Misleading Stats and Lazy Journalism.

Sometimes stats are helpful. Like the statistic on the scoreboard. That's about it.

For instance. NZ are poised to become the most successful test nation ever, by eclipsing France's record tally of victories overall. But given that France have wedged in nearly 30% more test matches than NZ somehow, this is an utterly futile record. Right up there with the worlds record test try scorer.

Another irritation of mine is the way Sam Cane is being talked up in NZ media right for having a high statistic for "arrivals at rucks". No doubt the guy is athletic and gets around the park, but more important is what he does when he gets there. I don't care if he hits 70 rucks a game if he's utterly effectless when he gets there, or gives away penalties. Not saying the guy can't do his job, just that this stat is meaningless.

The next one is winning streaks. SA have 9, NZ have 6 on the trot. So what? one team will lose and what bearing on the result will the previous games have? None.

Percentage successful kicks. Again, says nothing about the record of a kicker against a given opponent, or take into account the elements of altitude or weather involved.

These things get bandied around to no end in the media, and I think they smack of lazy journalism or a writer without the ability to express a coherent opinion based on his or her own knowledge and ability to read the game or analyse the factors that are likely to be pertinent.

For example, I put it to you that NZ are already the most succesful test nation ever. And one more victory, or not, isn't going to change that.

There have been loads of threads on this. Can we merge this to one of them?

Thanks.

GE, you used stats just the other day to prove Morne steyn is useless against the all Blacks?

Exactly Biltong. But I was just trying to irritate you.

Well your winning percentages are also skewed too GE.

In the old days SA and NZ would only play each other on average once every 2 years... with 4-5 year gaps (4-5 tests per tour year) often and allowing for the sporting ban of the 80s (when SA were very very strong) and the world wars etc.

Teams could be dominant for a decade but only rake up say 6-7 wins. These days you can rake up those number of wins in 3 seasons.

Prior to the boks re-introduction the boks won 54% of games vs. NZ over 70 years. NZ had won only 40%.... but had only played 37 matches. In 20 years subsequently they have already played more matches (48).

Lets say the 24 games every decade since their first game in the 1920s was in place then rather than the one every 2 years etc... we would have seen 168 matches rather than the actual 37 and using the same win rates the boks would have won 91 matches vs. NZ 67.

Even if we piled on the actual results from the last 20 years (those 48 matches since 1992) onto that which is 33 to 14 to NZ the number of games won would still be in SA's favour overall.... 105 wins to the Boks vs. 100 wins to NZ.

so as you said... statistics can be misleading.

Maybe the journalists are out there to irritate us.

GloriousEmpire wrote:Exactly Biltong. But I was just trying to irritate you.

so you admit it. you are a WUM artist

GE are you a journalist???

Mate! Real stats are bad enough, without making them up by unjustifiable linear extrapolations.

GoodinTightSpaces wrote:

GloriousEmpire wrote:Exactly Biltong. But I was just trying to irritate you.

so you admit it. you are a WUM artist

Just practising my Sun Tzu on engaging the enemy on all fronts It makes victory all the sweeter for BT if he can watch Morne potting an audacious 66 meter penalty to take the win and he can also think of how wrong I was.

I wasn't making up stats, it was an example to show that your figures are biased towards the present due to the number of games being played in the last 20 years superceding those played in the previous 70. It skews the data towards the most recent and isn't a fair reflection of total historical competition between the two nations.... or any other nations in fact.

Thats why, anyone who has been significantly better over the last 20 years compared to their history will seem better then they actually are.

Did referees in the 'old' days influence winning percentages?

fa0019 wrote:I wasn't making up stats, it was an example to show that your figures are biased towards the present due to the number of games being played in the last 20 years superceding those played in the previous 70. It skews the data towards the most recent and isn't a fair reflection of total historical competition between the two nations.... or any other nations in fact.

Thats why, anyone who has been significantly better over the last 20 years compared to their history will seem better then they actually are.

You can't make linear extrapolations based on very small sample sizes and then draw inferences from them. It's statistically invalid. Which constitutes making up the stats. So I'm right.

But I do take your general point. Which is valid, despite not being backed up by the numbers you are trying to use.

well given SA were also more successful in NZ then NZ were in SA (35% vs 25%) I would still stand by the point that whilst home referees were used, SA were the better side in the first 70 years of the rivarly.

Well unless we're suggesting NZ referees were less biased then their bokke counterparts.

but of course they weren't FA, only south Africans cheat. You should know that.

GloriousEmpire wrote:

fa0019 wrote:I wasn't making up stats, it was an example to show that your figures are biased towards the present due to the number of games being played in the last 20 years superceding those played in the previous 70. It skews the data towards the most recent and isn't a fair reflection of total historical competition between the two nations.... or any other nations in fact.

Thats why, anyone who has been significantly better over the last 20 years compared to their history will seem better then they actually are.

You can't make linear extrapolations based on very small sample sizes and then draw inferences from them. It's statistically invalid. Which constitutes making up the stats. So I'm right.

But I do take your general point. Which is valid, despite not being backed up by the numbers you are trying to use.

of course its statistically invalid... as it didn't happen, those games were never played. But it was never to suggest that the bokke had a better win rate... which I never tried to argue.... just that NZ claim via your statistics is skewed. The figures were there to suggest how they were skewed.

fa0019 wrote:well given SA were also more successful in NZ then NZ were in SA (35% vs 25%) I would still stand by the point that whilst home referees were used, SA were the better side in the first 70 years of the rivarly.

Well unless we're suggesting NZ referees were less biased then their bokke counterparts.

The NZ refs, precursors to Bryce Lawrence, were much fairer to visitors.

Biltong wrote:but of course they weren't FA, only south Africans cheat. You should know that.

never play poker with a boer thats all I know!

GE, by the way, you cannot wind me up.

I don't expect the boks to win tomorrow, so I won't be upset if we lose, I am merely looking to see if we have actually shown improvement over the last 12 months against a strong opponent as the others weren't.

GloriousEmpire wrote:

fa0019 wrote:well given SA were also more successful in NZ then NZ were in SA (35% vs 25%) I would still stand by the point that whilst home referees were used, SA were the better side in the first 70 years of the rivarly.

Well unless we're suggesting NZ referees were less biased then their bokke counterparts.

The NZ refs, precursors to Bryce Lawrence, were much fairer to visitors.

Completely sensible and supported with evidence statement there.

Well they made one verifiable decision, so if I project that over a hypothetical set of 10,000 decisions in 50 games, that's 500,000 decisions. Which is as close to 100% as you can get given that an average referee makes 100,000 decisions in a career and there are on average 5 referees at a time.

GloriousEmpire wrote:Sometimes stats are helpful. Like the statistic on the scoreboard. That's about it.

For instance. NZ are poised to become the most successful test nation ever, by eclipsing France's record tally of victories overall. But given that France have wedged in nearly 30% more test matches than NZ somehow, this is an utterly futile record. Right up there with the worlds record test try scorer.

Another irritation of mine is the way Sam Cane is being talked up in NZ media right for having a high statistic for "arrivals at rucks". No doubt the guy is athletic and gets around the park, but more important is what he does when he gets there. I don't care if he hits 70 rucks a game if he's utterly effectless when he gets there, or gives away penalties. Not saying the guy can't do his job, just that this stat is meaningless.

The next one is winning streaks. SA have 9, NZ have 6 on the trot. So what? one team will lose and what bearing on the result will the previous games have? None.

Percentage successful kicks. Again, says nothing about the record of a kicker against a given opponent, or take into account the elements of altitude or weather involved.

These things get bandied around to no end in the media, and I think they smack of lazy journalism or a writer without the ability to express a coherent opinion based on his or her own knowledge and ability to read the game or analyse the factors that are likely to be pertinent.

For example, I put it to you that NZ are already the most succesful test nation ever. And one more victory, or not,  isn't going to change that.

Presumably if you lose that does make a difference. Unless you're making a 'linear extrapolation' that imagines the ABs would have won one of the 30% extra games if they'd played them. And of course, as you say, you can't do that.

GloriousEmpire wrote:Well they made one verifiable decision, so if I project that over a hypothetical set of 10,000 decisions in 50 games, that's 500,000 decisions. Which is as close to 100% as you can get given that an average referee makes 100,000 decisions in a career and there are on average 5 referees at a time.

Is this the sort of extrapolation you meant when you said 'you can't make a linear extrapolation'?

No, my argument is that the "most successful" team cannot be evaluated by adding up the total number of wins.

Of course, it's winning percentage that matters.

Just as you can pick the winning % of the boks back then to somehow project it forward to allow the boks to win 96 tests we can do the same and push the recent results back over time to argue that the boks were probably lucky to win a match in the last 100 years? no? youre right fa...stats can be misleading.

Hood83 wrote:

GloriousEmpire wrote:Well they made one verifiable decision, so if I project that over a hypothetical set of 10,000 decisions in 50 games, that's 500,000 decisions. Which is as close to 100% as you can get given that an average referee makes 100,000 decisions in a career and there are on average 5 referees at a time.

Is this the sort of extrapolation you meant when you said 'you can't make a linear extrapolation'?

 

http://en.wikipedia.org/wiki/Reductio_ad_absurdum

Biltong wrote:but of course they weren't FA, only the Welsh cheat and south Africans eat meat. You should know that.

Fixed it for you.

I love the "fixed it for you" genre. It's always hilarious, isn't it?

It's almost as hilarious as you.

What is misleading is the influence of the old-timers on the overall winning stats that makes it seem that SA in the pro era have actually been more competitive than what they have been. The ABs have won 70% against SA in the pro era (ABs won 30/43). Surprised its that high. Yet, we include the old foggies so it's ABs at 56%.

GE I think you should do a pod cast on rugby. It would be entertaining anyway.

ebop wrote:What is misleading is the influence of the old-timers on the overall winning stats that makes it seem that SA in the pro era have actually been more competitive than what they have been. The ABs have won 70% against SA in the pro era (ABs won 30/43). Surprised its that high. Yet, we include the old foggies so it's ABs at 56%.

Exactly, all stats should be forgotten as soon as they are inconvenient.

I've seen this done, and it works. For example, all Welsh rugby fans can't remember anything that happened between 1982 and 2009, but still have razor sharp memory of a rainy day in 1905, having all been there personally.

England on the other hand, won the inaugural world cup final in 2003, competed in two further finals in 2007 and an undisclosed date and beat the All Blacks last year and have played no other rugby whatsoever.

GloriousEmpire wrote:Another irritation of mine is the way Sam Cane is being talked up in NZ media right for having a high statistic for "arrivals at rucks". No doubt the guy is athletic and gets around the park, but more important is what he does when he gets there. I don't care if he hits 70 rucks a game if he's utterly effectless when he gets there, or gives away penalties. Not saying the guy can't do his job, just that this stat is meaningless.

I've been banging on about this for ages, and it's a problem that has been a real issue in Scottish rugby. Players "leaning" on rucks rather than either (a) smashing into the ruck and clearing out the opposition effectively thus securing the ball, (b) smashing into the ruck in such a way as to disrupt or slow down the opposition call, or (c) not smashing into the ruck but rather accepting it's lost and making sure that you're in position to defend the line and compete more effectively at the next one.

I've seen Scots player manage to someone "attend" the ruck without achieving any of the above, and Al Kellock knows what's I'm talking about, but so do Ross Ford, Euan Murray, Jim Hamilton, Grant Gilchrist and David Denton. Putting one hand on the ruck for support so you can catch your breath is not "arriving at a ruck".

It's why players like Ryan Grant, Tim Swinson, Kelly Brown, Al Strokosch, John Barclay and Ross Rennie are so important to Scotland's progress. These guys invariably pick on of options (a) - (c) and make a real difference. The last England game in the 6 Nations was a classic example of us being too slow and passive in the rucks and contact areas, effectively standing off and allowing England to completely dictate the pace of the game.

GloriousEmpire wrote:

Hood83 wrote:

GloriousEmpire wrote:Well they made one verifiable decision, so if I project that over a hypothetical set of 10,000 decisions in 50 games, that's 500,000 decisions. Which is as close to 100% as you can get given that an average referee makes 100,000 decisions in a career and there are on average 5 referees at a time.

Is this the sort of extrapolation you meant when you said 'you can't make a linear extrapolation'?

 

http://en.wikipedia.org/wiki/Reductio_ad_absurdum

You don't really understand reductio ad absurdum, do you?

Hood83 wrote:

GloriousEmpire wrote:

Hood83 wrote:

GloriousEmpire wrote:Well they made one verifiable decision, so if I project that over a hypothetical set of 10,000 decisions in 50 games, that's 500,000 decisions. Which is as close to 100% as you can get given that an average referee makes 100,000 decisions in a career and there are on average 5 referees at a time.

Is this the sort of extrapolation you meant when you said 'you can't make a linear extrapolation'?

 

http://en.wikipedia.org/wiki/Reductio_ad_absurdum

You don't really understand reductio ad absurdum, do you?

Yes.

GloriousEmpire wrote:

Hood83 wrote:

GloriousEmpire wrote:

Hood83 wrote:

GloriousEmpire wrote:Well they made one verifiable decision, so if I project that over a hypothetical set of 10,000 decisions in 50 games, that's 500,000 decisions. Which is as close to 100% as you can get given that an average referee makes 100,000 decisions in a career and there are on average 5 referees at a time.

Is this the sort of extrapolation you meant when you said 'you can't make a linear extrapolation'?

 

http://en.wikipedia.org/wiki/Reductio_ad_absurdum

You don't really understand reductio ad absurdum, do you?

Yes.

Perhaps you're not familiar then with the fallacy of inconsistency?

To save you time in the future, there is a somewhat complete list of invalid argument techniques here:


http://www.don-lindsay-archive.org/skeptic/arguments.html


Presumably if you memorise them you can avoid the logical failings that usually cripple your arguments in the future.

Stats are useful in a defined context and correctly applied logic. The fault with misleading stats is with those presenting and interpreting them, not the statistics. Stats are not useless, they are very useful. See Scott Allen on the Roar for how brilliantly stats can be utilised

Though in general, people are criminally under educated in how to correctly interpret statistcs

Second hand stats are similarly a nightmare. But great points about Allen.

My pet bug bear is that the way stats are interpreted and presented in the media generally assumes their are no correlations between them.

Players are ridiculed for topping the list if handling errors for instance without mentioning their level of involvement comparatively.

For example a 10 throwing an intercept pass is bad, but a ball carrying 8 doing the same thing is inexcusable.

Faoo19 as you say Stats mean nothing BUT when talking AllBlacks v Boks it is true that Boks have always in the Amateur Era been strong.BUT the gap in 1981 was only 5 matches Bok
Rugby was strong BUT they weren't not playing Test Rugby in that era.England,Argentina,
A Barbarians style international sides and the NZ Cavaliers all played them 1981-92.
Indeed the Cavaliers was THE most successful NZ side 1928 apart to tour losing only a single
provincial match apart from the 3-1 loss result.THAT was the norm post 1956 between the 2 sides.Only in 1937 and 1949 when the AB`s had no Scrum until fixed courtesy DR,Craven in 1949.
Boks have been effected by an apartheid in reverse quota system since 1995,just as the All Blacks were up until 1981.
Token Whites were allowed 1970 and 1976,1976 final test and series lost instead of drawn
bcause of a politically motivated Ref[who admitted it to the AB team after the match].
George Nepia,Jimmy Mill and Bert Cooke stars of the 1925 side would certainly have made
a difference in 1928.1949 virtually half the squad including both half backs[9`s] would have
been Maori THE great Post War era then.
Hoodoo Grounds ,conditions etc mean NOTHING,win loss stats mean nothing,on the day you win or lose.
One thing is certain cutting away the reasons for it one stat remains very pertinent the AB`s
alone can tour SA and expect a near 50% win chance of winning there.
In 2008 after 20 minutes the ABs in the Second Test had 2 Locks one his second start,the other debuting in the Dunedin House of Pain.THE AB Fortress those two rookies outplayed
Matfield and Botha the whole match,the AB`s lost to a last minute solo try.
In 2008 post RWC the AB`s lost the best part of there Test Squad plus a lot of fringe players
they should have had no hope of winning the 3Ns.
Instead they went to SA and beat the Boks away to take the title so much for stats.in Super
and 3/4N Rugby Home win is a given for the top Nz/SA sides it`s the Norm.BUT the AB`s
are the exception with an away record in the Republic second to none.

Stats like these got the England arm chair fans with little knowledge of how the game works crying for Christopher Robshaw to be the Lions captain,
all they talked about was meters with ball and rucks!
He had very little effect at the ruck thou, yes a good hard worker but zero penetration.
Then the meters with ball stat,the English fans seemed unaware that he played the role of number 8 at kick off after Benjamin Morgan was injured, so you would expect a player who collects the kick off then to run up to ten meters to form a ruck ball out to clear the lines, after collecting three or four kick offs the stat would make him look like he had the foot work of Shane Williams.

I don't think that was any England fan's rationale as to why Robshaw should go on the lions tour... In fact I can't remember anyone claiming things using the stats you have suggested until you just did now

ChequeredJersey wrote:I don't think that was any England fan's rationale as to why Robshaw should go on the lions tour... In fact I can't remember anyone claiming things using the stats you have suggested until you just did now

It's a well know fact that people suppress memories of the events that lead up to a disappointment, so i imagine you you can't remember.

DeludedOptimistorjustDave wrote:

ChequeredJersey wrote:I don't think that was any England fan's rationale as to why Robshaw should go on the lions tour... In fact I can't remember anyone claiming things using the stats you have suggested until you just did now

It's a well know fact that people suppress memories of the events that lead up to a disappointment, so i imagine you you can't remember.

Well that at least was funny and a tad clever!

GloriousEmpire wrote:To save you time in the future, there is a somewhat complete list of invalid argument techniques here:


http://www.don-lindsay-archive.org/skeptic/arguments.html


Presumably if you memorise them you can avoid the logical failings that usually cripple your arguments in the future.

Thanks GE, think I covered your similar problems of logical meltdown on another thread. So perhaps pot and kettle?

Anyway, to your original point on stats - it's a fair one in general. A lot of people mention Parling's tackle count, but he NEVER stops a guy dead. You can guarantee that every tackle made probably gives away a metre a so and provides the opposition with momentum. I think issues like that are often an one of what is or isn't measured when collating stats.

funnyExiledScot wrote:

GloriousEmpire wrote:Another irritation of mine is the way Sam Cane is being talked up in NZ media right for having a high statistic for "arrivals at rucks". No doubt the guy is athletic and gets around the park, but more important is what he does when he gets there. I don't care if he hits 70 rucks a game if he's utterly effectless when he gets there, or gives away penalties. Not saying the guy can't do his job, just that this stat is meaningless.

I've been banging on about this for ages, and it's a problem that has been a real issue in Scottish rugby. Players "leaning" on rucks rather than either (a) smashing into the ruck and clearing out the opposition effectively thus securing the ball, (b) smashing into the ruck in such a way as to disrupt or slow down the opposition call, or (c) not smashing into the ruck but rather accepting it's lost and making sure that you're in position to defend the line and compete more effectively at the next one.

I've seen Scots player manage to someone "attend" the ruck without achieving any of the above, and Al Kellock knows what's I'm talking about, but so do Ross Ford, Euan Murray, Jim Hamilton, Grant Gilchrist and David Denton. Putting one hand on the ruck for support so you can catch your breath is not "arriving at a ruck".

It's why players like Ryan Grant, Tim Swinson, Kelly Brown, Al Strokosch, John Barclay and Ross Rennie are so important to Scotland's progress. These guys invariably pick on of options (a) - (c) and make a real difference. The last England game in the 6 Nations was a classic example of us being too slow and passive in the rucks and contact areas, effectively standing off and allowing England to completely dictate the pace of the game.

I think your rucking was often excellent under Robinson, far better than ours, but awful under Johnson and Ryan

GloriousEmpire wrote:Second hand stats are similarly a nightmare. But great points about Allen.

My pet bug bear is that the way stats are interpreted and presented in the media generally assumes their are no correlations between them.

Players are ridiculed for topping the list if handling errors for instance without mentioning their level of involvement comparatively.

For example a 10 throwing an intercept pass is bad, but a ball carrying 8 doing the same thing is inexcusable.

Are you now or have you ever been David Icke and or L. Rond Hubbard?

GloriousEmpire wrote:Second hand stats are similarly a nightmare. But great points about Allen.

My pet bug bear is that the way stats are interpreted and presented in the media generally assumes their are no correlations between them.

Players are ridiculed for topping the list if handling errors for instance without mentioning their level of involvement comparatively.

For example a 10 throwing an intercept pass is bad, but a ball carrying 8 doing the same thing is inexcusable.

But isn't that just human nature?
Keep checking the weather reports until you find one that says it will be sunny but ignore the other four that said it will rain?
Or just like when you find out your great granddad was a war hero then everyone in the family talks about hey they get their courage and valor from him, whilst no on claims to have anything in common with that feckless wife beating great uncle they have!!!!
People are massively selective in their reasoning in general.