The Eldritch game, years in development, has now been published by Goodman Games at a comparable cost to a Dungeon Crawl Classics module of the same size. You should definitely download the Quick Start as a primer, and the FAQ to save some confusion over typos (both are at Goodman Games).
Eldritch is a 96-page softbound book, with a colour cover and B&W interior in a simple layout with no graphic excess. Some art (grayscale reproductions of colour and quite dark) is used for illustrations and as design elements. Type is small but legible in good light (more so than, say, the latest incarnation of Tekumel). There is not much contrast between heading levels, but there are clear page numbers and running chapter headings.
Eldritch has five chapters and an appendix, plus bits and bobs: character sheet, quick reference pages, contents, index, foreword, afterword, and a preview of the official campaign setting.
With a title like 'Eldritch' you may have expected something to do with Lovecraft, but no squamous tentacled things appear. Eldritch is a plain fantasy role-playing game with epic leanings. It is deliberately generic, even bland, expecting (and relying on) the GM to add his own descriptive flair to characterise the action and the campaign. The high level of abstraction should let the rules work with most settings, flavours, or styles of play. Many RPGs these days come like premixed drinks, with those factors set (say, Hastur and Holsters: the Wild West Mythos -- "Nyartholep's three eyes scanned the dirt streets of Laredo, tendrils twitching above his guns..."). Those expecting that sort of thing will be disappointed: you have to mix your own drinks.
The game is skills-based. Characters are created using a points system, with possible advantages (e.g. ambidexterity) and disadvantages (e.g. missing leg -- seriously... a nod to KODT?). Characters can be any of the common 1st edition ADnD races. Each race has a typical ability package. Beyond that, any abilities can be chosen. For those wanting occupational paths, there is a range of archetypical careers defined by ability combinations. Characters don't have separate attributes as such. Some physical capacities such as agility and endurance are abilities. Others such as intellect are left to role-playing. Characters gain 'levels' by earning victory and role-playing points, getting points to purchase new abilities and improve existing ones, a rank at a time.
This works pretty well. New races could be created by defining typical strengths via abilities. Character archetypes (an important fantasy element) are preserved without using classes as such. It looks as if character advancement in a campaign should work well. Levels regulate progress rather than being an absolute measure of power.
Action and Abilities
The game hinges on the ability list and action system. Combat and magic elaborate on the same base mechanics. Abilities are broadly defined, with comprehensive list of mundane and adventuring skills. Rather than use a percentile or attribute-related value to measure proficiency, the game uses a range of dice, D4 (human average) through D12 (superb). Opposing rolls resolve challenges, the margin of success determining the exact results (GM describes). For static challenges such as climbing, the GM assigns a difficulty that determines which dice to roll. Basic abilities, except for a few that require special training, are free to all at D4 'die rank'. Very useful, this. You can improve depth as well as rank in an ability, gaining specialisation then mastery, each with its own ranks. This lets you roll up to 3 dice (base specialisation mastery) when using the ability. For example, a master archer might have Ranged (D4) > Bow (D8) > Long Bow (D6). He would roll D4 D8 D6 if using a long bow, D4 D8 for a short bow, and D4 if chucking a rock.
I really like the way this system works, although at first it seemed a weird way to do things (maybe that's where 'Eldritch' came from). Anyone has a chance to try anything, but the chance of success remains proportional to ability. It's possible, if unlikely, for a tyro to get the drop on a master. The dice seem to add just the right amount of risk.
In combat, an attacker rolls dice to determine potential harm, e.g. D4 D8 D6 for the master archer above. A defender chooses an active defense (e.g. evade, dodge, deflect) and any points getting through apply to passive defenses: armour (die roll), then resilience (against magic) or toughness (bodily harm). A flowchart shows how it works. In any continuous action, the active 'defense pools' are temporarily depleted until the target is worn down and taking real damage. All 'defense pools' derive from abilities, so boxers can evade blows well, and users of magic can better resist it. Characters can improve such abilities to get sturdier as they gain levels. Initiative is involved, with special rules and oddments for various things. Add to that multiple attacks and special combat manoeuvres, and combat is not rules-light, although adjustments are few, so calculators aren't needed either.
Combat plays quite quickly, and with practice should be very quick -- it seems as cinematic as claimed. I liked the die roll for armour and the way unarmed attacks fit sensibly and seamlessly into a fight. Players liked being active in both attack and defense, and immediately grasped the two-stage defense, parrying or deflecting as if born to it.
There is one magical ability, Arcanum, with specialisations (allowing spellcasting) in four power sources, and masteries as priest, wizard, psychic, and so on. Magic uses a spell points system, points deriving from the casterís ability. Spells are expensive to cast, but cantrips (minor effects such as detecting active magic or recharging items) can be cast for minimal cost. Spells other than Harm spells require a check. Casters can learn a certain number of spells. There are lists of example spells, and full rules for creating spells, and magic items. Spells are defined in a standard way, and have standard points costs, casting times, durations (usually instantaneous or maintainable by paying spell points), types of effect, maximum ranges, and so on. Summoning spells are complex to create, since they define the monster summoned. Items work on spell points, and permanent ones can draw on life energy.
I expect most players will take a while to get to grips with inventing spells, but magic looks promising. It integrates well with the rest of the system, and covers most things my group could think of offhand. Ideas for mean trap and item effects started to occur quite quickly.
Monsters and More
There are a couple of pages of weapon statistics: a decent basic range, rather than 57 varieties of polearm. There are no equipment lists as such (though some folk are doubtless tired of seeing them). There are no list of monsters either, although there are a couple of examples in the text, so the GM must create his own. Opponents are defined as one of four types: fodder, standard, exceptional, and full-fledged. Heroes are able to mow down fodder (e.g. goblins) like ripe wheat -- this is where the epic feel comes in. (There are suggestions to make character life tougher if you like.) To save bookkeeping, the first three types have simple statistics, with defenses rolled into one hit points score, and so on. Most legendary creatures are exceptional, but full-fledged opponents (e.g. dragon kings) are like characters. There is a system for working out the challenge monsters pose.
Simpler weaponry is good (most variations in other systems are minor compared to dice variability in combat anyway). I don't see why katana has a separate entry, though: at the level of abstraction in the game, it's just a long sword used two-handed. Creating monsters or adapting them from other systems should be easy for experienced GMs. There is not much game-mechanical contrast between monsters (or maybe this is just more obvious), so maintaining opponent variety over a campaign would depend mainly on the GM's descriptive skills.
Character Sheet. The Eldritch Character Sheet is seven pages long, so sheet is a misnomer. It is usable photo-reduced. It covers all abilities, spells, etc., and helps keep things clear in play, but a shorter version would be as useful until late in a campaign.
Ainereve. The taster for the official setting, Ainereve, is intriguing -- a dreamlike world with settled areas shaped by belief and other areas in flux, reminiscent of Jack Vance's tale Parapsyche. The text used examples of real-world religions, which I'd be reluctant to use in a game (and not just because GURPS Fantasy 2 was so awful).
Extras. Foreword, afterword, and side notes explain design decisions and philosophy. The simple basic mechanics are meant for story-oriented play, you are encouraged to add skill successions and otherwise tailor the system to your preferences, and so on. This material might have been better in one place. It is hard to find things buried in the text, so it is a pity the index is undifferentiated (no sub-entries or prioritisation) and incomplete (e.g. no entry for falls). The Contents is better. The Quick Reference is very handy, although it makes you wonder about differences in terminology between the tables.
I found the rules a tough read the first time through, and while some things such as the action system made perfect sense, others seemed impenetrably complex. Was it my aging brain? (Probably.) I then tried the Quick Start, which I could follow more easily, and ran a playtest of the sample adventure, which played very smoothly. The true test is in the polyhedra. Back to the book, and I understood it this time.
Some things prevented easy assimilation, I think. The writing was okay except for random pronouns, but the weird terminology (maybe that's where 'Eldritch' really came from), and perhaps some overexplanation, made the rules confusing for me. To avoid the unlovely "to hit roll", the authors came up with the worse "potential harm" and "threatpoints" -- the simpler "attack" or "threat" would have done, surely. I daresay I'll get used to the terms -- I can already say 'Ability-Dice-Chain' without grimacing. The rules also often try to run before walking. Complex material is introduced before it is needed, interrupting the smooth and ordered flow of information. For instance, the 'extra weapon attack' advantage in the first chapter details much about melee that might have been better left to the combat chapter.
As an aside, I believe a few changes in presentation might help the uninitiated. Advanced material could be marked in a way that allowed a graded approach to the game. The example spells in the appendix help those new to creating spells by supplying examples of how it's done, and lets games use magic straight away. If appendices of sample monsters and magic items were supplied on the same basis, it would help open the game up to the less experienced and those with less preparation time.
As is typical for many gaming products, there are typos. (Eldritch isn't that bad. My own Book of Names had so many mistakes -- lost formatting etc. -- that I couldn't open it without swearing for over two years.) Most errors are not too confusing: calvary for cavalry (a subtle pun on cross?), lead for led, aware for award, mourning star for morning star (a sad blow, that). Some text is misplaced, e.g. entries for p.23 in the contents, the title on p.41. Some pictures are stretched, e.g. the anorexic elf on p.81, sampled at a more normal aspect ratio on p.2. There are signs of rough revision: cross-references left at "see page xx"; mentions of EHF (Eldritch High Fantasy) and Falling DP from an earlier draft of the game. Some errors are confusing: a partial change of skill names from Intimidation to Coercion and (I think) Perform to Artistry; omitted stats for missile weapon ranges and for using a longsword one-handed (not good for the most popular weapon choice). As is true for many game systems, some worked examples have numerical errors, which is a bit of a lead lifebelt for the perplexed.
None of this is fatal, though, or even much of a problem. You can find some fixes in the FAQ download and the discussion forum at Goodman Games. Missile weapon ranges are not available as I write this, but ranges from other systems can be used with no trouble.
The way Eldritch works in practice shines through the cthulhoid cloud of quibbles above. The game shows the fruits of extended playtesting, in that actual play seems intuitive within minutes, and all the parts work together well (giving concinnity in action if not explanation). The die ranges really appeal. I was impressed by the way the more abstract approach and unified system simply made a lot of potential problems vanish. It just works. Two pages rather than twenty cover abilities perfectly well. No need for detailed class or race definitions. Weaponry and armour are straightforward with few fiddly bits. Combat fatigue arises naturally. Styles of spells and unarmed combat are easily outlined and play simply (no cries of "No! Not the grappling tables!"). I didn't have to throw my "rules lawyer stopper" D30 once.
The abstraction means the game relies on GMs adding vivid descriptions to round things out, improvising, and using common sense rather than rules to judge attempts to parry maces with penknives and so on. This is what good GMs have always done, though. This is not just a story-telling approach; it makes for better action too.
Eldritch's rules are not easy to digest at first, but make sense if you make the effort, and prove coherent in the end. The game is fun and not at all complex to play. This is far preferable to a game that reads well and plays badly! Overall, it is a good, flexible system with a lot to like, open to many gaming styles. You'll need your own campaign setting, especially if you want to focus less on action than interaction. I recommend it if you're an experienced GM looking for something a bit different.