This page describes some of the fundamental aspects of route planning as applied to the making of speedruns (and tool-assisted speedruns in particular, although most of the knowledge applies to both phenomenons).
Why route planning?
No speedrun should ever be attempted without first having thought out a sort of route. Although it may seem quite persuasive to simply grab a rerecording emulator and use your new-found power of slowdowns and save states to make a recording of a game, it is not going to get you very far. At some point, you will realize that even excellent play can get very boring when one takes wrong directions and, as a result, actually needs to unnecessarily run around from place to place and backtrack to earlier levels just to be able to progress.
The very best speedruns owe much of their low completion time to the fact that its author(s) planned out their attempt very carefully, and have tried out numerous possible paths to see which ones are beneficial. If you check out the history of some of the videos on this site (by clicking on the link that leads to the obsolete, old version), you'll find that many of the major improvements are caused by route improvements (sometimes as a result of a new glitch being discovered).
What is a route?
In the planning of a speedrun, there are both short-term goals as well as long-term goals to satisfy. Route planning pertains to both of these goals.
An example of a short-term goal would be the obtaining of a new weapon that's hidden in a level somewhere. The player may have various ways to reach the weapon, such as by shooting all enemies that defend it, or perhaps by stealthily sneaking past them. When given this choice, it usually seems like a good idea to shoot the enemies, but one might have to carefully check both possibilities; aggression may set off a reaction of enemy behavior that will ultimately cost more time than the stealthy approach.
Long-term goals are more broadly defined, and are generally more difficult to predict. An example would be the obtaining of a certain weapon during a speedrun that makes it much easier to pass a certain level much later in the game. In this case, a long-term goal would be to make certain short-term routes possible, such as the one described above (if shooting the enemies is much faster than sneaking past them, one will first have to get a weapon).
Other goals may include:
- investing time in certain characters (choosing to level up a particular character in an RPG speedrun),
- choosing a certain play style from the start (an example from a non-tool-assisted speedrun would be Cacodemon Leader's Hexen speedrun, in which the Cleric class is used; the Fighter class is much faster at passing the first few levels, but near the end of the game, the Cleric's versatile weaponry becomes ideal),
- choosing an order of events to take advantage of items obtained during certain events (such as the Mega Man series of games, in which every completed stage gives the player a new weapon).
Failing to recognize the optimal solutions to these things will cause a slowdown in your run. If you, for example, don't invest the right amount of time into leveling up your RPG characters, you'll notice the effect later on when it becomes difficult and time-consuming to beat bosses. Not knowing where or when to pick up which required item during a run of a platform game will require you to move around a bit more until you get all that's required.
These things are most easily fixed (perhaps by other runners), too, so if you don't want your run to become obsolete quickly, you'll have to do some thinking beforehand.
Deciding on a route
Which route to take depends entirely on the game itself. There is no such thing as a basic route that will always lead to modest success. It's good to keep in mind that the game designers' intended route is most likely not the best.
One may first want to consider events that will have the biggest effect on the gameplay, and then move on to the more trifling things.
The most important thing would be to decide on things that affect the whole game, like which character to use in case multiple characters can be chosen.
To runners who are experienced in playing the game itself, some of these goals will be easy to discover, as they know from experience that certain ways to play the game will be better than others.
Asking the community for assistance
It's recommended to ask for advice on the making of a route, especially in the case of short-term goals. It's quite easy to miss opportunities that could shorten a run's length, and posting frequent updates of the run's progress will allow others to help spot such alternatives.
TASVideos has an active community of people who might be able to help you. Some of the best videos presented on this site have been analyzed completely on our discussion forum. Some of the more extreme examples include a 170 page Ocarina of Time topic and a 157 page Super Mario 64 topic.
 The page count for both topics is recent as of July 15, 2007.