Wales V New Zealand - An experiment I ran

I have been designing, as part of my PHD, some software that analyses Sports performance in an intuitive and more specfic way than previous sports performance analysis programs before. It's early on in my test, but I fed the Wales - New Zealand game into it and got data for the Welsh team. The programme recognises simple things such as tackles made and missed, but also the previously undefined areas, like when a ruck was saturated and number of carries over the gain line. Each action such as a knock on or a line break is awarded a negative or a positive score of different values.

Here's some things I pulled out statistically (All data is relative to minutes recorded):

Wales' top 5 performers:

1. Sam Warburton
2. Alex Cuthbert
3. Toby Faletau
4. Scott Williams
5. Aaron Shingler

Wales 3 worst performers:

1. Rhys Priestland
2. Leigh Halfpenny
3. Gethin Jenkins

Most Tackles missed: Jonathan Davies

Most Tackles made: Aaron Shingler

Player beating the gain line the most: Toby Faletau

Most effecient at the breakdown: Sam Warburton

Least effecient at the breakdown: Scott Andrews

Most Penalties conceded: Liam Williams

Most Turnovers conceded: Rhys Priestland

Most tackles for a back: Scott Williams

Data based on what? what is TOP and WORST in relation to?

Here's some things I pulled out statistically (All data is relative to minutes recorded):

"top 5 performers:

1. Sam Warburton
2. Alex Cuthbert
3. Toby Faletau
4. Scott Williams
5. Aaron Shingler

Wales 3 worst performers:

1. Rhys Priestland
2. Leigh Halfpenny
3. Gethin Jenkins"

Interesting Lyn, but sadly regarding analysis is that until a programme is designed to rate the importance of each individual scenario, take into account all variables in every situation, and technical and tactical issues accounted for all of these stats are null in void.

For example regarding Warburton, and Cuthbert being Wales best performers, it's near total BS as Cuthbert missed numerous tackles, looked lost at times and had a very soft underbelly. The fact that he made lots of yards in a few breaks in the 2nd half when Wales were on top doesn't make his performance any good.

Also most tackles missed Jon Davies, I'd be missing tackles when trying to cover for Cuthberts wing too!!!

Don't get me wrong I love analysis software and use certain software all the time, but only in specific and isolated incidents, throughout an entire game it tends highlight trends and not actual performances.

What is different from a software perspective that I couldnt do with a pen and paper and a slo mo recording of the game and banging it into Excel to get those results?

Or is that what the software does? Chops and dices manually (significantly I would expect) gathered data into meaningful information.

Anyway, in terms of refining it into more useful information you could drill down further to get more out of it to get 'data about the data' so to speak.

For instance- the tackle count could be refined to measure the impact or effect of the tackle- ie a tackle that stopped a try- such as Daggs for example might be worth more than a tackle made where there were 3 others waiting anyway.

Missed tackles the same and perhaps the % a player misses. I'd rather have a player miss 3 out of 10 than one who misses none out of none.

You could do the same with all the other areas- metres gained- what context were they? the run that clinched the try etc

Significant more data is required but as bluesman says, trends only are derived from the high level stats. The real information (as we know) is in the detail and not only the quantity but also quality of the data being gathered.

Interesting all the same and good luck with the thesis- sounds like fun.

Boy, am I glad I don't watch rugby with a computer in front of me. It would really annoy me to discover the game I thought I was watching was statistically totally different to the way it actually was.

Maybe that's why all the modern coaches look miserable when they're up in the stands with their AppleMacs open in front of them!

Here's the only data Welsh fans need to be cognisant of... lose against Oz and ye are in the 3rd tier.

I guess the advantage of this is that it's live (is it?!) so that coaches can get live data and make quick decisions. Pen and paper and video would mean after game analysis. I guess it's hard for coaches to keep an eye on everyone during the game, so the analyst's job is to relay real-time info and messages to the coach. If someone's having a stinker, statistically, then maybe the coach can change them to tighten up their channel, for example.

However, as pointed out, what it doesn't do at the moment is account for the weight of the scenario. Missing one tackle in your own 22 at the end of the world cup final is worse (IMO) than missing 3 tackles in the first half in the opposition 22 (probably). Statistically the 2nd chap would look worse.

However, the game is moving on so fast that these sort of things are really the way forward. The extra 1% that the coaches love to talk about. Embrace it or get left behind I think is the feeling. If you're not using it but the Kiwis are then they're getting a slight performance advantage, no matter how small. So everyone does it. Some of the stuff that is coming out is mind blowing. I know Under Armour have got training jerseys (and soon to be playing jerseys I believe) with transmitters in them to tell you player's heart rate, breathing rate, distance covered, accelerometer to assess change of direction off both sides, etc., etc. Live info is fed back to the coach so that he can make decisions during the game rather than scratch his head and wonder why his star centre is struggling (e.g. lack of sleep, illness, etc. leading to higher Heart rate, lower energy levels, quicker to fatigue, etc.).

Modern, see!

Yeah, but is it rugby Griff?

Not sure you'd need sophisticated software to know that Rhys Priestland was the worst player on the pitch??? (Bless him, not his fault etc).

This kind of stuff is, of course, mumbo jumbo. If players can't tackle, pass and kick straight then the "extra 1%" is just putting fancy icing on an unbaked cake.

Having said that, plenty of people make a living at it these days, so good luck to you and hope the PhD goes well.

Not sure you'd need sophisticated software to know that Rhys Priestland was the worst player on the pitch??? (Bless him, not his fault etc).

Thanks for the kind messages of encouragement. I appreciate I hadn’t explained the formula or criteria for analysis very well previously, but it has been interesting to see what rugby fans thought of Sports Analysis Software.

In answer to Taylorman, the formula for the software is much the same as you described but easier and about two hours faster!

The challenge of course is to make the program as intuitive and rugby aware as possible. There is lots of software around at the moment, and a lot of it is too general. Statistics such as body weight lost and km ran as well as tackles made and goal kicks completed, don’t give the depth of analysis that some of you guys have hinted at. The challenge for me is to make this software, understand things like what is an efficient ruck contribution, positional errors (still having difficulty) and the importance of a try saving tackle and the difference between that and a tackle.

I have already made progress. Some examples of how the program has become more rugby intelligent are: Recognising a try saving tackle, spotting overlaps and judging if executed or misused and also offload success.

The software basically divides actions up:

Line break

Occasions over the gainline

Tackles made

Line out won with throw

Line out won without throw etc

The next step is to preset – and + points ratings to an action. This then gives a player a rating for the match based over all the different categories. As the points ratings are preset by the user or coach, they retain an element of control. So if a red card was set to deduct -18 and the offence occurred in the last minute, then the coach can retrospectively change the value of one or more category.

What I am going to do, is compile the ratings from every Wales game in 2012, and try and compile a list of player rankings. This will give me an in-depth analysis of their contributions. I am already finding that certain players are performing differently to how I perceived them, before using such analysis.

Sounds cool. I know Wayne Smith developed over several years a very sophisticated data capture and analysis system that goes far beyond anything we ourselves get given privy to.

We kept hearing comments from GH like 'Smithys finished his analysis on Wales and will have England finished middle of next week' sort of thing.

What I am going to do, is compile the ratings from every Wales game in 2012, and try and compile a list of player rankings. This will give me an in-depth analysis of their contributions. I am already finding that certain players are performing differently to how I perceived them, before using such analysis.

Again this is dangerous as players miss games regularly, so wether your counting numbers of X or % of X it'll be distorted by number of games played, fitness condition, psychological condition etc...

I think New Zealand could do their analysis on an old Gameboy most of the time, and stiill have no worries for the match day!

thebluesmancometh wrote:What I am going to do, is compile the ratings from every Wales game in 2012, and try and compile a list of player rankings. This will give me an in-depth analysis of their contributions. I am already finding that certain players are performing differently to how I perceived them, before using such analysis.

Again this is dangerous as players miss games regularly, so wether your counting numbers of X or % of X it'll be distorted by number of games played, fitness condition, psychological condition etc...

The software should calculate efficiency in terms of minutes played and mathematically determine averages. The condition of a player is a variable that I doubt I can account for. That's a great point. A player could soldier on with a bad calf and not contribute to their usual level. This is certainly a flaw in statistical analysis.

You would have to make assumptions. One would be that a player is 100% fit on the field all the time. If they arent, they themselves, the team and the coach presumably know this but regardless its their 'contribution' to the match that is being monitored.

If they soldier on injured, then just like a bad selection, its a choice that is likely to affect performance. Any 'poorer' result due to the injury counts in the same way as anyone else uninjured. And so it should. Its about measuring performance, not ignorance or stubborn-ness.

RuggerBoy wrote:Yeah, but is it rugby Griff?

Eh?! Course it is. It's called progress. They don't play with rugby balls made of pig skin anymore, or wear hob nailed boots, or jerseys made of hessian sack either. All that stopped in the 80's, bit it didn't stop it being rugby.

Sport has always evolved and moved with the times, as does life in general. Advances in sport science have made the game the way it is today, whether good or bad, but it's still rugby.

Taylorman wrote:You would have to make assumptions. One would be that a player is 100% fit on the field all the time. If they arent, they themselves, the team and the coach presumably know this but regardless its their 'contribution' to the match that is being monitored.

If they soldier on injured, then just like a bad selection, its a choice that is likely to affect performance. Any 'poorer' result due to the injury counts in the same way as anyone else uninjured. And so it should. Its about measuring performance, not ignorance or stubborn-ness.

Yeah-case in point-Radike Samo's try of the year last year. Thomson was hurt and never should have carried on. Samo palmed him off like a twelve year old lad in his third game of footy.

But then Mr McCaw played the last 3 tests of the World cup with an injured foot- didnt even train in between matches. And still played blinders. GH recalled how he was horrified when he saw his foot after the matches...

Some people are just destined for greatness ay? Can take anything in pursuit of a cause...