How to Analyze Action in Film

Aim of Analysis

One of the benefits of being able to categorize action scenarios is the ability to isolate moments of action in a film with much greater precision. It is also important to share those methods of analysis so that the results are more understandable and can be replicated by others who are also interested in quantifying action in films. In this post, I will summarize the methods I have employed when analyzing action moments and how I go about quantifying them.

What Counts as Action?

The first step in measuring action moments in a film is the ability to identify those moments. In my book, Action Scenarios in Film: The Essential Guide to Action in Film, I identify those moments as seven historically recurrent scenarios, namely: rescue, escape, capture, heist, fight, pursuit, and speed. This means that any instance of these seven scenarios appearing in a film counts as an instance of action. Other action scenarios do exist, most notably the transfer – the occasion when a character transfers over from one means of transport to another normally at speed – as prominently used in The Fast and the Furious (2001) and Carter (2022). However, the transfer scenario is less prominent and recurrent in the history of action cinema.

An instance of action can manifest high typicality, meaning that it possesses a high number of features that define an action scenario prototype. Or the instance of action can manifest low typicality, in other words, possess only a few features of the prototype. In between these two poles is a range of instances moving from high to medium to low typicality, creating a typicality gradient for the action scenario in question. In my book, I detail the features of the prototypes that define the seven action scenarios through which a typicality gradient can be derived.

A few examples from films can illustrate this point. Consider the typicality gradient for the fight scenario. A fight scene in a film normally consists of two parties who engage in physical conflict. That conflict can be a one-on-one fight, or be depicted as opposing combat forces. Instances of the fight scenario that fit these two prototypes are the one-on-one fight between James Bond and Red Grant on the Orient Express in From Russia with Love (1963) and the beach landing sequence in Saving Private Ryan (1998) featuring multiple characters as the American forces combat the entrenched Germany army [Figure 1 and 2]. In both examples, there are minimally two opposing fight participants, be they individual or aggregate, who engage with each other with lethal force. Further down the typicality gradient of the fight scenario, one can place the tussle between Daryl Bates and Jed Eckert in Red Dawn (1984). There is a reciprocity of aggression as they shove, slug, and throw objects at each other but at no point does their fight approximate the physical risk that defines the fight prototypes [Figure 3].

Another medium example would be the unidirectional fight variant. Instead of the reciprocity of physical attacks between participants in a fight, individuals are attacked by other characters but no such reciprocity occurs. In such instances, the attack is unidirectional instead of bidirectional. An example of the unidirectional variant can be found in the opening scene of Face/Off (1997). FBI agent Sean Archer is with his son on a carousel. His nemesis Castor Troy aims his sniper rifle from a distant hilltop and fires, injuring Archer, as the bullet goes through him, but kills his son. As an instance of the unidirectional variant, the example shares the dimension of physical risk associated with the fight prototypes but lacks the reciprocity of attacks as Archer grieves over the death of his son instead of retaliating. At the bottom end of the typicality gradient of the fight scenario, one can place instances that neither possess reciprocity of physical attacks or high levels of physical risk but are marked by some form of aggression directed at another character. An example of such low-typicality occurs in Logan (2017) when Logan unidirectionally slaps a mug out of Caliban’s hand, smashing it to the ground [Figure 4]. All of these examples are moments of action regardless of their typicality given they are instances of the fight scenario.

Measuring Action Moments

After identifying a moment in a film that manifests action, the next task is to isolate that moment to measure it. The easiest way to do so is to use the edit, be it a cut or other form of editing, just before the action commences as the start point and the edit just after the action finishes as the endpoint. The length from the start point to the endpoint represents the duration of the action moment. An illustration of such measurement can be drawn from the prison escape sequence in San Quentin (1937) The action sequence begins with the shot in which Hansen, one of the escapees, reaches into his shirt to pull out a gun that will be later used to kidnap a guard [Figure 5]. That shot commences at 57:42 minutes. The sequence ends with the shot when Kennedy, the other escapee, climbs onto a train to elude police capture, with the edit appearing in the form of a dissolve [Figure 6]. That shot ends at 1:02:38 minutes from which the sequence’s duration can be derived as 4 minutes and 56 seconds.

For films that consist of long takes, or digitally seamed shots combined to appear as a one-shot long take, such as 1917 (2019) or Carter (2022), other measuring techniques must be used. In such cases, the start point is established when the action moment precisely commences, such as a character who starts to move their arm to deliver a blow, whereas the endpoint is fixed at the moment when the action ends, such as the terminus of a character falling to the ground after being punched. Like instances of action falling within a typicality gradient, the duration of an action moment has no bearing as to whether it is to be counted as an instance of action. An action moment can consist of a 3-second instant when a character accelerates on their motorcycle, as is the case in A Better Tomorrow (1986), or can be over 44 minutes long as is the final battle sequence that ends 13 Assassins (2010).

Action Combinations

The next task after one has isolated and measured a moment of action is to log the types of action scenarios that are manifested in them. As discussed in my book, action sequences rarely manifest one single type of action scenario. More often they appear in combination, either sequentially (horizontal combination) with other action scenarios and/or concurrently (vertical combination). The task then is not to simply note the different types of action scenarios manifested in a particular action sequence but to also observe how they combine together. An illustration of such action scenario combinations can be found in the second shop robbery in Bonnie and Clyde (1967). The sequence commences with the heist scenario as Clyde robs a shop by ordering the shopkeeper at gunpoint to add goods to his bag [Figure 7]. Suddenly Clyde is attacked by a butcher wielding a meat cleaver instantiating the fight scenario [Figure 8]. Since the robbery is still ongoing then the action combination is vertical in nature. After Clyde and the butcher scuffle in the shop, Clyde knocks him out and then hops on the running board of a vehicle as Bonnie drives off, instantiating the escape scenario [Figure 9]. At this point, the heist has yet to truly come to an end since Bonnie and Clyde are still at risk so the action combination is again vertical in nature. When Bonnie and Clyde are safely out of town limits the heist scenario ends but since they are traveling at speed the combination is horizontal in nature as the heist and concurrent escape scenarios segue into a speed scenario [Figure 10]. As an action combination form, the sequence exhibits the heist, fight, escape, and speed scenarios in both vertical and horizontal combinations.

However, more complex combinations can be found in action cinema. As discussed in the action profile for X-Men (2000), extended final act action sequences often feature action complexity, with one of the main features of such complexity is action scenario combination.

Action Moments

After identifying the action moments in a film, measuring their duration, and noting the nature of the action combinations they exhibit, the next task is to tally up their number and their duration. Total action moments in a film can vary significantly. Whereas San Quentin has only 10 action moments, The Guns of Navarone (1961) has a total of 32. In itself, the number of action moments is only a partial measure of the extent of action in a film. More important in that regard is their duration. Carter, for instance, has only 14 action moments but five of these are over 10 minutes long, with the final act sequence over 25 minutes in duration.

In addition to numbering action moments in a film, their duration can be used to calculate Action to Story Duration ratios and Action to Act Duration ratios, topics I discuss in this website’s Introduction. By quantifying action in film in the ways summarized one can derive a set of useful data that offers a rich and unique approach to understanding action cinema.