The Making Of Thief: Deadly Shadows

[This time we turn our attention to the development of the third Thief game. It’s worth noting this is the first making of where the person I interviewed wasn’t the effective Project Lead. This leads to a very different interview. I’m speaking to Jordan Thomas, who’s got a way with a quote. I’ve interviewed Jordan a few times before: here’s him on the Cradle and here’s him on lighting in Bioshock. EDIT: When I was putting the article online, I somehow snipped a whole paragraph and a half when formatting it. It was the bit after the word “Academic”, and actually one of the key sections of the whole interview. Excuse? Er… I was deeply hungover. Will that do?]

When Looking Glass shattered, your correspondent, along with the vast majority of Thief’s sizable, fanatic fanbase, got more than a little despondent. Was there any hope for a continuation of the greatest stealth game the world had ever seen? Well, yes, there was, as otherwise we wouldn’t be doing a post-mortem of Thief III and instead continuing to weep hot tears into our foaming mead. The game arrived in the hands of Ion Storm Austin, fresh from their success in making the original Deus Ex. With a new team, mixing veterans of Looking Glass and new staff, they faced the challenge of matching their forefathers.

In many ways they did. Thief: Deadly Shadows isn’t a perfect game – but then neither were any of the Thief games. But the journey from concept to completion was clearly a difficult one, with some noticeable scars on the final game. Jordan Thomas was second to last man into the Design Team. He was there for the long journey from November 2001 until the game actually shipped, ending up as Lead Designer. What were things like when he arrived? “It was probably pre-alpha,” he recalls, “Very rough. You could move Garrett around. He had some of his basic abilities, like you could mantle on things. Technology was nowhere near where it needed to be. We had a guy who was working on a real-time lighting renderer. And that’s Herculean. The fact is, we got it out before Doom, and we were trying to take on some of the same issues: Normal Mapping.”

“Things were going pretty well at that point,” he continues, “We were all really idealistic at that point, even though they’d been through some level building hardships. Because Randy [Smith, Project Lead – Ed] was a phenomenal design lead, and at that moment he was acting like a Design Lead in every sense of the word. We learned so much from him, very quickly. When things started to solidify, with a given space starting to feel great and the lighting technology came online, our enthusiasm was definitely at an all time high. But it was a rough project, and we had a lot of trials ahead from when I was hired.”

Many of the difficulties centred around technology. “If you’re going to be a middle-ware company – if you’re going to be a content company, and say your strength is to come up with game systems and content that are groundbreaking… don’t also attempt to be a technology company,” Jordan argues, “We decided to write our own renderer. The company had no experience with that. They’d barely gotten the Unreal Tournament technology to be subject to their whims for Deus Ex… and when that renderer finally did come online, it was somewhat nightmarish. Not the fault of the guy who wrote it, just that it would have needed another year of hashing out issues before it was really ready. If there was one thing that was the Achilles heel of both Invisible War and Deadly Shadows it was the technology. It was hard for us to use, and we could never really get the performance up to what it needed to be for retail. And we fought it the whole way there.”

In terms of other problems, it’s perhaps telling that there was a time that Ion Storm were discussing a change to the name “Manifesto Games”. “There were some radical departures from traditional structured business at Ion Storm, which often lead us in a direction which people like to call “Organic” but really meant that mostly you were… well, self-congratulating while the game wasn’t making as much progress as it needed to. It was so academic. Amazing as the design culture there was, it sometimes felt more like a school.”

In other words, a over-emphasis of theory over practice. ”To the credit of the guys who were the progenitors, the philosophy of design that you would learn there was very sound, and you could apply it all the time,” Jordan elaborates, “But I often got the feeling that we spent so much time discussing a given… well, for example, we’d say “This is exactly how the game should work. This is how the core game should flow”. And we’d discuss it at great length, take down notes and finally implement… but at that time there wasn’t enough time really to iterate upon it. A much more sound process, in an Art-for-hire industry, is to prototype earlier. Make something you can get your hands on, and then iterate rapidly on that and allow to evolve into your game, rather than believe you can really put your game down on paper. I don’t believe that. I believe the best games are the ones that they could play and tune as early as pre-alpha”

Not that their formalist rigour didn’t mean they were pushing consciously in a progressive direction, much of which is evident in the game’s many high points. “On some very basic level, we were trying to do the right thing with game design,” Jordan notes, “We were pushing the right boundaries. The notion of simulation and free-play spaces, with prone to emergent, systemic design… that absolutely is the wave of the future. People are going to start demanding that the world behave as if it’s a collection of inter-reflection behaviours, rather than “Hey – look! These are our deterministic rules for how things works. If you don’t use the chainsaw on the desk, which we’ve told to respond to the desk, it’ll never, ever function”. But at this stage, we were sort of on the frontier and kind of stumbled. But we paved the way.”

“If you want to look at a good example – and the Ubi guys will admit this freely – the Splinter Cell series owes a lot to Looking Glass and Ion Storm games,” Jordan adds, “That’s a great example of what you can do if you take that same sort of design philosophy, at its core, and apply it to a much better positioned franchise with a more mass-market core fantasy”. Which does make you wonder how they dealt with the obvious difficulties and compromises of co-development, with Deadly Shadows appearing simultaneously on PC and Xbox.

“As PC intelligentsia traditional systemic-design nerds, which is what we were, we found that the technical restraints of a console were relatively challenging,” Jordan answers, “On top of that, there was “Hey – how easy does this game have to be? At what level are we patronising?” For example, console gamers in general prefer a little more clear direction while PC gamers tend to prefer to be thrown into a playground and told “Hey – Go! Find the other side”. We constantly tried to find that sweet spot. I feel that Deadly Shadows is mostly mainly a PC game on the console… though you can’t truly please the hardcore PC fans with any form of compromise. From the very beginning, we tried hard to strike a balance so the more moderate PC fan and the more moderate console fan – the console player who doesn’t need crazy amounts of direction – would have a good time playing the game. It was definitely challenging. I think all of them have become better console developers having come through it, but it was a long road”.

Any particular lessons which others would be wise to steal from the Thief designers? “Make intelligent use of what technology you choose to undertake and prototype your gameplay experience very early,” Jordan argues, “Don’t try and front load the process with too much theory, essentially, and get it into practice. Make it feel good in your hands, because that’s what we’re doing. We’re making interactive experience, not narrative.”

Note that despite the lacklustre sales of Deadly Shadows, it isn’t necessarily the last we’ll ever see of Garrett. Maybe it just needs to be tailored to the needs of its more limited audience. ”I believe honestly that, in the long term, Thief will be better off as episodic content,” Jordan theorises, “Smaller release portions for a smaller, but very regular, audience. I’d love to see the series continue in that vein, but that’s very much up to the people who now own it”. This actually puts Thief:DS in the tradition of the Thief games. Thief, while not being the sales disaster as it’s occasionally painted, certainly never reached the mass audience, focusing on a real a real cult appeal. “They had exotic interests, and were looking for something new in terms of fiction as well as play experience,” Jordan describes Thief’s players, “Thief absolutely made a huge splash among that audience… but never got particularly larger than that. The console people, the guys who buy Splinter Cell, aren’t really interested in playing a Thief in a Medieval city. And that’s sad.”


  1. Willy359 says:

    Where do I sign up for a lifetime subscription to Thief: Episodes? Especially if they’re going to go back to catering to the old-school Thief fans, of whom I am one. I enjoyed Deadly Shadows, and the Cradle is an undisputed masterpiece, but what I really want is a few more levels like Life of the Party.

  2. The Sombrero Kid says:

    extremely unfair on a brilliant game, but I can see where you’re coming from, not saying there weren’t mistakes, but not as much as Deus Ex 2 imo & worth playing for the cradle alone.

  3. MisterBritish says:

    DS was definately a great game, but there was always the feeling it could have been much better if they hadn’t tried to straddle the PC/Console design fence.

    They also lost points for having a value in the config for ‘MouseLag’.

    And yeah, ‘The Cradle’ is a fantastic, atmostphereic, unnerving level, but ‘Life of the Party’ is what the Thief experience is all about.

  4. John P (Katsumoto) says:

    To be honest, I saw nothing in T:DS that implied it was being developed cross-platform. I genuinely think it is the best in the series and levels such as (the one when you sneak into the pirate captain’s mansion) really stand out for me, even over the cradle, which just scared me shitless. So yeah.

  5. The Sombrero Kid says:

    I just noticed Jordan Thomas quoted Session 9 as an influence on the cradle in the interview it, that was the main influence for the group project we had to do at Uni, it was a prototype of a game called The Mind of Kaminsky & didn’t exactly live up to our expectations, especially as now i realise the cradle was what I was trying to make :(

  6. simonkaye says:

    So no sooner do I finally complete Bioshock than I am reminded of Deadly Shadows. And then I read all the reviews and the special feature on the Cradle. And now I’m scrabbling around for the disc.

    So much for the dissertation.

    Metal Age was clearly the best in the series, btw. Those robot things… god.

  7. Meocene says:

    I’ve never said this anywhere else before because I knew i’d get slammed, but i really didn’t think much of the cradle level. The cradle level scared me twice, and that was it. The first time was the knocking noise coming from up that spiral staircase. dear god I’ve never taken so long to climb a set of stairs. the second time was when you briefly glimpsed something dart across the reception area, after that it was kinda pants i thought. Not the level design, that was exceptional, but the scare factor. I liked the electroshock patients as a concept, but in practise they were pants. I think the cradle level would have been supurb if it had been devoid of any real enemies, but things like the knocking and the little critter running in front of you just kept happening. How freaking scary would it have been if the electroshock guys were simply ghosts that caused the lights to flicker, and them something very solid darted right in front of you which you only just catch during a flicker of the lighting. if that level of tension had been maintaned throughout the level it wouldn’t have mattered that the level was empty. it’s what you can’t see that scares the pants off you. It could have been a proper nightmare to play level that left you utterly relieved to finish, as it was I soon saw through the flawed AI and just walked through the level. Still, it was a good game, but only good. When i irst heard the T3 was going to have a fully explorable city I couldn’t wait. I had visions of the Thiefs highway rendered in glorious dynamic lighting… pity it just wasn’t to be.

    And in my opinion, Inv War, although highly flawed in many, many ways was a master piece of story telling, and i prefer it to the origonal for that reason alone.

  8. simonkaye says:

    I think it’s possible to ‘break’ the tension and horror of the Cradle, just as in any level in any Thief game. You almost need to approach it as a role-player. Total avoidance of the electroshock bads – the most realistic response to them – would ensure that you barely reveal their limitations. They only work for as long as you can’t quite be sure of their capabilities. After that, they’re just jittery bots.

    Also, not sure I’m with you on IW.

  9. Meocene says:

    i know what you mean, but i am a BIG role player. I love to loose myself in a game and play it as if I was actually there (god bless BioShock, it’s touched on something, but I’m also disappointed in it, mainly due to the fact i let my imagination run away with me, but I blame Levine for that. He bigged up the AI was beyond what it was actually capable of. Ecology my ass). Back to T3 though, I’d say the game broke my illusions, and i’ve never had any problems filling in the blanks.

    I personally think the next big think in game’s is not physics, I think (hope) it’ll be AI. So we can dedicate 1 of either 2, 3, 4 or 8 cores to physics, whoopty do. What i want to see is what Levine claimed was in Bioshock – a proper AI ecology (Stalker came way closer to this than Bioshock imo).

    To quote the dictionary:
    1. the branch of biology dealing with the relations and interactions between organisms and their environment, including other organisms.

    Focus on “including other organisms”

    Now where exactly was that in Bioshock?
    There is only one example of this in Bioshock, the daddies and sisters. I remember when i saw the first few vids of Bioshock in action, where you had a bigD looking after an littleS, whom were then attacked by a splicer, and that was very cool. I envisioned Bioshock as a game in which a fully fledged ecology had evolved from the ruins of Rapture, in to which the player suddenly finds themselves, and i thought your actions would slowly but directly affect how aspects of that ecology reacted to you (ad then on top of that a story)
    Some would learn to fear you, recognising you as a foe, others a friend. Say for example you come in to a room in which there’s a bigD, a littleS and some splicers. The bigD is doing his best to protect the little sis by _not_ wading straight in to the fight, but by taking the hits. Much like a father or mother would for a child. The bigDs primary function, as programmed in to his basic but combat effective mind is to ensure that the LS is not harmed in any way – so he backs into a corner, hides the LS behind him and starts getting pounded – he knows that due to the number of foes if he goes on the attack the chances of the littleS getting killed are too high (that’s what I mean by combat effective). He has no concept of death and so this appears to him to be the best course of action. His demise is inevitable, as is hers. Now you have a choice, either help finish off the BigD to get to the Adam, thus siding with the splicers, or wade in to the splicers to help the big D. at the end of the first, whichever side is left standing starts to recognise you as a friend, so on and so forth – much like the gangs in GTA. The moral decision was not whether or not to rescue the littleS’s, they’re already lost to Rapture, but how you manipulate the ecology.
    The truly morally sick moment could have been finding that Fountaine had programmed the bidD’s to see him as a friend even while watching him butcher a littleS to get a higher yield of Adam.
    Now that, in my humble opinion is an example of emergent game play based around an ecology

    As you can probably tell, I’m no programmer, and I’m a little drunk, but i do find it hard to believe that this kind of AI is impossibly complex, especially after seeing what Black and White’s AI was capable of… but then that was just one cow i suppose. roll on a core dedicated to AI, that’ s what i’m really trying to say.

  10. steven peterson says:

    just another name in the long litany of consolized crap

  11. Halo Jones says:

    Episodic Thief content?


  12. Sören Höglund says:

    “Metal Age was clearly the best in the series, btw. Those robot things… god.”

    I’d still rate the first one the highest. The story, atmosphere and gameplay gel the best there. The Metal Age has all those great open cityscapes to explore, true, but they come at the expense of a more poorly developed story, and the art direction is pretty jarring considering what came before and the short time that’s elapsed in the game fiction.

    DS was a fine game, althouogh the technical constraints hampered them quite a bit, and I really wish they’d found the time to explore vertical space more with the climbing gloves.

  13. Anyone says:

    How could you possible screw up Thief 3? All you had to do was to make Thief 2 w/ better graphics & build a more solid level editor. DONE! Yet, somehow, w/ more advanced tech then they had w/ Thief 2, they built some piece of consolized crap w/ no rope arrow, blue mist loading zones, 3rd person perspective, faction mini-games, no swimming…

    Here’s to hoping The Dark Mod is actually finished & that Thief fans finally get the “sequel” that is worthy of the Thief franchise…

  14. Winterborn says:

    “I’d still rate the first one the highest. The story, atmosphere and gameplay gel the best there.”

    I agree.

    I also agree with Meocene on ecology, it’s really what I need to enjoy an action game. Good AI goes so much further toward immersing me than good graphics.

  15. Winterborn says:

    Also… new site design, interesting.. did that just go live?

    I miss the old headers though.

  16. John P (Katsumoto) says:

    Rar! I still don’t see how it was consolised? Rope replaced with the gloves which work even better imo, 3rd person perspective was entirely optional and actually helped in areas, and it was entirely optional in the console version as well. Blue mist loading zones? How is that dumbing down? Its just a transition. Faction mini-games? Huh? You mean to get your rep up? They were hardly “mini games”, it was just extra stuff you could do around the city.

    All of those are new features but none of them to me seem to be explicitly to do with consoles. As i say, for me, TDS is the best in the series by a long shot. I just think it gets panned by the same people who unfairly panned Invisible War – just because it was also on xbox doesn’t automatically make it shite! Admittedly the console roots of IW were more obvious than DS, tiny levels being the most obvious problem in comparison to the original. But Thief’s levels managed to portray some really large areas, even if they were broken up with some (really short loading) portals.

    TDS remains one of my favourite games of all time, up there with Deus Ex et al. Then again, I think Call of Juarez is a work of genius but I definately seem to be alone on that one! OH well!

  17. Kieron Gillen says:

    Randomly, here’s some nasty errors in the piece which I presume happened when I was formatting it hungover yesterday. When my computer is working again – my 3D card may have just died – I’ll go back to the copy to get the fixes.


  18. Jim Rossignol says:

    “What i want to see is what Levine claimed was in Bioshock – a proper AI ecology (Stalker came way closer to this than Bioshock imo).”

    Aye. I think Stalker was actually a more interesting game than Bioshock for precisely these kinds of reasons.

  19. Acosta says:

    From a pure gameplayviewpoint, yes, I am agree, S.T.A.L.K.E.R is more interesting that Bioshock and nails the ecology part much better. But the way Bioshock is crafter is fantastic and it has other virtues.

    But I still hope some huge change in following the way of S.T.A.L.K.E.R. It´s a key for getting more interesting virtual worlds.

    Great interview. I think some PC players need to calm down, they start seeing blue footprints and they feel terribly insulted. Deadly Shadow had a good compromise, (very far of Deus Ex 2, that was heavily compromised in my opinion).

  20. Nickbjorn says:

    Personally I loved all 3 Thief games and struggle to choose between them, the whole concept is so good and the ‘world’ was so cool that all I can say is that I want more Thief!

    Episodic content sounds great, but, like ‘Anyone’ says, I’m very excited about the Dark Mod

  21. Meocene says:

    never heard of this dark mod thing, just checked it out… understandable excited! the rope arrow physics look ace, mucho potential.

  22. Eric Blade says:

    The STALKER AI was just totally garbage, I don’t know why everyone says it was awesome. It was as bad as Land of the Dead, and that’s really damned bad. Seriously. I was wandering around in that giant bug-ridden mess they called a world, and dudes would stand inside things, shooting me. Through the walls. They’d be on one side of the wall, their guns and legs would go right through the wall, you could see them running, and shooting, even thoguh they were on the other side of a wall, and you’d be getting killed.

    You could stand there with a pistol at as far a range as you could see (or farther, really) and pick off people, and their companions sitting right next to them wouldn’t even notice.

    I ran the first 3? or so patches on the game, and none of them fixed any of the super obvious bugs (like, oh, a sky that was nothing but a mirror of what was on the ground), and never finished that game. It was the absolute worst $40 I’ve ever spent on a game. Ever.

    Deus Ex 2 was the worst game ever created. Period. The first one was a masterpiece, the second one they fixed all the things that weren’t right in the first one, then threw out all the things that were right in the first one, and then wrote a story that took my total excitement about playing a sequel to one of the greatest games ever, and continuing on in the world post J.C. Denton, and totally destroyed it. By the time I was halfway through DX2 (one of two games that I have started and not finished in the last 10 years or so – the other being Stalker), I was so apathetic to the plight of my character, and so frustrated by the incredible flaws, that after it crashed going into the second half of the game, i uninstalled it, and read a walkthrough online to find out if the story EVER tied back into the Dentons or the events of DX1 (and when it does, **SPOILER ALERT** you kill the Dentons! WTF?!)

    I probably should dig up the Thief series, I played the first one, but that was so long ago I can’t even remember it.

  23. roBurky says:

    But there wasn’t any ecology in the released Stalker was there? There was just a warzone, as if it was a WW2 game.

  24. James says:

    > Deus Ex 2 was the worst game ever created. Period.

    I think it was an accomplished game in its own right. I can’t think of anything super-wrong with it off the top of my head. If you’re comparing it to its predecessor, then sure, but the commercial sequel very rarely lives up to expectation. There was still a very good game there.

  25. Richard says:

    I disagree – I thought Deus Ex 2 was garbage at the time. Writing, story, world design, quests, mechanics… just crap, and crap with some very bizarre mistakes – right from the opening scenes, where you’re supposedly making judgement calls on people and organisations you’ve never even heard of, right up to the ending.

    I replayed it a while ago after needing some shots from the original, and it’s still horrible. It failed miserably at what Deus Ex had made a decent stab at, and was a rubbish shooter/RPG in its own right as well. The handful of genuinely fun bits in it (NG Resonance, um.. er… aaah…) were drowned by some of the most insipid cyberpunk crap it’s ever been my misfortune to play.

    Without the license and the names attached, it’d have been down in the 50-60% bargain bucket for obscure Russian imports and clones that just didn’t get why their original worked. At best.

    i uninstalled it, and read a walkthrough online to find out if the story EVER tied back into the Dentons or the events of DX1 (and when it does, **SPOILER ALERT** you kill the Dentons! WTF?!)

    Well, on a couple of paths you do. You don’t have to.

  26. Richard says:

    Grr. Blockquote not working then…

  27. Jim Rossignol says:

    >> The STALKER AI was just totally garbage…

    Funny, I experienced none of the bugs Eric mentions here. I did experience bugs, but they were almost all with the scripting.

  28. Andrew says:

    Same as Jim re: STALKER’s AI. And I definitely saw AI-ecology going on. It wasn’t just a warzone, and it wasn’t scripted stuff. It could mess up the scripting, in fact.

    I also greatly enjoyed Thief 3, warts and all. Thief 2 is probably my favourite of the series, and it was the one I played last of all. I somehow skipped it originally, getting Thief 1 way back when then going straight to Deadly Shadows. All three are great games, though.

  29. Ben Hazell says:

    Stalkers AI was very alive – I never saw enemies fail to react to firing. If anything, I saw a single noise stir whole camps full of troops out into the open to hunt for you. Only the odd respawns felt clunky to me – the Duty checkpoint from the garbage where a full squad always reappeared.
    If only Stalker had really let you work with the factions rather than just odd jobbing. You could break the atmosphere through legitimate roleplaying choices. Did anyone else manage to get to the Barkeep after annoying Duty?

  30. Alec Meer says:

    Yes. But only barely, and by having to murder pretty much everyone else in the game so I could get to him. Most annoyingly, I hadn’t actually annoyed Duty – a (pre-patch) bug just caused them to decide I had.
    I may post about it at some point actually, as it made Stalker a very different experience than the one I was supposed to have.

  31. John P (Katsumoto) says:

    Hmm! That reminds me of the time I was hanging around near the bar, and suddenly bullets started flying – i thought there had been an invasion by the mutants/another faction and was really excited! It was only after about 15 minutes that I noticed all the bullets were aimed at me. I’m not sure what to make of that

  32. Ben Hazell says:

    Still, it was quite an experience to have to clean out a whole town, then have to sneak quietly into the one room where I wasn’t allowed a weapon.
    Lets have a post where we can mutter about Stalker in the comments rather than hiding in the Thief zone…

  33. Robert Yang says:

    Regarding Ion Storm being too “academic”… I know the feeling. I’m looking into my university’s “New Media” program and it’s all this silly academic jargon about redefining the act of play and whatnot. Do they ever put all this theory into practice? Do they actually go out to design games? Most of them don’t, and it’s really frustrating when I’m trying to discuss the gameplay systems at work and they keep going back to a cultural context that often has little relevance to anything.

  34. Rock, Paper, Shotgun » Blog Archive » Dungeon of Regret says:

    […] Kieron’s splendid Thief 3 retrospective attracted a whole bunch of comments discussing how both it and Deus Ex spawned sequels that didn’t tickle their respective fanbases in all the right places, and thus caused the apparent end of their respective lineages. […]

  35. i dont give two shits says:

    i dont care what some of you guys think about it thief 3, the point is i dont give two shits about the graphics or the short story line, to me thief 3 is a sweet game its all about the experience and for me, it was sweet :P

  36. GIFT OF GAB: Jordan Thomas Interview « Don't Lose Your Day Job says:

    […] This is a quote I had pulled from an interview you did for ‘Deadly Shadows,’ “When pushing the right boundaries, promotion of simulation and […]

  37. chinook says:

    That last paragraph is so amusing (and sad) to read now when we know the direction Thi4f is headed.