The title of a thriller is paramount, and the title of a talk about thrillers is no less so. Luckily (and perhaps a little cynically), neither has to have much bearing on the contents. After long and careful thought, I've opted for:

Chills & Spills at the Thrillsville Grill

Before that it was: "By the Numbers: Using 20th Century Mathematics to Deconstruct 20th Century Crime Fiction". And before that it was: " 'Rithmetic': The Third R In 'Thriller'". As you can see, that I've had a little trouble in this area.

And, to be completely honest, that's not the only one. This is the first time I've been invited to speak on a subject that isn't specifically related to speculative fiction. Usually people want to know about what SF is, or how to come up with and develop ideas, or how to sell it. Very few people ask science fiction writers -- perhaps genre writers in general -- about the craft of writing, because either the details that distinguish the genre from the mainstream overshadow any other aspects of the writer's work that day, or else it's assumed that the writer has no knowledge of the craft.

The last is obviously not true -- in the case of most of the writers I know, anyway. But it is true that I have had no formal training in the art of writing or analysing literature. I therefore found it hard to determine exactly what it was I wanted to say today -- and even when I did, I didn't have the references in my personal library to back me up, apart from a dictionary, some mathematical textbooks and back issues of New Scientist.

This explains the original titles, anyway, and the focus of this talk. As that ancient Japanese proverb goes: "When all you've got is a hammer, everything looks like a nail."

Irrespective of this, I'm both honoured and pleased to be here today. For the thriller/mystery form is not just one of my personal favourites, but is, along with the quest, one of the most prevalent in the world. Without it, the publishing industry today would look very different indeed.

*

The best place to start would be to address the question of: "Who decides whether something is a thriller, or a mystery, or neither?" In the case of my own work, the back of my first novel, Metal Fatigue, displays, in bold print, the label: "Thriller/Science Fiction". The "Thriller" part even comes first. My latest novel, The Resurrected Man, doesn't have the same label on the back, maybe because the science is a little tougher, but it, even more than Metal Fatigue, is constructed along these lines. Certainly the feedback I've had has judged it to be squarely in this camp.

Clearly, then, the publishers of my books and the readers all have a fairly large say in what is a thriller or not, whether I, the author, happen to agree or not. It's the same with all books. But on what do we base our opinions? Is there a benchmark, archetypal "mystery" or "thriller" against which all others are judged, in order to prevent person A's version from being completely different to person B's?

We should look for some definitions. My trusty Collins English Dictionary defines a thriller as:

"a book, film, play etc depicting crime, mystery, or espionage in an atmosphere of excitement and suspense".

That sounds pretty reasonable to me. My aim is to do everything possible in order to stop the reader from putting down the book. That's one of the two things I want from a thriller, anyway. (The other is to take me completely by surprise.) The dictionary isn't much use helping me find ways to describe what "suspense" is, at this point, or how it might help me go about my job beter, but its definition of "intrigue" is insightful:

"to make interested or curious, to conspire".

That's exactly what I'm doing: I'm conspiring against the reader to disturb their sleep patterns for a week. There are lots of ways to keep a reader curious, some of which I'll touch on later. Suffice it to say that adrenalin is a key point: anything that sets the heart pounding, or sinking, or soaring, is fair game.

Compare the thriller, then, with the mystery, which according to Collins once again is:

"a story, film etc which arouses suspense and curiosity because of facts concealed".

Here, the concealment of information -- identity, most often -- is the central premise of the work, the principle method of keeping the reader reading.

(And note also that here, again, the emphasis is not solely on books. Thrillers and mysteries are as prevalent in cinema as they are in the library. Since I believe they demonstrate the same qualities of structure and flow as books, I will be referring to movies fairly frequently throughout this talk.)

The classic mystery plot is deceptively simple: the reader (through the eyes of a sleuth, or sleuths) must determine who is the guilty party from a set of known suspects. Facts are doled out as the investigation proceeds, and a clever author will ensure that the reader realises the truth at the same time as, or just after, the protagonist. There's nothing worse than solving a mystery in the first third of the book, except perhaps a boring thriller.

One of the great things about good thrillers is that they can be read over and over again, with little diminishing returns. Mysteries, however, hang on that sole central fact which, once known, devalues the reading experience.

In both, the payoff occurs when the writer has hooked you, started to reel you in, let out just enough slack so think you're getting away, then blown you out of the water and left you gasping for breath. That's when you know that all those times you looked at the clock and thought "Just another ten minutes" were worth it.

But how exactly do good thriller and mystery writers manage this? I'm not sure I can answer that to everyone's satisfaction. All I can do is describe what I look for in a story, and therefore what I try to emulate in my own writing. Looking at my own work, in turn, and trying to see how I've done what I've done, I've been able to make some sweeping generalisations about thrillers that might, fingers crossed, ring true.

*

I should say first, though, that I don't regard "crime" to be a separate genre, in the same way thrillers or mystery are. That is, I don't see that police procedurals, for instance, are markedly or uniquely different from a police procedural that also happens to be a thriller (or indeed a "post-Apocalyptic police procedural", as Metal Fatigue was once called). Maybe that's due purely to my own limited experience of what is marketed as "crime", but I do know that I couldn't find a word in my dictionary describing the genre or the payoff one gets from reading "crime" books.

The closest I could come to a definition of such is to borrow the definition from "crime" itself:

"A crime novel is a depicition of unlawful acts; that is, acts or omissions prohibited and punishable by law."

Since any thriller or mystery could be defined that way too, as can an awful lot of fiction, maybe the "crime" label could serve as an umbrella term to cover it all.

Either way, my two books do fit under that umbrella. They both have, at their heart, science fiction ideas that are explored and extrapolated and interact with the major characters, who generally work in or work to avoid complex legal and political systems that owe much to the ones our present-day police enjoy. Even though they follow future police procedurals, the patterns of investigation, accusation and vindication are recognisably the same.

In both there is also a mystery of identity requiring resolution -- that is, one of the major players turns out to be the antagonist -- but to me this aspect of the novel is secondary to the sfnal idea, and is treated accordingly. In Metal Fatigue the identity of the criminal is signalled by the way that character is drawn; in The Resurrected Man, I actually tell the reader who he is in the first hundred or so pages.

So my books are crime-based: they exploit the mystery form as well explore sfnal ideas, and I personally regard them to be thrillers. As far as cross-genre hopping goes, that may seem a bit of a handful, but I don't believe it is, necessarily. What is true for a mystery-thriller will, I believe, also hold for a sf-thriller, or a western-thriller, or even a romance-thriller.

*

The commonality between the various types of thrillers in the marketplace arises not via a deliberate effort on behalf of the writers, but, in my opinion, because it is an inherent property of what a thriller is. It's certainly not something I work through consciously during the development of my books, although I may be aware of it happening at times. It's an intuitive and organic process I've learned mostly from reading, and the rest by trial and error. Because it's therefore a product of my own experiences, it may not be relevent to anyone else. But I do believe that it is a useful way of looking at things. All writers, after all, undergo a similar process as they learn their art.

Without exhibiting evidence of two underlying principles, I believe, a thriller may not work. Like all such universal principles, they sound deceptively simple.

In writing a thriller, structure and flow are everything. Specifically, the structure must be fractal in nature, and the flow must be chaotic.

We've all probably heard these buzz-words in the last few years, even though they're not words usually associated with fiction of any kind -- except perhaps science fiction, and even then only with respect to what the story is about, not how it works.. Like that other word swooped upon by pseudo-science -- "quantum" -- they describe some of the most fundamental discoveries about nature made this century. The more deeply scientists look into the world around us, the more they seem to be finding evidence of fractal design and chaotic behaviour. Writing, the productive of decidedly natural creatures -- humans -- is no exception.

Just in case you're not familar with the terms, I'll try my best to explain, briefly ... (and apologies to any mathematicians in the audience for any mistakes I make along the way) ...

*

New Scientist suggests this simple experiment to demonstrate chaos in action:

"take a simple pendulum, push it repeatedly at regular intervals and, despite the simplicity and rhythm of your pushes, its motion can turn out to be very irregular."

The interaction between the pendulum's natural rate of oscillation, determined by the length of the string, and the timing of your pokes at it, cannot necessarily be predicted: there is a fundamental uncertainty that, despite the simplicity of the functions operating on the system, can take tiny changes in the way you start the experiment and magnify them to the point where you end up somewhere quite different to where you expected to be.

Examples of chaotic behaviour are to be found in EEGs and ECGs, where the lack of chaos in either might signal the onset of epilepsy or a heart murmur. It's found in the way populations rise and fall, in the orbit of the planets, anywhere a small variation in initial conditions can lead to widely varying outcomes.

The concept that simple, comprehensible patterns of behaviour, can lead to unpredictable results should be an intuitive one since it's so prevalent in the world around us. The fact that it isn't shows how deeply flawed our commonsense understanding of the world is, and explains why we're always surprised when the weather doesn't turn out the way we expect.

Fractals, on the other hand, look at the way things organise themselves, and can be summarised simply:

"As above, so below."

Any structure that repeats itself at different scales -- be it over time or with increased magnification -- is said to be fractal. This structure usually emerges a result of some sort of innate property belonging to the components of which the structure is composed. In other words, whether you're stacking oranges or water molecules, the shape of the objects determines the patterns they'll eventually fall into.

Fractal geometry is found in coastlines and the money market. It explains why the veins in your arm look like the Nile from orbit, and why the Mandelbrot Set looks something like a beetle. Together with chaos theory, it has helped me see the processes and patterns that help make a thriller ... thrilling.

*

I'll start with chaos theory first. A good thriller should grab me immediately. If my interest can be likened to a spring, the author must take hold of one end and stretch it, creating tension within the spring, within me. I want to know what's going to happen; I want to see how the conflict in which the protagonist has been embroiled is resolved. Throughout the course of the book, the tension I'm feeling will ease or increase, depending on what's happening. The protagonist gets deeper into peril, I grip the book a little tighter; the antagonist takes a temporary fall, I relax. The tension doesn't ease completely until the end, when the author's job is done, and the spring inside me is released, more or less.

This isn't a new thought, or a new metaphor. And as it stands it is, in my opinion, somewhat limited. A couple of years ago, West Australian film critic Robin Pen published a concise guide to the plot of every James Cameron movie ever made. The plot progression not only applies to many other films of the thriller type, but is appallingly simple. Robin charts a regular rise and fall on what he calls the Action/Emotion Level graph.

[show diagram from The Secret Life of Rubber-Suit Monsters, page 62]

This is the simple bouncing spring in action: perfectly made, potentially still enjoyable, but not the very best pattern possible. Indeed, it becomes slightly tedious once you're aware of it. Such recurring patterns of tension/release can emerge in prolific writers or other artists, who we then say are working to a formula. As well as character, setting or theme, this aspect of their work can be a major distinguishing factor.

So what makes a better thriller? A good start is to keep the reader from guessing what is going to happen next -- whether she is going to be excited or relaxed by the words on the page ahead. Sometimes you can just tell when things are going to hot up -- and if they don't, you're more likely to be startled. In other words, the tension in the spring, its bouncing up and down -- in the same manner of the the pendulum from before -- must become unpredictable, or chaotic.

This, to me, is a perfect representation of how the plot of a thriller should work. I don't mean plot in the sense of the complete story arc -- the interactions between the characters and their imaginary environment, complete in the author's mind and fixed once the page is printed -- but the unfolding, intuitive sense the reader enjoys as the book progresses. It's more of a rhythm than a structure, a metre dictated by the interactions between all the different aspects of the book, evinced by the ebbing and flowing of tension, of conflict.

*

Underlying these aspects, driving the characters along, are usually a handful of simple motivations -- love, greed, hatred, patriotism, revenge, and so on. These point to another field of mathematics which can be relevant here. Complexity theory deals with the behaviour of systems composed of a large number of units obeying a handful of basic, known laws. These units could be ants, or shoppers in a department store, or subatomic particles. What complexity theory shows is that large-scale order -- ant hives resting during the day, rushes for checkouts, crystals -- arise spontaneously from the mess of units even though there is no overseer. These so-called "emergent properties" are like the story of a novel arising out of the combined behaviour of the characters.

The knowledge we acquire of the main characters' behavioural codes, or goads, does not mean that we can predict the behaviour of the collective plot. Not even the author can do that, at times. (As many will relate, characters have a tendency to do their own thing, not necessarily what the author wants.) But when we do know the entire plot, we can see how it emerged from the combination of these rules. The plot is therefore internally consistent. It is neither random, which is a completely different thing, nor contrived.

This last for me is very important. For something to come out of nowhere -- a deus ex machina -- without justification or prior explanation, is cheating. This very thing happened in the movie Armageddon, in which a shower of meteorites happened for no apparent reason, serving only to use up some of the special effects budget, to kill a couple of characters and, ultimately, to damage a device upon which the success of the protagonist depended.

I hate being cheated, and I hate being lulled to sleep by predictability. I love being surprised, and I love knowing afterwards that, even though I couldn't predict that something was going to happen as it happened, I could see precisely why it did afterwards. And I think that here lies the reason why thrillers do stand up to re-reading: there's always the feeling that the story isn't fixed -- that, with just a little nudge at the beginning, it could go off anywhere.

*

So the overall story arc of a chaotic thriller could look something like this. [draw a story arc here, lots of different ups and downs] Or starting with the same assumptions, it could be something like this. [draw another story arc] It could be anything. But it will almost certainly be more believable than the mechanical bouncing of the plot Robin Pen theorised.

Pulp Fiction is a good example of what I'm talking about. An undeniably a crime-based story, although may not be a thriller, it has a wonderfully organic flow. The odd leaps and twists are unexpected yet quite natural; I found it completely impossible to guess what would happen next. In that sense, it is one of the most satisfying movie experiences I've ever had.

Whether you liked Pulp Fiction or not, I'm sure that you'd all be able to name a thriller you like and sure also that it would exhibit a measure of complexity and chaos. They are fundamental parts of all living things, and we therefore unconsciously regard things that exhibit chaotic behaviour as being more alive. By making their work unpredictable yet comprehensible, thriller writers help the reader believe in their work.

*

But this only describes how the plot unfolds. How it is told depends on the fractal nature of the text. It's all very well to have a story-line that is internally consistent and progresses naturally, but that doesn't guarantee that the reader will feel the tension they should.

Here we come back to suspense, and my trusty Collins:

"the condition of being insecure or uncertain, in a state of mental anxiety" due to the "approach of a climax" of some sort.

How does the author achieve that state in his or her readers? Definitions of "suspension" use evocative words like "deferment, postponement, interruption". "To suspend" means to "postpone" or to "hang from above so as to permit free movement" (there's our weight on a string again).

If an author gives a character a problem, a reader will be interested to see how that character deals with it. That, at its most fundamental level, is how plotted fiction works. It depends of course on a number of other factors, such as whether the character is one the reader will associate with, or whether the style of the prose will be one the reader will enjoy. If they find nothing to dislike in these, they will probably keep reading.

Beneath the peripherals lies the basic situation requiring resolution. Who killed Mr Burns? Does Frodo successfully manage to destroy the One Ring? Will Harry Stamper save the Earth from the giant asteroid?

On it's own, though, a single premise makes for a fairly dull story. As I mentioned before, it's the presence of sub-plots that results in a more complex plot-line. Even Armageddon had these. That the main character's daughter had fallen in love with a man he deemed unacceptable was one such situation requiring closure. Regardless what one feels about the nature of the sub-plot or the way it is handled, the fact that it is there, jogging along beside and occasionally interfering with the main premise, indicates an awareness that complexity adds to the story-telling experience.

*

We can continue to dissect the story of Armageddon. Each act of the movie has its own miniature story arc. Will Harry Stamper agree to save the world? Can Stamper and his team survive training and make it to the asteroid in one piece? Will they manage to destroy the asteroid before it hits?

Within these acts there are sub-dilemmas: How will AJ escape from the disintegrating Russian space station? Can Stamper convince the nasty army guy to let them do their job? Will the same army guy -- now good -- defuse the warhead in time? Do AJ and his pals manage to jump from one side of the asteroid to the other without drifting off into space and being lost forever?

And below these there are more still, some of them blatant cliches: Red wire or blue? Fix the problem using brute force or thought? Can Stamper push the button at the last possible split-second?

There are problems that are not resolved within the confines of the story. "Is Stamper a macho, self-serving jerk?" was a recurring one, for me. It lingered beyond the frame, nagging at me long after the viewing experience was over.

Just as the time-frame in which the resolution occurs varies, so too does the nature of the conflict. The obvious source of conflict is the one to which the main plot is pinned, but overuse of this will desensitise the audience. It need not be something facing the protagonist, or even an example of what the script-writers of Armageddon would seem to regard as being "conflict" -- that being outright physical confrontation. Perhaps "discord" is a better word.

The ways of creating "discord" are as numerous as the stories available to tell. Any unanswered question leaves a gap in the puzzle the reader is mentally trying to make in her head.

The important thing is that this presentation to the audience of a situation that needs to be resolved, then either resolving it immediately or deferring the resolution until later, happens continuously throughout the story. A situation that has not been resolved adds to the reader's feeling of suspense. It nags at them. It adds to the force tugging on the internal spring.

*

If we chart the sum fluctuations of discord throughout a story -- the tension quotient, if you like -- we arrive at the overall plot line we had before. The major dips and crests correspond to the moments during which the story's main conflict was predominant. [show chaotic diagram]

But if we magnify a portion of the curve, we'll see a series of smaller dips and crests. [another diagram] These are the motions of the sub-plots, the smaller matters that come and go through the various acts.

Magnify these, too, and you'll find further examples of this sort of motion, following the tension levels in every scene. Tension can even dip, to a much lesser degree of course, between lines of dialogue. This property of a pattern being repeated at different scales is what is exibited by something that is "fractal".

Bound demonstrates this property well. Tension within a scene, and throughout the movie, is finely calculated to keep the viewer on edge at all times. Discords come and go; they accumulate and overlap; one by one they fall away, until at the very end, most if not all have been resolved.

Since information is power, and a story-teller has absolute control over the dissemination of information, this on-going push-me-pull-you dance between question and answer -- conflict and resolution, discord and harmony, tension and relaxation, whatever -- lies at the heart of a good thriller, and is, I think, the principal tool by which an author keeps tugging a reader along.

*

But it clearly isn't enough on its own. Look at Armageddon, which had it but certainly wasn't a satisfying story, in my opinion. The plot was decidedly ho-hum, and the execution cliched.

It's the same with chaos. A plot we can't anticipate -- such as in The Pelican Brief, or Jade, or The Net -- doesn't guarantee its believability, or that we'll enjoy its execution.

We have to have both and chaos and fractals to hit close to the mark.

How does the author go about instilling both in her work? I don't think there's any one sure-fire method. If there was and I knew it, I'd be a bestseller a dozen times over. And I certainly wouldn't be broadcasting it. But what I can do is tell you how I go about trying to write such a book. Whether it works or not is for my readers to decide. Those of you who are writers might see parallels in the way they write, or they might find a new way of looking at their work. Or not. As I said earlier, this is very personal stuff. I'm fully prepared to accept that my recipe for thriller soup -- chaos with a dash of complexity bubbling away in a fractal pot -- might be a very local delicacy.

The rules I follow are fairly simple. Not all of them are specifically related to thrillers. One applies to science fiction as well as it does to any other form of literature, and that is to have a core idea that is both intriguing and fertile (in the sense that it inspires a large number of secondary ideas). In RM, that idea is to take the common-place trope of "d-mat" and use it as a means to facilitate the perfect, and perfectly ambiguous, murder. The novel revolves around that idea, and, to an extent, exists solely to explore it. In the same way, Patricia Cornwell seeks out and exploits unusual means of dying; as the sum total of detective novels increase, authors must pursue ever-more arcane routes to find ways to be original.

This is the framework of the novel, if you like. To it, I attach the mechanisms of creating tension.

*

There are some that apply specifically to plotted novels. Making sure the characters have clearly-defined problems is one such. That way the reader will have no doubt what the conflicts are. In RM, for instance, Jonah wants to know who killed his father, while Marylin wants to know if Jonah is a serial killer. The need to resolve these problems underlies every action they take in the book. This doesn't mean that the characters are stereotypes, though; Jonah was never particularly fond of his father, and Marylin isn't entirely sure whether she wants Jonah to be guilty or not. There are few genuinely black and white characters in the real world, so that should be reflected in novels too. Saying that the problems should be clearly-defined simply means that we give complex characters more room to evolve around them, in a plot that, in turn, becomes more complex as the agents occupying it increase in number.

During the first half of any book, I multiply problems, on the whole, so that my protagonists have their hands well and truly full. By mid-way through Jonah still doesn't know who put him in a coma for three years, and Marylin is having a hard time dealing with the administration over-riding her investigation. As the book unwinds, the secondary problems are stripped back to core issues, which are either dealt with or destroy the characters they apply to. By the end of RM Jonah does know who killed his father, and Marylin does know if Jonah is guilty or not. Anything that happens after that point is, arguably, unnecessary.

In order to handle the many conflicts I put in my novels, I follow two basic rules: Resolve Nothing and Anticipate Everything. These are two rules in particular that, I feel, lend a finished work the evidence of chaos and fractals it needs to function as a thriller.

*

By avoiding the temptation to let the protagonists solve any of their problems too early, or at least until a new problem arises, I ensure that I obtain the maximum effect from that problem. It also allows me to multiply problems, so that at any given time a character will have several things on her mind. I find the problem/solution, new problem/new solution progression of some novels very tedious, even when the problems themselves may be fascinating. Solve problem A before problem B has appeared, and the reader's more likely to think "Oh well, that's that then" and go find something else to do. If, on the other hand, problem A isn't solved until problem E is on the field, while in the meantime problem B has caused C and D, which is still making itself felt, the reader is much more likely to keep reading.

The conflicts don't have to be resolved in the same order they appeared, or in order of ascending importance. It's more realistic, in fact, if they're not. And as the problems may be both small scale and large, their treatment will inevitably be different. Jonah and Marylin, for instance, spend the entire book trying to sort out their feelings for each other, and even at the end it's not completely resolved. Along the way, they face and try to deal with numerous other challenges, from catching a serial killer to catching a simple virus. All contribute to the flux in the "tension quotient" throughout the book.

This overlap of investigations and dilemmas gives both a chaotic and complex shape to the plot, while giving us plenty of opportunities to create a fractal flow through the story.

*

The anticipation of all possibilities is just as crucial, although it may not be as evident in the plot. The author must give the reader no opportunity to wonder why the protagonist simply didn't call the police, or fire the gun, or tell the truth, or any number of sensible things that might ruin a perfectly good climax. The reasons can be many -- ranging from simple mechanical failure to deeply-buried childhood traumas -- but they must be there. An author must also ensure that, if the protagonist couldn't call the police or whatever in chapter two, but they can in chapter twenty-two, the reasons why are obvious and plausible. This ensures internal consistency.

Keeping Things Simple is another rule I try to follow, although, looking at my books, it might be hard to see this principle in action. In RM, for instance, the characters not only have clearly-defined goals, but those goals can be met quite simply; it's finding out what they need to know that's difficult for them. Ideally, the entire mess of plot and sub-plots should revolve around a bare minimum of facts which, when known, unravels the lot. Jonah, in solving the case of the Twinmaker, has to come to three separate realisations, but once he has them, he has enough. All that remains from then on is summation.

[In the case of RM, this summation is quite long. This is partly because I have an enduring love for Agatha Christie-style endings in which the sleuth shows how clever she is by forcing the antagonist's hand in front of all the other major players. It's also because I found it very difficult to ensure that the full significance of the three important facts was realised by the reader. They had to be prepared for, or else their significance wouldn't be obvious, and after several hundred pages of concealing or down-playing them, the habit was hard to break.]

Ideally, the ultimate thriller revolves around a single piece of information which is given on the opening page of the book -- but the reader doesn't grasp its true significance until the very end. And since events, places and people are allowed a greated inter-connectivity in novels than in real life, that's not as impossible as it might sound.

*

Just as the journey through a novel can be plotted in terms or tension, or action, or emotion, it can be analysed in terms of revelation, too.

The author must follow a practice of selective information dissemination, if you will, doling out facts in carefully measaured doses, a trickle of data designed to lead the reader to the intended conceptual conclusion: who did what, or how, or why. Readers tend to follow the protagonist -- I like to allow the characters to wonder what the hell is going on, making the reader wonder with them, both infodumping and creating tension at once -- but it isn't the only method available. Having the reader know something a character doesn't -- especially if it foreshadows something bad -- is an effective tension-raising trick .

The ways of leading a protagonist to the required realisation are many. They can follow a plodding, patient path from fact to fact, until they have accumulated enough to see the truth. Or they can experience a blinding revelation when all seems most hopeless. Or they can be passive while the antagonist reveals all. Often it's a combination of all three. However it happens, it must make make both logical and intuitive sense. It must satisfy on all levels. Otherwise it would fail to be a complete resolution, and the reader will feel cheated.

There are also numerous ways to prevent the protagonist (and the reader) from realising the truth without simply withholding information. The most obvious is the red herring, revealing something that turns out to be false or misleading. If you have to tell your readers something, do so in a way that will only muddy the situation further: by tossing them a crumb, you'll remind them that there's a whole loaf out there waiting to be found; toss them the crumb in a way that makes them think the loaf is in the wrong place, and you'll put them in a worse position than before.

The most efficient use of a red herring is to reveal information in a way that suggests it relates to problem A, to which it later turns out to have no connection at all; it does, however, relate to problem B -- so the one piece of information serves two purposes in the book: first to mislead, then second to resolve.

Red herrings are, however, something of a cliche. Every thriller or mystery reader is wary of them, and expects early suspects to be misleading. Aware of this, I tried to turn the cliche on its head. By pointing the finger at the killer within the first hundred pages of RM, as I mentioned earlier, I hoped to rule out that suspect in the reader's minds, only to surprise them later with the truth. Attempting to work against the reader's expectations can work for or against an author, since readers have varying tastes for novelty and surprise, but it is another way of investing a plot with just a little more chaos.

In many ways, I also regard thriller writing as being a battle with the readers' expectations. Warping cliches or stereotypes in general is a useful way to keep the readers on their toes. Give them a character that looks like a typical cop in his mid-forties, and many people will assume that to be the case through and through; gradually reveal a cyborg ex-soldier in his mid-nineties, as I did in MF, and you'll not only surprise them but you'll have plenty more opportunuties for conflict between that character and others in the book.

*

The last rule I follow is as important as the first one (which is to have an interesting premise), and that is to leave the reader wanting more. There are lots of good reasons for this. The first relates to the fractal nature of the text. Just as some conflicts can be dealt with quickly, there must be some that will be left unresolved; the story is too confined to contain both their emergence and their dismissal. Assuming that the major issues have been resolved, the odd dangling thread won't matter.

By dangling threads, I don't mean failing to explain from where the bad guy obtained the nuclear weapon with which he planned to blackail North America. Details like this must comprehensively covered, or at least the lack of explanation must be acknowledged. I'm suggesting that it might be better to keep your love interests slightly confused after the killer has been caught. Leave some motivations a mystery for your protagonists to ponder beyond the frame.

There's nothing worse than getting the whole story, the whole truth -- for, just as nature hates a perfectly straight line, so too does it despise a perfect ending. By imitating nature the author can come closer to something which truly lives.

*

And that's what it's all about, really. I want to write a story that could actually happen, somewhere, featuring characters that breathe on their own. Trawling through mathematics textbooks might seem an odd way to go about meeting this end, but I think it has helped me understand what exactly I'm trying to do, and why it works when it does.

As an old English teacher of mine used to say: "You have to know how to obey the rules before you can break them properly." Just as the best golfers understand the physics of their swing and the properties of their clubs, so too must the all writers attempt understand the field in which they aspire. Whether it's a case of Japanese hammers and nails or not, every diverse viewpoint adds to this understanding. In the long run, I hope it has brought all of us, as readers and writers, just that little bit closer to finding the perfect thrill.