There have been many attempts to predict the outbreak of war.

If one tries to predict war, one should first decide on what time-scale one tries to predict war, to decide what kind of tools one should use.

The tools can be divided roughly by the time-scale applied: Weeks, Years, Decades or longer.

• Weeks

On a short timescale, one can often predict wars simply by being in the country of the aggressor, and keeping ones eyes and ears open. Weeks before the attack by Argentina on the Falkland Islands, it was an open and popular threat in the news.

In the short term, a political journalist that is fluent in the local language and customs, based in the country of interest may be the best alarm clock.

• Years

The Sipri Yearbook, among others, display information on troops all over the world. By simply noticing the changes in the buildup and composition of the Iraqi Armed Forces into mountain troops and panzer troops, and noticing that these would be almost useless in any defence of Iraq, almost only useful for a fast attack war against any neighbours with mountains and/or large deserts, like Iran and Saudi Arabia.

Sipri and others list interesting data such as expenditures on the military, types of military: This is interesting since certain types of military equipment or training are a lot more useful in attack than in defence, and some types may even only be useful in the territory of a neighbouring country.

One can also find lists of disputed areas that are disputed over natural resources or simply because the population happens to be of another nationaliy, as in the case of Turks in Cyprus.

• Decades

Lewis Fry Richardson made an arms race model, trying to "predict" World War I, where he showed how two countries would go to war if more money was spent in the arms race than in trade. There is an interesting economic idea here, but since the model was made after the war, it did not have the prophetic power it might have had. The model was very accurate, but the parameters may well have been tuned to fit the data, since it was after the fact.

Econometric models.

A number of small and large scale simulation econometric models have been tested after Richardsons model, to try and predict when war might come.

Statistics.

In the Correlates of War project, and other projects, very few realiable basic correlates of war were found, although a number of interesting ideas seem to have had some some basis in fact. The only strong result was support for Democratic peace theory.

These models tend to focus on what how particular countries work in theory. An implicit idea seems to be that there may be some fundamental flaws in certain countries, and that one tries to unconver these.

• Very long term

Themes regarding the world systems The Democratic peace theory is one idea about how a political system, national and/or global should look like to avoid war. Other long term ideas have mainly suggested that whatever nations look like, they must do away with the need for expanding for world peace to ensue.

These ideas tend to focus on how the entire world system ought to work, which world systems that would eventually lead to a new world war, and which do not.