Harte Signs

[Snowball]

Mind you LOL @ Snowball for taking a 6 game sample (post Gylfi 2010) and drawing a comparison with a 23 game sample (Gylfi games under McDermott).

EDIT. My bad, tbf to Snowball that post Gylfi comparison was requested by someone else. Still daft though.

I compared ALL non-Gylfi games = 16 games

I compared ALL non-Gylfi league games = 12 games

I looked at all games, Rodgers games, McDermott games

16 games is a very good sample. 12 games is a very good sample

As you point out, I posted the six games because somebody asked.


LOOK, LOGICALLY I would have expected a fall-away in results. Gylfi is a class act.
But last season, when he didn't play we had better results. FACT

So I checked ALL results with/without Gylfi and the stats held firm


There is no question that we have done and currently are doing, better without Gylfi. We have a different personality, other individuals getting their chance, players not thinking automatically "give it to Gylfi". It was NOT what I expected, but the effect is clear and real. Once an effect is shown (and this is shown - we have not missed Gylfi results-wise) then we start to consider WHY.

[Vision]

Still too many variables for you to draw the definitive conclusion you are, I'm afraid.

[Snowball]

The problem I have with this analysis is the small number of "after Gylfi" games you are analysing. Lets assume that we play brilliantly in our next match dominate but fail to score and in the 93 minute the ball flies into our goal off of a defenders backside. Your after Gylfi stats then become

P07 W03 D02 L02 09-05 1.57 ppg = 72.28 Point 46-Game Season League -- Games this season since Gylfi left

In other words not really any different to the with Gylfi figures on on a points per game basis.

Now would this prove Snowball wrong and others right - of course not.It merely proves that a decent sample is needed before making sweeping statments. So Snowball I really dont think you should be so vociferous based on such a small sample.

TRY AGAIN. I used a sample, NOT of six games. I used one of SIXTEEN games. We have played sixteen times WITHOUT Gylfi
and I used those stats first. The six-game thing was after someone requested it.

Despite PERCEPTIONS games where the God Gylfi played did NOT produce an amazing points tally.
We averaged "1.39 ppg" over 43 games (including cup games), a 64 point season

Yet, despite ten games where Gylfi DIDN'T play we were Reading 46 10 7 6 39 22 7 5 11 29 41 +5 63

So it is UTTERLY clear that at worst, the non-Gylfi team was a single point IN A SEASON worse than Gylfi (including cup games)

However those 43 games include cup game where we beat Burnley, Liverpool, WBA and had two draws and only lost to Villa (P6 W3 D2 L1)
and the Gylfi games become P37 W14 D11 L12 and that maps to 62 Points. So the non-Gylfi side gained a point in 9 games.
Not much? That's 5 points in a season. That would have meant being 2 points outside the play-offs compared to 7 points off.

And lo and behold, without Gylfi right now, we are on equal points with the the last play-off place, even though we started slowly (with Glyfi) with 5 points from 4 games.

[Snowball]

The problem I have with this analysis is the small number of "after Gylfi" games you are analysing. Lets assume that we play brilliantly in our next match dominate but fail to score and in the 93 minute the ball flies into our goal off of a defenders backside. Your after Gylfi stats then become

P07 W03 D02 L02 09-05 1.57 ppg = 72.28 Point 46-Game Season League -- Games this season since Gylfi left

In other words not really any different to the with Gylfi figures on on a points per game basis.

2.3 POINTS IN A SEASON IS NOT TRIVIAL

Your point about the sample size is fair enough (which is why I used 16, not 6) but to hypothesize
and say "IF" we lose, then we'll only be on target for a 2.3 point better season..." is silly

You could phrase it like this.

"We are so far in front of the Gylfi averages that we can afford to lose our next game and STILL be in front."

P07 W04 D02 L01 10-04 2.00 ppg = 92.00 Point 46-Game Season League -- Games this season since Gylfi left. Beating Swansea
P07 W03 D03 L01 10-05 1.71 ppg = 78.86 Point 46-Game Season League -- Games this season since Gylfi left. Drawing with Swansea
P07 W03 D02 L01 09-05 1.57 ppg = 72.28 Point 46-Game Season League -- Games this season since Gylfi left. Lose to Swansea

Gyfli season = 70 points, was worse than all three of these.

I KNOW six games is not enough, which is why I was using SIXTEEN...

12 Games, over a quarter of a season, however IS enough

Say we L2 D3 W1 of our next 6 (TERRIBLE, right?) Win just one in six

P12 W04 D05 L03 =1.42 ppg = a 65-Point Season. Gylfi last year was a 64 Point Season

[Wycombe Royal]

Snowball just out of interest, and because you seem to like taking requests can you show the AVERAGE league position of the opposition (at the time we played them) for the with and without Sig samples you are using.

And also can you say how many were home and away.

[Woodcote Royal]

Still too many variables for you to draw the definitive conclusion you are, I'm afraid.

But sufficient, surely, to support your earlier point in that any adverse effects from Gylfi's sale have been greatly overstated and, so far, we have consistently gained more points without the player than with, even going back to when he was still here.

However imperfect Snowball's stats maybe, he, at least, offers some evidence to back his views.

That's a lot more than we get from others who use nothing more than their tried and failed gut instincts to predict doom and gloom when an expensive is player is sold and/or to rubbish the likes of Harte, because they were signed for next to nothing.

Frankly, I'll take Snowball over those who slagged off Harte, Kitson etc with nothing whatsovever to support their views other than the size of their fee and that's before we get started on those who were trashing Kebe not long ago but are now bricking it for fear he might be sold in January!

[facaldaqui]

Davies made his debut for England at 33. Harte's not too old for Reading at the same age.

[Hoop Blah]

Davies made his debut for England at 33. Harte's not too old for Reading at the same age.

True, although there is of course the glaringly obvious difference in that Davies hasn't been given a two year contract on £Xk per week and as such is part of our squad for the foreseable future.

[Nomad_Royal]

Your point about the sample size is fair enough (which is why I used 16, not 6) but to hypothesize
and say "IF" we lose, then we'll only be on target for a 2.3 point better season..." is silly

You could phrase it like this.

"We are so far in front of the Gylfi averages that we can afford to lose our next game and STILL be in front."

P07 W04 D02 L01 10-04 2.00 ppg = 92.00 Point 46-Game Season League -- Games this season since Gylfi left. Beating Swansea
P07 W03 D03 L01 10-05 1.71 ppg = 78.86 Point 46-Game Season League -- Games this season since Gylfi left. Drawing with Swansea
P07 W03 D02 L01 09-05 1.57 ppg = 72.28 Point 46-Game Season League -- Games this season since Gylfi left. Lose to Swansea

Gyfli season = 70 points, was worse than all three of these.

I KNOW six games is not enough, which is why I was using SIXTEEN...

12 Games, over a quarter of a season, however IS enough

Say we L2 D3 W1 of our next 6 (TERRIBLE, right?) Win just one in six

P12 W04 D05 L03 =1.42 ppg = a 65-Point Season. Gylfi last year was a 64 Point Season

I really should keep to my principles and not respond to people who put lots of capital letters in their posts - but ...

Snowball I think at various times you have used 6,12 and 16 games to prove your point (ok the 6 was in response to a question from someone else )- you have also earlier used an anlysis of ( I think) 7 games to try and prove a point about Harte. You acknowledge that 6 games is not enough despite using that number to make the statment:-

"....we are clearly better?"

You then gone to try and draw conclusion on 7 games by promptly extrapolating on a win,draw and lose basis against Swansea.( Before you say that is what I did I was using it as an example of the uselessness of doing so). You have shown yourself as soneone who will use small number to try a prove a point.

You also make the defintive statement that 12 games is enough to draw a conclusion - are you sure- I take you have done proper statistical analysis to prove that.

Comparing the results of games last season when we underwent huge changes (non Gylfi related changes) with this season aren't valid.

The point I am making is simply that I dont think that we are yet in a position to fully analyse the effect losing Gylfi and no amount of stats on such a small number of games can provide you with evidence that really stands up to scrutiny.

[Vision]

Still too many variables for you to draw the definitive conclusion you are, I'm afraid.

But sufficient, surely, to support your earlier point in that any adverse effects from Gylfi's sale have been greatly overstated and, so far, we have consistently gained more points without the player than with, even going back to when he was still here.

However imperfect Snowball's stats maybe, he, at least, offers some evidence to back his views.

That's a lot more than we get from others who use nothing more than their tried and failed gut instincts to predict doom and gloom when an expensive is player is sold and/or to rubbish the likes of Harte, because they were signed for next to nothing.

Frankly, I'll take Snowball over those who slagged off Harte, Kitson etc with nothing whatsovever to support their views other than the size of their fee and that's before we get started on those who were trashing Kebe not long ago but are now bricking it for fear he might be sold in January!

A world of difference between saying we won't miss him as much as some suggest to Snowball's "we are definitely a better team without him"

Using the stats to enhance a point is one thing but the numerous variables dictate that making definitive statements about what we WILL become based on those stats is deeply flawed.

EG
How about seperating games where he played in his most influential position of behind the front man as opposed to left midfield where certainly many of his earlier games were played.

Or we could throw in the fact that as a young player whether his individual form would get better with each passing season or not given his improvement from 08/09 to 09/10. How can any stat definitively predict how good 2010/11 Gylfi would have been for us?

I dont dispute that as ever there are one or two on here who greatly exaggerate the doom whenever a certain player is sold or when another is bought in but thats not the point I'm making.

[brendywendy]

You acknowledge that 6 games is not enough despite using that number to make the statment:-

"....we are clearly better?"

but he did the stats for all games last season with and without the gylfmeister.

[Vision]

You acknowledge that 6 games is not enough despite using that number to make the statment:-

"....we are clearly better?"

but he did the stats for all games last season with and without the gylfmeister.

He started 19 games under McDermott and there were 6 where he didn't. Its a bit skewed to say the least.

He started 23 league games in total under McDermott reign , whilst so far 12 he has not.

At least give it another 11 games so that the starts are equal (conveniently it would amount to the games as a full League season) and the statistics would be more comparable rather than simply multiplying an average.

Even then there are so many other factors at play but at least the numbers would at least be equal.

[Woodcote Royal]

Still too many variables for you to draw the definitive conclusion you are, I'm afraid.

But sufficient, surely, to support your earlier point in that any adverse effects from Gylfi's sale have been greatly overstated and, so far, we have consistently gained more points without the player than with, even going back to when he was still here.

However imperfect Snowball's stats maybe, he, at least, offers some evidence to back his views.

That's a lot more than we get from others who use nothing more than their tried and failed gut instincts to predict doom and gloom when an expensive is player is sold and/or to rubbish the likes of Harte, because they were signed for next to nothing.

Frankly, I'll take Snowball over those who slagged off Harte, Kitson etc with nothing whatsovever to support their views other than the size of their fee and that's before we get started on those who were trashing Kebe not long ago but are now bricking it for fear he might be sold in January!

A world of difference between saying we won't miss him as much as some suggest to Snowball's "we are definitely a better team without him"

Using the stats to enhance a point is one thing but the numerous variables dictate that making definitive statements about what we WILL become based on those stats is deeply flawed.

EG
How about seperating games where he played in his most influential position of behind the front man as opposed to left midfield where certainly many of his earlier games were played.

Or we could throw in the fact that as a young player whether his individual form would get better with each passing season or not given his improvement from 08/09 to 09/10. How can any stat definitively predict how good 2010/11 Gylfi would have been for us?

I dont dispute that as ever there are one or two on here who greatly exaggerate the doom whenever a certain player is sold or when another is bought in but thats not the point I'm making.

Fair enough but I think there's a strong case for saying that in signing Harte and Zurab on the back of Gyfi's transfer has made for a stronger all round squad, especially with the added bonus of Armstrong's surprise return in midfield.

[brendywendy]

You acknowledge that 6 games is not enough despite using that number to make the statment:-

"....we are clearly better?"

but he did the stats for all games last season with and without the gylfmeister.

He started 19 games under McDermott and there were 6 where he didn't. Its a bit skewed to say the least.

He started 23 league games in total under McDermott reign , whilst so far 12 he has not.

At least give it another 11 games so that the starts are equal (conveniently it would amount to the games as a full League season) and the statistics would be more comparable rather than simply multiplying an average.

Even then there are so many other factors at play but at least the numbers would at least be equal.

yes. its a slightly skewed and smallish sample.

as long as you remember that when reading the stats its fine.
suffice to say that the stats we have available at this time do point towards there being a positive to no difference since he left

[Ian Royal]

I think I'll just leave it that drawing a conclusion and stating it as an absolute certainty from a sample of statistics which take account of zero variables other than one player's appearance in the team (or possibly only starting line up) is about as retarded as anything I have ever seen on here.

It shows a complete lack of understanding of how to make a reliable statistical analysis on even the most basic level.

[Snowball]

You acknowledge that 6 games is not enough despite using that number to make the statment:-

"....we are clearly better?"

but he did the stats for all games last season with and without the gylfmeister.

Shush, Wendy. There's selective statistics, and then there's selective reading of what is actually posted.

You are quite right, of course

[Snowball]

Snowball just out of interest, and because you seem to like taking requests can you show the AVERAGE league position of the opposition (at the time we played them) for the with and without Sig samples you are using.

And also can you say how many were home and away.

You gotta be kidding me!

Look, if you play 46 games you play all the teams, right?

And 37 is a a very large proportion of that 46.

[Snowball]

Snowball I think at various times you have used 6,12 and 16 games to prove your point (ok the 6 was in response to a question from someone else )- you have also earlier used an anlysis of ( I think) 7 games to try and prove a point about Harte. You acknowledge that 6 games is not enough despite using that number to make the statment:-

"....we are clearly better?"

I tried to use ALL games since Gylfi began to be selected. However, some of those games were league cup or FA Cup and these can skew the data. For example, Gylfi didn't play in the two league cup games this season and we won one, drew the other (I don't count penalty shoot-outs in stats as they are a lottery). The cup games this year would bias AGAINST Gylfi, so I removed them. That's why we have

16 games without v 43 with (a HUGE sample, more than a season's worth)
12 games without v 37 with (a very large sample. 12 is a quarter of a season -and they are randomly spread, not a close run 12. 37 is 80% of a season)

Those games involved two managers so I also posted stats for McDermott-Managed games only
(because people tried to allege Rodgers games skewed the data. They didn't)

There is also a comparison on "just this season" 4 Gyfli games versus 6 since he left. i know AND HAVE SAID that this is a small sample
but the sample size grows every game and at least people can't bleat about Rodgers


You then gone to try and draw conclusion on 7 games by promptly extrapolating on a win,draw and lose basis against Swansea.( Before you say that is what I did I was using it as an example of the uselessness of doing so). You have shown yourself as someone who will use small number to try a prove a point.

NO. I don't rely SOLELY on the 6/7 games but ALSO on my knowledge of ALL the Gylfi/non-Gylfi games.
There is NO set of league games you can decide to use that will show Gylfi Games doing better than non-Gylfi.

To discard the ten non-Gylfi games last season is plain daft. What if we had lost them all 2-0 each and I refused to use them?

the six games this year are ANOTHER EXAMPLE of Gylfi being shown NOT to improve our points-scoring. That's all.

You also make the definitive statement that 12 games is enough to draw a conclusion - are you sure- I take you have done proper statistical analysis to prove that.

YES I AM SURE. 12 games from 46 (more than 25%) is a large sample.

A single 12-consecutive game run MIGHT be argued to be less certain because runs can end
but I have never heard a reasonable pundit suggesting that 12 games from 46 is not a good measure.


Comparing the results of games last season when we underwent huge changes (non Gylfi related changes) with this season aren't valid.

I DISAGREE. Just "stating" that means nothing

Imagine that last season we averaged 2.2 points per game with Gylfi
and 1.2 ppg without, and I said, "I'm ignoring last season's games
as they aren't valid. There'd be an uproar.

of COURSE they are valid. The most recent results are MORE valid, but last year means something, of course it does.


The point I am making is simply that I dont think that we are yet in a position to fully analyse the effect losing Gylfi and no
amount of stats on such a small number of games can provide you with evidence that really stands up to scrutiny.


But if we had had W3 D1 with Gylfi in those first four games

and W1 D2 L3 in the six since he left, THEN the comparison would be shoved down my throat!!

[Snowball]

A world of difference between saying we won't miss him as much as some suggest to Snowball's "we are definitely a better team without him"


IT is an ABSOLUTE FACT that in terms of goal-difference and points per game we are better. END

Using the stats to enhance a point is one thing but the numerous variables dictate that making
definitive statements about what we WILL become based on those stats is deeply flawed.

EXCUSE ME? When have I EVER said anything about what we WILL become?

I have given season's equivalents but I have NEVER said "That is what we will achieve"

I post 46-game equivalents because they are easier to absorb and understand than 1.534 ppg

ANYTHING could happen. We could be done for fielding an ineligible player and lose points.
We could have a bug go thru' the team. We could get 3-4-5 serious injuries to key players,
we could have a run of bad luck. We could simply lose form.

Or on the positive side, we could get a brilliant loanee or two, get in a VG striker, have three 1-0 wins
when we should have lost, have a newcomer like Antonio or Taylor burst into the team and have a purple patch.


All the projection is AND ALL I HAVE EVER SAID IT IS, is "what would happen if we maintain this average"



EG
How about seperating games where he played in his most influential position of behind the front man as opposed to left midfield where certainly many of his earlier games were played.

TOO DIFFICULT.

Or we could throw in the fact that as a young player whether his individual form would get better with each passing season
or not given his improvement from 08/09 to 09/10. How can any stat definitively predict how good 2010/11 Gylfi would have been for us?

But there ARE stats to show if that occurred.

You can look at Gyfli's first 18/19 games in the league versus his second 18/19.

You can ask, were we picking up lots of extra points this season?

We faded badly at the end of last season (when Gylfi should have been IMPROVING)

and we started fairly poorly this year

[Snowball]

He started 19 games under McDermott and there were 6 where he didn't. Its a bit skewed to say the least.


What is "skewed" about that? That is a classic "random sample"

I do not know, beforehand which games Gylfi will play in or will not.
Having six non-Gylfi games interspersed with Gylfi-Games is a statistician's dream



He started 23 league games in total under McDermott reign , whilst so far 12 he has not.

And 23 v 12 is a very very decent sample to make judgments from

and we are averaging 58% wins versus 40% wins. That is a monster-difference



At least give it another 11 games so that the starts are equal (conveniently it would amount to the games as a full League season)
and the statistics would be more comparable rather than simply multiplying an average.

There is no reason whatsoever to require "equal number of games"

All that matters is that both sets of games are of a reasonable size