An Introduction to Analytic Narratives for Coaches and Students

A photograph of Aboriginal Whalers at Eden, NSW.

Background

I received an alert to a paper today that has sent me off to revisit Donald Polkinghorne‘s and Philippe Mongin‘s discussion of narrative and the process of historical analysis … and to contemplate pedagogy.

The paper that started my journey today is titled ‘The Cooperation of Humans and Killer Whales (Orcinus orca): The Application of a Simple Fuzzy Rule-Based Model to a Historical System‘. The authors of the paper are Emery Coppola, Ryan Jones, Jack Owens and Ferenc Szidarovszky.

They present:

an historical model application that is pedagogical in nature, in that it presents the methodology for constructing a simple fuzzy model for a vague  but complicated social cooperative network along with example model-simulation results.

Their paper has an immediate empirical appeal for me as they discuss activities in a geographical area four hours to the south of my home in New South Wales.

Once I was hooked by the accident of geography, I became intrigued by their approach to bring together fragmentary data sources to create a model.

I believe their paper, and its connections with others interested in ‘narrative knowing’ and ‘analytic narratives’, raises important issues for the discussion of sport analytics.

A Narrative of Cooperation

A picture of a whale hunt at Eden, NSW.

In their paper, Emery, Ryan, Jack and Ferenc study “a complicated social cooperative network in Twofold Bay, southeastern Australia, over a century until 1930″. They note “Surviving sources document that pods of killer whales or orcas worked cooperatively with human bay whalers” to pursue and kill baleen whales.

Jack has been involved in the field of geographically-integrated history since the late 1960s. The Twofold Bay research provides an excellent opportunity to pursue this kind of history project.

The research has to deal with some fundamental issues about data. “The whalers are long dead, and there is no systematic collection of records from which we can draw”. There are subsequent studies of “orca behavior in different regions of the world done over the past 40 years” and there is “significant research on Australia’s Aboriginal peoples”.

Despite these constraints, their goal was to develop a fuzzy rule-based model that predicts the likelihood of the success of the social network in killing a whale”. In this case study “prediction means that the model will simulate the outcome of a whale hunt for each event in our narrative”.

They share the process of developing their fuzzy rule-based model and report:Our attempt to represent a complicated social network with a simple rule structure falls far short of plausibility. At the same time, our initial efforts, however modest, compel the historian/modeler to formulate a set of linguistic rules that quantify often highly vague variables and conditions, qualitative and/or quantitative in nature, in an attempt to represent and simulate a complicated system of interest.

I liked their exploration of the pedagogical issues in their research. I liked too their reflection on their practice:

As we learn from our initial models and accumulate more data, information, and understanding, we can formulate and test new models against the surviving record, allowing us to consider alternative hypotheses and to see more clearly what additional information we need to acquire … in an attempt to explain better this fascinating cooperation between orcas and human whalers for at least a century to hunt successfully large baleen whales. (My emphasis.)

As I read the concluding paragraphs in their paper, I was struck by the generic issues Emery, Ryan, Jack and Ferenc raise:

  • Imperfect information
  • Flawed understanding of processes that are often so complicated that no model will ever accurately capture the underlying dynamics
  • Narrative sharing
  • Acceptable prediction accuracy
  • Models as a first step to forming a theoretical or heuristic framework for analysis
  • Refining and improving understanding through additional data collection, model development, and testing.

These issues are central to a scholarship that embraces “new forms of research organization and rapidly evolving types of information management and analysis” (Owens, 2010).   They connected me with Donald Polkinghorne and Phillipe Mongin.

Sharing Stories With Practitioners

A picture of the fins of two orca whales

After reading Emery, Ryan, Jack and Ferenc’s paper, I thought about how I might share with coaches some of the take-aways from their “flawed understanding”.

I wondered too how I might share the pursuit of heuristic frameworks with students as they develop their understanding of analytics.   Donald Polkinghorne’s (1988) exploration of narrative knowing places significant emphasis on the importance of “having research strategies that can work with the narratives people use to understand”. 

Most of the coaches with whom I work are able to locate themselves within an historical context in sport, in terms of the sport in general and in terms of their own career paths.

I think they would find the story of the orcas as fascinating as I do.

My hope is that this might lead to conversations about understanding and transforming performance. I think I might be very selective about what I share and would gloss over the fuzzy-logic part of the story.   I think the orcas would be a great lead in for students too but the context of our conversations would enable me to explore what constitutes fuzzy logic and its potential to model behaviour.

With both groups, coaches and students, I would be mindful of Donald’s observation:

History’s function is to describe the events of the real world as they have actually happened and to explain why they have happened. … Historical narrative is supposed to be factual – that is, it is supposed to be made up of true sentences that represent actual past events. The sentences of historical discourse are expected to pass a correspondence test based on the evidence of the traces of events left in documents. (1988:57) (My emphasis.)

I take our ability to develop actionable insights to be informed by a rigour in how we collect and analyse data that can be fragmented and partial as well as comprehensive.

Could He Have Won?

Farmland in Belgium that was the site of the Battle of Waterloo

I enjoy returning to Philippe Mongin’s 2009 paper, A Game-Theoretic Analysis of the Waterloo Campaign and Some Comments on the Analytic Narrative Project.

In the paper Philippe presents a game-theoretic model of Napoleon’s last campaign, which ended dramatically on 18 June 1815 at Waterloo. It looks in particular at the decision Napoleon made “on 17 June 1815 to detach part of his army against the Prussians he had defeated, though not destroyed, on 16 June at Ligny”.

In his discussion of events in the Waterloo Campaign, Philippe observes:

At three key moments – June 17, around mid-day on June 18, and in the final hours of this same day – Napoleon could have departed from the line of events that his previous decisions had set in motion, and he did not (2009:15).

Philippe is able to include much more detailed data in his analytic narrative compared to the orca paper. His discussion of the process of constructing an analytic narrative provides an explicit opportunity to explore how history might have been redefined and to think critically about ‘the culture of the unique’.

In 2016, Philippe revisited the process of constructing an analytic narrative. He notes that “the transformations that standard narratives incur to become analytic narratives bears some relation to the transformations they incur to become computational narratives” (13:11).

I take the essence of this tranformation to be the understanding that “analytic narratives are narrative texts, which include, among their parts in non-narrative form, the statements of formal models and their consequence” (13:9).

Philippe used this approach in his study of Waterloo, Emery, Ryan, Jack and Ferenc did too in their use of models within a case study with much less documented evidence.

Narratives and Audiences

A picture of Mongolian wrestlers and their coaches.

My aim in discussing analytic narratives is to open conversations about evidence and models.

It is an attempt to extend the epistemic reach of sport analytics in the connections we make with coaches and students.

I am attracted to the qualitative nature of analytic narratives but am mindful that they provide an excellent platform for engagement with quantitative models. Emery, Ryan, Jack and Ferenc used fuzzy logic with fragmented historical accounts; Philippe used game-theoretic tools with an extensive textual record.

I am hopeful that the epistemic reach of sport analytics can be enriched by a pedagogical leap too. Jack Owens (2010), in his work on a Masters course at  Idaho State University, developed a capstone internship experience that allowed tutors to ‘coach’ students “in ways they can interact more effectively with others”.

As Donald suggests, narrative will be at the heart of vibrant interaction with practitioners. Imagine where a story that starts “Did I ever tell you about Old Tom?” or “How could you snatch defeat from the jaws of victory?” might lead us.

Photo Credits

The Aboriginal whalers of Eden (ABC South East NSW)

Return of the killer whales of Eden (Australian Geographic)

Orcas (Ed Dunens, CC BY 2.0)

Waterloo, Belgium (cjlvp, CC BY-NC-ND 2.0)

P1140782 (WhatsAllThisThen, CC BY-NC-ND 2.0)

References

Coppola, E., Jones, R., Owens, J. & Szidarovszky, F. (2015). The Cooperation of Humans and Killer Whales (Orcinus orca): The Application of a Simple Fuzzy Rule-Based Model to a Historical System. NOAH LLC and the Geographically-Integrated History Lab (ISU).
Mongin, P. (2016). What Are Analytic Narratives? Proceedings 7th Workshop on Computational Models of Narrative. B. Miller et al. (eds), pp. 13:1–13:13. Dagstuhl.
Mongin, P. (2009). A Game-Theoretic Analysis of the Waterloo Campaign and Some Comments on the Analytic Narrative Project. Paris: Groupe HEC.
Owens, J. B. (2010). Graduate Education in Geographically-Integrated History: A Personal Account. Ann Arbor, MI: MPublishing, University of Michigan Library.
Polkinghorne, D. E. (1988). Narrative knowing and the human sciences. New York: State University of New York Press.

Entangled narratives: sport performance analysis and sport performance analytics

Introduction

This post explores the stories about data we share in sport performance analysis and sport performance analytics.

It is an attempt to think aloud about the narratives we construct when we share our observations about performance in training and in competition.

The trigger for this post is Emma Uprichard and Leila Dawney‘s (2016) discussion of data diffraction.

I am distinguishing between analysis and analytics as occupations but increasingly these are becoming entangled in sport settings … particularly when performance analysts extend their learning to include data science skills.

This continuing learning is enhanced when we ask second order questions about our practice. Such questions help us clarify the why, what and how of our performance data responsibilities.

Diffraction

Emma and Leila consider how we respond when the data we collect, through a variety of mixed methods, diffract rather than become integrated.

Diffraction splinters and interrupts performance data … “it provides an explicit way of empirically capturing the mess and complexity” that is intrinsic to “the social entity being studied” (2016:1).

Emma and Leila share their understanding of diffraction as “a process of paying attention to the ways in which process produces ‘cuts’ that can interrupt and splinter the object of study” (2016:2).

They point out:

To be clear, in arguing for diffraction as an alternative to integration, we not wish to negate or undo the efforts colleagues have made regarding data integration. Rather, we emphasize the need for an approach that explicitly supports instances where data do not integrate or ‘cohere’ and argue that this may be due to the messy nature of the object of study. In doing so, we provoke a discussion around the orthodoxy of integration as a goal of mixed methods research (2016:3).

They add:

an implicit assumption to integration is that the empirical data will depict a particular social phenomenon. Yet there is no a priori reason that this is necessarily so. Mixed data could equally reveal multiple phenomena that are entangled together, even though they appear to singular or whole. Mixed data could instead multiply the partiality, increase the uncertainty and further entangle the subject (2016:7).

In addition to adding ‘diffraction’ to my thought horizon, Emma and Leila have pushed me to think about ‘cuts’ and ‘entanglement’.

Cuts

In their discussion of the process of observation, Emma and Leila note “Different methods may produce very different cuts, but the same method may well do too” (2016:11). They add “These cuts produce different ‘matterings’; they make some aspects visible but not others and this process has social effects” (2016:11). 

They continue:

the process of doing research, of making cuts, will always be partial and will always bear the traces of the research process undertaken. Moreover, the act of making these cuts contributes to how we understand what it is we think we are researching (2016:12).

Cuts are boundary-drawing processes that, through what they reveal or conceal, come to matter (2016:13).

Entanglement

One of the important discussions in Emma and Leila’s paper is related to entanglement. Their argument draws upon earlier work by Karen Barad . (“Existence is not an individual affair. Individuals do not preexist their interactions; rather individuals emerge through and as part of their entangled intra-relating” (2007:ix.)

Emma and Leila propose that “observers and phenomena are always entangled” (2016:12). Sensitivity to methodological cuts opens us up to what is “messy, fuzzy and multiple in the social world” (2016:14).

An openness to diffraction encourages awareness of difference and entanglement. They argue that the process of collecting, analysing and sharing performance data can make visible “its own interference and its various material effects” (2016:14).

The Stories We Share

I am hopeful that awareness of Emma and Leila’s paper will encourage us to think more about the stories we share and how we share them.

Their discussion of diffraction, cuts and entanglement is an important counter to uncritical acceptance of data capture and analysis.

Their work took me back to Donald Polkinghorne’s (1988:11) contemplation of narrative knowing and his suggestion that “Narrative is a scheme by means of which human beings give meaning to their experience of temporality and personal actions”.

If we accept that our methodological cuts offer a partial perspective on performance,  I wonder how we address uncertainty in our stories we share with coaches and players.

Donald again …

Narrative is a meaning structure that organizes events and human actions into a whole, thereby attributing significance to individual actions and events according to their effect on the whole. (1988:18)

Through experience, we learn to craft the stories we share. Gradually, we come to appreciate the difference between headlines and granular detail.

My interest in Emma and Leila’s work speaks to my fascination with personal differences in sport contexts.

During my time of sharing performance stories with coaches and athletes, I have tried to balance the spoken and written with the unspoken and unwritten.

Some of my most memorable conversations with coaches have started with a coach’s question “Is there something you are not telling me?”.

I see these as wormhole moments (bridges through time-space) … the question opens up opportunities to go wherever the coach wishes to go and where our intra-relating might take us. Prompted by absence in the story rather than presence.

This is the start in our professional learning journey that moves us from a chronicle to a story.

Making sense of our entanglement in this process and it products offers us the possibility of learning how to welcome diffraction.

I think this sense making is energy giving and invites us to have a repertoire of stories for diverse audiences.

Photo Credits

Reading (jwyg, CC BY-SA 2.0)

Emma Uprichard (Twitter)

Leila Dawney (University of Brighton)

Osheaga 2011 (Tony Felgueiras, CC BY-NC-ND 2.0)

Augsburg (Keith Lyons, CC BY 4.0)

Postscript

Jamie Coles mentions two images in his comment below. They are contained in this tweet:

Conversations about informatics and performance analysis

I noticed a tweet by Rob Carroll earlier this week

Rob’s post raises a profound question about generational change.

One of my teachable moments was when I read Donald Polkinghorne’s (1988) Narrative Knowing and the Human Sciences. I tried out some of his ideas at Porto in 1998.

I think this slide resonates with Rob’s argument:

fr3

By coincidence, this week, I have been reading Arthur Samuel’s (1953) paper that explained computers to a lay audience.

It has this preface:

samuel01

Arthur concludes his paper “Computers are here to stay, and it is high time for us to be learning more about them”.

I take Rob’s point about the forms of representation we can (and should) use. I imagine there is a point in all relationships when we can extend proximal development. This opportunity is at the heart of learning organisations.

As I was reflecting on Rob, Donald and Arthur, I happened upon Nathan Kinch’s post with this diagram connecting value, meaning and engagement:

nk1

Sometimes we achieve this in our communication. I am becoming more and more interested in story sharing as a way to make this possible … which is a fascinating return to Donald via an early morning tweet.