Although he needs no introduction for the majority of the OSG readership, Steve Turner was the driving force behind Graftgold as well as being directly involved in the development of several classic games like Magnetron and Uridium on the 8bits which makes him an authority on game design and development. Here are some of Steve’s general thoughts on design, taken from his website and used here with permission.

At the heart of every game, whether it be a multi CD extravanganza or a retro 16k, is the game design. Game play is really what games are really about. Cool graphics and sound immerse the player in the illusion that the game creates and should not be underestimated in importance but when it comes to the crunch it is the game play itself that keeps the player coming back for more. It is often stated that retro games are stronger on gameplay than current games. I think that is a myth , there is good and bad in both., however when graphics and sound were pretty basic a greater percentage of our time concentrated on the thing we could improve, the game play.

What is a game? I define a game as a subset of reality, with a set of rules, that interactively entertains a player or players by setting defined goals and gives them a possibility ability to achieve those goals by their actions. The entertainment comes from getting the “possibility” correct. Lets look at these elements one by one.

Intensity (Spectrum)
Intensity (Spectrum)


In board games its taken for granted that the world can be abstract. A chess board does not look realistic, and the players have limited movement and shape. The simplification of reality, or the removal of all the unnecessary elements allows a focus on what is left. Video games have tended to be more realistic over the years widening the subset of reality. They still are a subset, you can switch them off at any time and return to real life. Games range from a very restricted subset of reality, ie abstract games all the way to simulations. The gameplay tends to decrease with the increasing of the subset of reality, because the amount of non gameplay interactions increase exponentially. For instance in a huge world you can wander around for ages to find something of interest. A small restricted would focuses the player on the active gameplay. So designing a game becomes a question of deciding what to leave in and what to leave out. The trick is to leave enough in so the player has enough interest but remove the unnecessary.


Its the rules that formally define the subset of reality. Its surprising how few the rules can be for a good game. The earliest video games had very few rules. Once you have enough rules to maintain interest looking at the number of rules required is a guide to the strength of the game design. Strong designs need few rules. If a game needs loads of rules its generally means that the core idea does not work too well. It has to be patched with many rules to plug holes in the design. I am not saying you should only have a few rules. You can add interest by additional rules, but the core design should stand up with few rules.

Take a Doom type shoot ’em up for example;

  • You can move in two dimensions along the floor.
  • Basic physics of solidity and gravity exist, ie you follow the floor, and cannot go through walls.
  • Basic temporal and positional physics apply, ie things positions and times are maintained.
  • Limited interactions between things apply. Bullet versus person. Pickup versus person.Person versus person etc., they give you energy or take it from you.
  • When you have run out of energy the game is over.

Things like secret doors, pickups are additional rules to add interest.


Thats the whole point of a game. It can do this by amazing them with videos but it’s not a game unless these are a reward for the player doing something, whether it be a simple button press or a completion of a level. Interactive entertainment comes from responding to a players actions.


Setting the player a goal they think they can achieve is the real crux of gameplay. The better defined a goal the stronger the came play tends to be. The goal may be to stay alive as long as possible, to reach a high score, to get to the last screen. It may be a game has several goals. Usually it is better if the game does not have conflicting goals, that is more like real life. A game can have subgoals but the whole point is if the play the game well you must be able to get nearer whatever the ultimate goal is.

The goal should not be easily obtainable or impossible. Good gameplay lies in between and presents the player with a feeling that he could achieve the goal and rewards good play. The converse of rewards are penalties. Just as goals reward the player for playing the game well penalties punish the player. They need to be present but should be in balance with the rewards. A penalty may be the loss of energy for getting hit with a bullet, or loss of a life. The clearer the reason why a reward or penalty has occurred the more instant the game is. Obscurer rewards (or even penalties) provide additional interest and add to longevity of a game. The Rainbow Islands reward system is an example. If you collect gems in the correct order secret rooms appear.

The possibility of achieving goals is the prime factor a games designer has to consider. It does not have to be constant. Like the tension in a good movie it will build up in hardness to a sub peak ease off a bit build up to another peak , ease off , build to a mini climax, then all this will repeat until a major climax at the end of the level or game is reached. Take Space Invaders for example. As the player progresses through a wave of aliens they get nearer and faster. This increases the hardness of the game to a peak of excitement. Then the next wave comes. For a brief time its critical to take the end ones out but is easier than the hectic few moments of the previous level. If the player is skillful at the start they can relax a bit as the level goes through its middle phase. The pressure continually builds towards the end of the level. The overall cycle of each level starting with the invaders nearer the player ensures each climax is greater than the last.

Managing the difficulty in a game that has fewer restrictions is more difficult but can be done. In an 3D game like Doom the layout ensures there are real hotspots difficult to get past, followed by quieter areas. The players route through the game has to be considered. It is generally better to have areas of light and shade , difficulty wise. If meanies are distributed chaotically in groups it is better than an even distribution. Note this still can be achieved randomly . I have found random placement is often better than inexperienced placement by hand as humans tend to place things evenly which is less interesting. I have often used chaos maths in programs for added interest.


In the early days our games evolved in the following manner.

We investigated the platform and designed and wrote graphics architecture to exploit some feature of the machine.
We considered game environments that utilised that feature.
We create a player in a prototype level and worked on the control and movement.
We decided on the game idea and added enemies to the test level.
We worked on real levels constantly coming up with new ideas as we progressed.
We reviewed the levels , adding new ideas in where necessary.

Paradroid (C64)
Paradroid (C64)

Take Paradroid for example. Andrew started off investigating scrolling on the C64. He had a working all direction scroller from Gribbly’s Day Out. Using this for a top down game ensured the game would be different . As I was working on Avalon set in the past I suggested a futuristic scenario. He started off by programming the player movement , then player firing. He the added the enemy robots. At this stage the game idea was very much crystallised but a targeting system for the bullets did not work. After a few weeks of programming ever more complex intelligent targeting systems he ripped them all out and limited firing to 8 directions. He tried a gunsight but found the gunsight was often not in the right direction (ie if you wanted to rapidly fire behind you.) It just did not fit in with the necessary speed of play. The sub game when taking over other robots was added later although the idea of being able to use the robots was a core game idea.


Paradroid balances the gameplay with the use of both positive and negative feedback. Positive feedback is where the player becomes more powerful for doing something right ( a reward). It is useful for giving the payer a progression of ability. Positive feedback has been used in many Shoot’em’ups but suffers from the problem of exponential effect. Any positive feedback system allows the system to be more effective, that produces even more feedback. The result is a runaway system such as an population explosion, or chain reaction. Negative feedback has long been used by nature or designers as a controller. Negative feedback uses the output of a system to restrict the ability to produce the output . For instance when a population increases it needs more food. Thus a limited food supply acts as a negative feedback.

In Paradroid negative feedback is obtained as the player eliminates enemies. The enemies are also the player tools. This makes things get more tricky as robots are eliminated. The other balancing factor is the burning out of robots that the player takes over. The life of a host robot is inversely proportional to its strength. Thus powerful robots can only be controlled for short periods.