Ball not in play

Ray Williams’ book Rugby for Beginners was published in 1973. I first read it as a postgraduate physical education student at Loughborough College. By coincidence, Ray had been a student there too. Both of us were from North Wales.

Years later when I got to know Ray, I was able to explain how important the book was to me in my development as a player, teacher and coach.

The cover of Ray Williams' book.

Huw Richards (link) wrote about Ray’s career and noted his appointment as the Welsh Rugby Union’s first national coaching advisor. In that role he “drove that transformation through his promotion of conferences, teach-ins and courses which gave Wales more than 300 qualified coaches by the mid-1970s”.

I was fortunate to be one of those coaches and delighted in late night conversations with Ray in the bar of the National Sports Centre at Sofia Gardens in Cardiff. It was like being with the Oracle at Delphi.

One conversation became quite heated. I asked Ray about a line in his book that suggested “no player has the ball in his hands for more than one minute” in a game. The essence of Ray’s argument was that each player had a responsibility to support the ball (one of Ray’s game principles).

Even at my time playing rugby at Loughborough, I was sure I did not have the ball in my hands for that amount of time. I suggested to Ray that I ought to investigate what time the ball was in play and not in play.

I did follow up on this for the much of the next two decades. My operational definition of ball in play time was when the game was started and restarted by the referee either by a whistle or when the play was put back into play. Ball out of play was measured by a referee’s whistle or when the ball visibly left the field of play or was waiting the restart of the game.

It took some time to stabilise the recording of ball in play time. I monitored ball in play time from live broadcasts. One of my first successful attempts was on 16 January 1982, in what was then the Five Nations rugby tournament. Scotland played England at Murrayfield in a game refereed by Ken Rowlands (Wales).

  • The first half game time was 42 minutes and 33 seconds. The ball was in play for 10 minutes 28 seconds.
  • The second half game time was 44 minutes. The ball was in play for 13 minutes 10 seconds.
  • In the whole game, the ball in play time was approximately 27% of the available time.

It took me a further three years to develop a template to record each passage of ball in play in real time in addition to the other data I was collecting with hand notation. From this time on I termed passages of ball in play activity cycles.

My record of the Scotland v Wales game played on 2 March 1985 (video link) was:

For the first time, I was able to have a detailed account of game play. I recorded 97 distinct activity cycles (49 first half, 48 second half). Scotland had 52 of these (25 first half, 27 second half) and Wales 45 (24 first half, 21 second half). The game was refereed by Rene Hourquet of France. Wales won 25 points to 21 points.

The activity cycles were:

My record of the 97 activity cycles indicates a total ball in play time of 25 minutes 46 seconds (12 minutes 01 seconds first half, 13 minutes 45 seconds second half). Scotland had 13 minutes 20 seconds of ball possession and Wales 12 minutes 26 seconds.

I shared these data with Ray and we corresponded about the implications of such data for coaching and playing. I continued to share my data with him and he in turn passed it on to colleagues in coaching.

I have returned to these data this week as I researched the concept of dwell time (link). I was delighted to discover that Herbert Levinson (1983) was undertaking similar real-observations of performance … in the context of transit travel times. He concluded “transit performance should be improved by keeping the number of stopping places to a minimum”. That sounds like a fascinating pedagogical insight for rugby union.

A rugby union hand notation archive

I am using Google Sites to curate an archive of two decades (1980-2000) of my real-time hand notation of rugby union football.

I am hoping that in sharing these notations with a Creative Commons Attribution 4.0 International licence (CC BY 4.0) they may be of use in secondary data analysis. The Google Sites provide links to my Google Sheets for the data collected.

I aim to add to the site whenever I have an opportunity. At present there are two pages that share data:

  • 1980 (my first attempt at real-time notation)
  • 1992

My hand notation journey started in order to inform and transform my coaching. I tried to see what coaches might see and thought my notations might help me have conversations with other coaches.

I was mindful that coaches observed games in real-time. At the time I started my notation, video recording was still quite novel. I chose not to use video in the early days of developing my systems.

I spent a lot of time practising the notations whilst watching live broadcasts of games. Gradually, I think my intra-observer reliability improved substantially. In the 1980s, I did not conduct any inter-observer reliability studies.

I do think that any errors in my data might be minor and within the accepted limits of reliability. I see these early notations rather like a dance notation. They are records of performance that make the performances more permanent and amenable to conversation.

This was particularly the case in 1982 when I started to indicate ball in play time as distinct from total game time. And in 1985 when I started to look at activity cycles.

In the 1990s, I was fortunate to work with the Welsh Rugby Union. My data in that decade combined real-time hand notation to support coaches within games and lapsed-time forensic analysis of the world game.

Performance Against Ranking: Rugby

I have been looking at performance against previous year’s ranking in rugby union and rugby league.

In the charts that follow, the green (win) and blue (lose) boxes indicate that teams conform with their previous year’s ranking, gold (win) and red (lose) indicate unexpected result based on the previous year.

In all examples shared, visually the expectation is that the bottom half of the table is predominantly blue and the top half green. It is interesting note how different the Top 14 Orange matrix is. Lower ranked teams win at home on a regular basis.

Super 15 Rugby Union (after 4 weeks)


Aviva Premiership Rugby Union (17 Rounds)


Greene King Championship Rugby Union (17 Rounds)


Top 14 Orange Rugby Union (19 Rounds)

Top14 19

Six Nations Rugby Union (3 Rounds)


NRL Rugby League (1 Round)