Realizing Hoichi has been bewitched by ghosts, the priest vows to save his friend from further torment. He paints Hoichi’s body with the text of a holy sutra to protect him from the ghostly samurai. The priest also instructs Hoichi to remain motionless at all times; any sound or movement, and the spell will be broken. That night the samurai calls for Hoichi as usual, but Hoichi is nowhere to be found. In his place, the samurai finds only a pair of floating ears. Though the sutra has rendered Hoichi’s body invisible to the ghost, the priest has neglected to paint


A statue of Hoichi at Akama Shrine

the characters on Hoichi’s ears, making them visible. The samurai, attempting to comply with his order to retain Hoichi to the furthest extent possible, rips Hoichi’s ears off and leaves with them. Though he is scared out of his wits and bleeding profusely, Hoichi does not dare move or make any sound until the priest finds him some time later.

Though Hoichi loses both of his ears, the priest’s spell is effective at freeing him from the clutches of the dead clan. He is thereafter never bothered by ghosts again, and goes on to become a famous lute player.

The story of Earless Hoichi is important because it is widely known. It is also interesting because embedded within it is another widely known story, the Tale of the Heike, which is based on real events. To understand the significance of some of the details of this story, it’s good to have a little background.

In the 12th century, the rival Minamoto and Taira clans had an extended conflict called the Genpei War. The final battle in that war is called The Battle of Dan-no-ura, which ended in victory for the Minamoto clan, giving them control of Japan. Dan-no-ura was a navel battle, and when the Taira elite realized that victory was impossible, most of them (including the six year old Emperor Antoku) chose to jump into the sea rather than face defeat. The word Heike means “Taira clan” (“hei” is the alternate reading for “taira”), and the Tale of the Heike chronicles the conflict and eventual demise of the Taira.

The decisive Battle of Dan-no-ura took place in the straits of Shimonoseki, which is where Amidaji is located. During the Meiji Restoration shinto became the state religion, and at that time the Buddhist temple of Amidaji was converted into a shinto shrine. The temple was probably originally constructed in memory of the child Emperor Antoku and the others of his clan who drowned themselves; today it also pays tribute to Hoichi.

Considering the location of Hoichi’s story and the specific reference to the Tale of Heike, I think it is safe to assume that Hoichi’s ghostly assailants are indeed members of the dead Taira clan. Thus their interest in Hoichi’s ability to recount the sad story of their demise is easily understood. This story is a classic Japanese ghost story in many respects, but I think the most telling theme is that of intense sadness: the Taira who died in the Battle of Dan-no-ura feel so strongly about their end that their emotion is able to reach out from beyond the grave and interfere with the real world, even though the Minamoto, their enemies in life, are long dead.

Holy Bandwidth, Batman

It seems that I’ve been discovered by reddit.com. Hi there, reddit folks, thanks for stopping by! You might also be interested in my guide to Japanese horror, my article on the prehistory of the survival horror genre, the games I’ve reviewed, or maybe my assorted blog posts about game design. Apologies if the site seems a little sluggish–there are a great many of you at the moment.

Update: Well, the reddit,com traffic has died off and been replaced with many visitors from spiegel.de. Hi there, German friends! The article from GEE is now being hosted at Spiegel, and if you are like me and don’t speak any German, you can read it (slightly mangled) via machine translation.

Hey, I’m Famous

If you read German and have access to magazines from that country, the latest issue of GEE contains an article on horror games with references to yours truly. I spoke with the author at GEE while in Japan last month, and though I’ve not actually seen the article yet, Michael from frightening tells me that it is now on sale. I wish that I had some German ability so that I can read the article (it sounds fascinating), but perhaps if it is posted online in the future I can read it with the aid of machine translation.

Anyway, check it out if you can!

Ratings Changes


Somebody set up us the skull

Ever since I started thinking about what game scores really mean, I’ve been bothered by the warped scale that the game review industry seems to use. If you follow sites like metacritic.com or gamerankings.com, you have probably noticed that most games fall into a very small range of scores. There’s no difference, for example, between a score of 30% and a score of 40%; both mean that the game is terrible and you shouldn’t buy it. On the other hand, there is a huge difference between 70% and 80%; one means that the game is so-so and the other means that it’s very good. Even the delta between 79% and 80% is enormous. And finally, ratings over 90% are just fodder for angry fanboys; is a 96% game really perceptibly better than a 95% game?

In short, the ratings curve for video games is extremely warped. Interestingly, aggregate review systems for other types of media don’t seem to have this problem; a review of 50% on rottentomatoes.com means what you would expect: that the film is throughly average.

Anyway, I had been thinking of changing my scoring system on this site to better reflect what I think reviewers mean when they select scores. I wanted to get away from actual numbers (too granular) and compress the range of possible scores down into a smaller set. While I was thinking about how best to do this, 1up.com switched from using numbers to letter grades. Letter grades seem like a pretty good system, but I’ve decided to go for something even less granular.

As of now, all games on this site are rated with a system that I ripped off from Leonard Maltin: 4 stars or a skull. I’ve detailed the individual ratings in the FAQ, but generally, four stars means “awesome,” one star means “just ok,” and a skull means the game is terrible. I think that this system will better communicate the value of individual games without getting too specific about how many percentage points better one game is vs another. And, in following with 1Up’s policy, the internal table for converting between my star rating and the warped scale used by the rest of the world will not be published.

I’ve changed the Horror Games and Reviews pages (and some others) to reflect the rating system changes. Check it out and let me know what you think.

Fatal Frame II and III Less Impossible To Locate

Destructoid is reporting that Fatal Frame 2 and Fatal Frame 3 have been (finally) issued reprints. These games, especially the third one, have been notoriously hard to find. But apparently if you act now you can find copies of both games fairly easily.

Destructoid also has an interesting note about the popularity of Resident Evil on the Wii. Astute readers might recall that I expected the Wii to be the place for horror games, and it appears that customers, at least for the moment and at least in regards to Resident Evil, agree.

Flower, Sun, and Rain


Shot from the DS version

In Japan I picked up a copy of Flower, Sun, and Rain (花と太陽と雨), an early game by Suda51, the brains behind Killer7, No More Heros, and, unfortunately, Michigan. Like Killer7 and No More Heros, the game uses a flat, cell-shaded style to tell a crazy story. Fans of Suda51 games will notice his fingerprints all over this game: everything from the excellent style to the creepy voices to masked pro-wrestlers are present here in protozoic form.

The game plays a lot like Hotel Dusk, another game that I enjoyed that mixes stylized art with adventure mechanics. The protagonist, Sumio Mundo, is a professional searcher; people employ him to find things that don’t want to be found. He’s visiting the tropical island of Lospass (“Lost Past”–the game is full of oddly contorted English) for work; the owner of the local resort hotel wants him to track down a bomb that terrorists have hidden at the airport. Before he can even begin his mission, however, Sumio becomes distracted by events within the hotel and the plane explodes. The next day he wakes to find that time has reset; though he doesn’t realize it right away, he’s been given another chance to stop the plane from exploding. And so the game consists of a series of days, each occurring on the same date and each ending with the destruction of a plane. Each day Sumio is faced with a new puzzle, from the fantastic (“my room was rearranged by ghosts!”) to the mundane (“how can I get passed this guy who is blocking my way?”). Each puzzle is eventually solved by entering numbers into Katherine, Sumio’s rotary computer-in-a-briefcase. Jack Katherine into some object, enter the correct sequence of numbers, and proceed to the next day.

Like Michigan (and maybe No More Heros–I haven’t played it), Flower, Sun, and Rain is sort of nonsensical and funny. Or rather, it tries to be funny. As a non-Japanese who can just barely parse the excessive text in this game, the humor seems out of place to me. And the game is tough to play, not because the mechanics are hard (you don’t actually do much besides walk around and talk to people), but because the puzzles involve reading a huge amount of Japanese text. This game is just slightly beyond the level at which I can read in Japanese, so while I’m able to get enough to keep playing, I don’t always have much clue as to what the heck is going on. Of course, this is a Suda51 game, so confusion may be intentional rather than the result of my failed reading ability.

Flower, Sun, and Rain is headed to the DS sometime soon, which is a good thing. Not only is the laid-back adventure style a good fit for the DS (games like Hotel Dusk and Trace Memory have already proved this market), but it means that somebody will have to translate the damn thing into English. And though some of the sites on the net are complaining about the reduction in graphical quality for the DS version, I don’t think any of them have actually played the PS2 version; while the art style is excellent, the actual graphics look incredibly dated.

Flower, Sun, and Rain isn’t a horror game (at least, I don’t think it is–I’m only a few hours into it), but it’s a pretty interesting game from a pretty interesting designer who tends to make games that are hard to put into concise categories. I’m looking forward to playing this game on the DS when it comes out–even if it doesn’t make any sense, at least I’ll know that it’s not just me.

Props and Respect to my Homies

Just a quick shout out to some awesome game development blogs on the net. Last night my referer log lit up with links from Bruce on Games and Japanmanship. I’m going to follow Japanmanship’s lead and keep this blog roll rolling.

Japanmanship additions:

My additions:

  • Greggman. Game design, programming, and thoughts about Japan from an industry veteran.
  • Insert Credit. Obscure games, crazy indy games, hentai games, something for everyone.
  • UK Resistance. Hilarious critiques of gamers and the industry.
  • Three Hundred Mechanics. Idea prototyping with finesse and style.
  • Neko Games. A huge number of mini-games and handling experiments, all expertly done.

Why Racing Games Have Checkpoints


39 seconds to the next checkpoint.

Most racing games have checkpoints. The typical implementation gives the player a time limit to reach the next checkpoint, and adds time to the clock when the checkpoint is passed. If time runs out, the race ends prematurely and the game is over.

Ever stop to wonder why racing games have this checkpoint system? I mean, the game is usually about racing; there’s already a well-understood way to win (come in first) and way to fail (come in forth or worse). Even more perplexing is that the checkpoints in a racing game very quickly become irrelevant; as soon as the player gets good enough to hit checkpoints consistently, he can accumulate enough seconds on the clock that there’s very little chance he’ll run out of time. Real car races don’t feature checkpoints or time limits, so where did the idea for checkpoints come from and why are they so prevalent?

A historian will probably tell you that the checkpoint mechanic is a hold-over from arcade game design, which was often focused around getting the player to lose so that they would insert more quarters into the machine. And while that may be true, I think there’s a better reason that checkpoints in racing games have survived this long: they provide a short-term goal mechanism.

Consider an absolute beginner who plays a racer for the first time. He doesn’t know how to drift yet, so the turns are too hard. He keeps running into the walls or driving off the track. Anybody who has played pretty much any racing game has had this experience; at first the mechanics are unwieldy and it takes several races to begin to get into the groove. For this kind of player, coming in first is initially impossible. Unless the AI really significantly cheat to help the player, one crash into a wall is probably enough to ensure that the player is going to finish last in the race. But it’s going to take an hour or two for the player to get better, and finishing last over and over again can really be a blow to a new player’s ego. So instead, the racer gives the newbie player a short-term goal: don’t worry about finishing the race, just try to get to the next checkpoint before time runs out. This mechanism has several benefits: it makes the other cars on the track irrelevant, and pits the player against a fixed and predictable challenge. It also provides a way for the game to end early if the player is really sucking; prolonging failure is never a good idea, and if the player has no chance of winning it’s best just to end the game early. Finally, it gives the designers more control over the difficulty of a single track. Once the player can hit checkpoints consistently, he then has to deal with the other cars; checkpoints are a prerequisite to actually participating in the race. That means you can get a lot of play value out of a single track, even if the other cars always drive the same way.

The genius of the checkpoint system in racing games is that it is self-deprecating. I don’t mean that it makes jokes about how fat it is, I mean that it’s a mechanic that automatically becomes irrelevant with time. As the player gets better and is able to hit each checkpoint consistently, he accumulates time so that future checkpoints are easier. Eventually the checkpoints themselves have no effect on the game; the player hits every one of them, and they cease to be an important game mechanic. But by that point, they’ve already done their job: they’ve taught the player how to play well enough that he doesn’t need them any more.

So what’s the lesson here? The lesson is that games have to provide a level of difficulty that is appropriate to a wide range of player skills. The racing approach accommodates both newbie players and veterans by providing two separate challenges within the same game space. Players who know what they are doing can jump right in, but players who are new to the game still have a way to get enjoyment out of it.

So if you’ve read this whole thing and have been patiently waiting for me to tie this back in to horror games, this paragraph is for you. Horror games often (but not always) produce a similar effect by combining game play challenge with interesting content. The story, art, and characters in horror games are often the reason that the newbie player is willing to die over and over again without giving up on the game; instead of providing a type of challenge that is specifically geared towards inexperienced players, horror games seek to avoid frustration by being thematically interesting. Sometimes this works; Resident Evil has a very steep difficulty curve and obtuse controls, and yet the game is loved by all sorts of gamers because the story succeeds in hooking the player long enough for them to get good at the game mechanics. And that’s the ultimate goal: to keep the player from giving up on the game before he’s had a chance to really understand how to play.

Snow Over Shibuya

I just got back from a month-long trip to Japan. Though I lived in Japan a number of years ago when I was in school, and though my wife and I go back at least once a year, this was the first trip in about eight years that lasted more than two weeks. Normally I haven’t been able to take enough time off of work to really stay long enough to overcome jet lag, but the new job I started last year happens to have an office in Tokyo, which is extremely convenient. It snowed off and on last month, which was actually quite nice; living in California, it’s not often that I get to see snow, let alone trudge through it in the middle of a megacity.

While I was there I tried to find the Silent Hill arcade game, and while I was able to locate the machine, I didn’t get a chance to play because there was actually a line. Not a long line, mind you, just one sort of suspicious-looking middle-aged man, but the arcade was smokey and I couldn’t hear my friend over the Guitar Freaks racket, so I left. I was surprised to find anybody playing the thing, let alone people waiting to play it. I looked around for another machine but wasn’t able to find one. Oh well, maybe next time.

I also picked up THE Tairyou Jigoku, which is pretty much as terrible as it looks. Still, it’s probably quest-worthy for the same reasons that (equally terrible-looking) Escape from Bug Island is included. I also tracked down a copy of Dino Crisis for $1, which was a deal I could not pass up.

The month in Japan was a month without video games (except for the highly, highly excellent Phoenix Wright: Justice For All), so I’m behind on my Quest progress. But bear with me, we’ll return to regularly scheduled industry rants, game design ramblings, and random horror media reviews shortly.