Alonso and Vettel

[pokerman]

Well, with selective figures, you can show everything.

We could focus on 2016 only for the Vettel-Räikkönen comparison, for instance.

why would you?

In order to twist the analysis to be as void as the selective one by KingVoid.

It’s selective because otherwise the entire cross comparison doesn’t make sense.

Alonso had 0.287s over Massa on average but 0.528s over Raikkonen

This would imply that Massa was as dominant over Kimi as Alonso was over him. We know that isn’t true. Massa and Kimi were evenly matched at Ferrari.

Alonso and Massa spend 4 years together while Alonso and Kimi spend only 1 year together, so it’s likely that Alonso vs Massa was more representative because we have a lot more data.

If you want to believe that Alonso is actually over half a second faster than Kimi on average, then you are free to believe that.

That's because your methodology is flawed, I have it:-

Alonso > Massa 0.27s
Alonso > Kimi 0.28s

[Zoue]

I guess we all have different methodologies I had Alonso's average for Kimi at 0.28s for the season.

Vettel's 2015 average was 0.24s.

which means they are pretty much even, I guess

Yes if you want to highlight individual years, Vettel and Kimi are into their 4th season as teammates.

we've been through this before. Things can change year on year. In their first year as Kimi's team mate Seb and Alonso had virtually identical stats

No if you just want to use 1 year and ignore the other 3 then that's what you call cherry picking.

No it's called taking an equivalent sample

[Paolo_Lasardi]

Ignoring available information is called a flawed sample.

[KingVoid]

Ignoring available information is called a flawed sample.

Webber beat Rosberg by an average of 0.140s and a median of 0.235s in 2006. That would imply Vettel is close to half a second faster than Rosberg, and therefore about 4 tenths faster than Hamilton.

Of course, that Rosberg was a rookie is considered a valid excuse, but the fact that Kimi (who has a reputation for being a very sensitive driver) hated the handling of the F14T is not a valid excuse.

[Zoue]

Ignoring available information is called a flawed sample.

depends whether you want the results to be honest or not. It also depends on what you are looking to measure. So if you want to measure Vettel against Kimi, then it's right to compare their whole time together. But if you want to compare how both Vettel and Alonso fared against Kimi, then it's only fare to make the comparison as equal as possible. Which means taking the most equal sample which is a close in size to each other as possible

[Exediron]

That's because your methodology is flawed, I have it:-

Alonso > Massa 0.27s
Alonso > Kimi 0.28s

Just because he's using a different methodology than yours does not give you the right to call it flawed. KingVoid has explained several time the method he's using, and as far as I can see he's applying it correctly. It gives a different answer to the one you use, but that does not make it flawed. If he was also trying to calculate a mean average and getting a different result than you, then you could say he's doing it wrong. He's not trying to do that.

Personally, I think it's astonishing how much stock you put in the mean average over a season down to the hundredth of a second, when a single race can easily throw those hundredths away.

[davidheath461]

Ignoring available information is called a flawed sample.

Webber beat Rosberg by an average of 0.140s and a median of 0.235s in 2006. That would imply Vettel is close to half a second faster than Rosberg, and therefore about 4 tenths faster than Hamilton.

Of course, that Rosberg was a rookie is considered a valid excuse, but the fact that Kimi (who has a reputation for being a very sensitive driver) hated the handling of the F14T is not a valid excuse.

do you think Alonso liked the handling of the F14T ?

[KingVoid]

Ignoring available information is called a flawed sample.

Webber beat Rosberg by an average of 0.140s and a median of 0.235s in 2006. That would imply Vettel is close to half a second faster than Rosberg, and therefore about 4 tenths faster than Hamilton.

Of course, that Rosberg was a rookie is considered a valid excuse, but the fact that Kimi (who has a reputation for being a very sensitive driver) hated the handling of the F14T is not a valid excuse.

do you think Alonso liked the handling of the F14T ?

He is better at dealing with understeer than Kimi. Andrea Stella himself said that Kimi has a very small operational window.

[davidheath461]

Ignoring available information is called a flawed sample.

Webber beat Rosberg by an average of 0.140s and a median of 0.235s in 2006. That would imply Vettel is close to half a second faster than Rosberg, and therefore about 4 tenths faster than Hamilton.

Of course, that Rosberg was a rookie is considered a valid excuse, but the fact that Kimi (who has a reputation for being a very sensitive driver) hated the handling of the F14T is not a valid excuse.

do you think Alonso liked the handling of the F14T ?

He is better at dealing with understeer than Kimi. Andrea Stella himself said that Kimi has a very small operational window.

Right, and that's Kimi's problem - a clear weakness he has as a driver.

[KingVoid]

That’s precisely why assuming that Raikkonen is some kind of constant when comparing Vettel to Alonso is tricky. He has a small operational window and his speed fluctuates a lot.

We don’t have all the information on how Vettel compares relative to Alonso. Ferrari do, and they appear to be very happy with Vettel as their number 1.

[Zoue]

Ignoring available information is called a flawed sample.

Webber beat Rosberg by an average of 0.140s and a median of 0.235s in 2006. That would imply Vettel is close to half a second faster than Rosberg, and therefore about 4 tenths faster than Hamilton.

Of course, that Rosberg was a rookie is considered a valid excuse, but the fact that Kimi (who has a reputation for being a very sensitive driver) hated the handling of the F14T is not a valid excuse.

do you think Alonso liked the handling of the F14T ?

He is better at dealing with understeer than Kimi. Andrea Stella himself said that Kimi has a very small operational window.

Right, and that's Kimi's problem - a clear weakness he has as a driver.

I don't think anyone is arguing Kimi doesn't have a weakness, so that's a bit of a strawman

[davidheath461]

That’s precisely why assuming that Raikkonen is some kind of constant when comparing Vettel to Alonso is tricky. He has a small operational window and his speed fluctuates a lot.

We don’t have all the information on how Vettel compares relative to Alonso. Ferrari do, and they appear to be very happy with Vettel as their number 1.

How do Ferrari have a comparison? the 2 of them never drove the same car at the same time.

[davidheath461]

I don't think anyone is arguing Kimi doesn't have a weakness, so that's a bit of a strawman

Except that i never claimed that anyone was arguing that Kimi doesn't have a weakness.

[Exediron]

We don’t have all the information on how Vettel compares relative to Alonso. Ferrari do, and they appear to be very happy with Vettel as their number 1.

You haven't gotten very much disagreement on that point. It's where you try to take that and make from it that Ferrari believes Vettel is just as fast that I have an issue. They could very easily be happy with him despite knowing (or suspecting) that he gives up a tenth to Alonso in pure speed.

[KingVoid]

That’s precisely why assuming that Raikkonen is some kind of constant when comparing Vettel to Alonso is tricky. He has a small operational window and his speed fluctuates a lot.

We don’t have all the information on how Vettel compares relative to Alonso. Ferrari do, and they appear to be very happy with Vettel as their number 1.

How do Ferrari have a comparison? the 2 of them never drove the same car at the same time.

Ferrari engineers have telemetry and data from both drivers. They might not know exactly how much faster one is than the other, but they likely have a much better idea than we do

[Zoue]

I don't think anyone is arguing Kimi doesn't have a weakness, so that's a bit of a strawman

Except that i never claimed that anyone was arguing that Kimi doesn't have a weakness.

then it's a bit of a non-sequitur in regard to the conversation being made, wouldn't you say?

[davidheath461]

That’s precisely why assuming that Raikkonen is some kind of constant when comparing Vettel to Alonso is tricky. He has a small operational window and his speed fluctuates a lot.

We don’t have all the information on how Vettel compares relative to Alonso. Ferrari do, and they appear to be very happy with Vettel as their number 1.

How do Ferrari have a comparison? the 2 of them never drove the same car at the same time.

Ferrari engineers have telemetry and data from both drivers. They might not know exactly how much faster one is than the other, but they likely have a much better idea than we do

Telemetry will only tell you about driving styles - not which driver is quicker. For that, you need a stop clock and both drivers to be driving the same car at the same time.

[davidheath461]

I don't think anyone is arguing Kimi doesn't have a weakness, so that's a bit of a strawman

Except that i never claimed that anyone was arguing that Kimi doesn't have a weakness.

then it's a bit of a non-sequitur in regard to the conversation being made, wouldn't you say?

how so?

[Black_Flag_11]

That’s precisely why assuming that Raikkonen is some kind of constant when comparing Vettel to Alonso is tricky. He has a small operational window and his speed fluctuates a lot.

We don’t have all the information on how Vettel compares relative to Alonso. Ferrari do, and they appear to be very happy with Vettel as their number 1.

How do Ferrari have a comparison? the 2 of them never drove the same car at the same time.

Ferrari engineers have telemetry and data from both drivers. They might not know exactly how much faster one is than the other, but they likely have a much better idea than we do

Telemetry will only tell you about driving styles - not which driver is quicker. For that, you need a stop clock and both drivers to be driving the same car at the same time.

I saw a story not too long ago about how they measure a drivers performance in the race, they called it an effort rating.

They basically took the fuel level etc. Into account and worked out what time the car was capable of at any point in the race and then scored the drivers on how close they were to achieving that as a percentage (so achieving the best possible laptime every lap of the race would give a score of 100%).

The article was from an engineer who said that Schumacher was always in the high 90% bracket and basically no-one else was. Based on that a team could compare two drivers even if they weren't driving at the same time, but obviously it wouldn't take into account performance fluctuation with different cars etc.

Of course the big flaw with that is that I assume they calculate the 100% time based largely on practice sessions so its almost scoring a driver against their own potential times. They may get a good idea of race pace but not absolute pace.

[pokerman]

which means they are pretty much even, I guess

Yes if you want to highlight individual years, Vettel and Kimi are into their 4th season as teammates.

we've been through this before. Things can change year on year. In their first year as Kimi's team mate Seb and Alonso had virtually identical stats

No if you just want to use 1 year and ignore the other 3 then that's what you call cherry picking.

No it's called taking an equivalent sample

It's only equivalent if the partnerships both only lasted 1 year, let's take a theoretical sample, driver A beats driver B, also driver C beats driver B, but after the first year driver B beats driver C three years running, so the overview is that driver B is better than driver C and because driver A was better than driver B he is also better than driver C, but according to your equivalent sample driver C is at a similar level to driver A, that's just nonsense because you're basically saying that driver B after losing to driver A in his first year then would have beat driver A every year after that.

[pokerman]

Ignoring available information is called a flawed sample.

Webber beat Rosberg by an average of 0.140s and a median of 0.235s in 2006. That would imply Vettel is close to half a second faster than Rosberg, and therefore about 4 tenths faster than Hamilton.

Of course, that Rosberg was a rookie is considered a valid excuse, but the fact that Kimi (who has a reputation for being a very sensitive driver) hated the handling of the F14T is not a valid excuse.

Rookie drivers improve, the Hulk got beat easily by Barrichello in his rookie year.

[pokerman]

Ignoring available information is called a flawed sample.

depends whether you want the results to be honest or not. It also depends on what you are looking to measure. So if you want to measure Vettel against Kimi, then it's right to compare their whole time together. But if you want to compare how both Vettel and Alonso fared against Kimi, then it's only fare to make the comparison as equal as possible. Which means taking the most equal sample which is a close in size to each other as possible

How you deal with numbers, ignoring data, is not a honest way with dealing with numbers at all, plus I would say that you're not really interested in crunching numbers yourself just in disproving data that give results that you don't like.

I dare say if you swapped 2015 and 2016 around then you would come up that 2016 was more representative because Vettel was new to the team in 2015.

[pokerman]

That's because your methodology is flawed, I have it:-

Alonso > Massa 0.27s
Alonso > Kimi 0.28s

Just because he's using a different methodology than yours does not give you the right to call it flawed. KingVoid has explained several time the method he's using, and as far as I can see he's applying it correctly. It gives a different answer to the one you use, but that does not make it flawed. If he was also trying to calculate a mean average and getting a different result than you, then you could say he's doing it wrong. He's not trying to do that.

Personally, I think it's astonishing how much stock you put in the mean average over a season down to the hundredth of a second, when a single race can easily throw those hundredths away.

If you're using wet qualifying sessions, sessions were drivers did not take part, ignore sessions were a driver had a problem, then I can see how you come up with head scratching figures were 2 drivers that were so evenly matched are then 3 tenths apart when matched with another driver.

[Rockie]

Ignoring available information is called a flawed sample.

depends whether you want the results to be honest or not. It also depends on what you are looking to measure. So if you want to measure Vettel against Kimi, then it's right to compare their whole time together. But if you want to compare how both Vettel and Alonso fared against Kimi, then it's only fare to make the comparison as equal as possible. Which means taking the most equal sample which is a close in size to each other as possible

How you deal with numbers, ignoring data, is not a honest way with dealing with numbers at all, plus I would say that you're not really interested in crunching numbers yourself just in disproving data that give results that you don't like.

I dare say if you swapped 2015 and 2016 around then you would come up that 2016 was more representative because Vettel was new to the team in 2015.

Why is it that with Vettel the representative year is always the abnormal one?

[pokerman]

Ignoring available information is called a flawed sample.

depends whether you want the results to be honest or not. It also depends on what you are looking to measure. So if you want to measure Vettel against Kimi, then it's right to compare their whole time together. But if you want to compare how both Vettel and Alonso fared against Kimi, then it's only fare to make the comparison as equal as possible. Which means taking the most equal sample which is a close in size to each other as possible

How you deal with numbers, ignoring data, is not a honest way with dealing with numbers at all, plus I would say that you're not really interested in crunching numbers yourself just in disproving data that give results that you don't like.

I dare say if you swapped 2015 and 2016 around then you would come up that 2016 was more representative because Vettel was new to the team in 2015.

Why is it that with Vettel the representative year is always the abnormal one?

Ask Zoue I don't deal in representative years, I use all the years available.

[KingVoid]

Ignoring available information is called a flawed sample.

Webber beat Rosberg by an average of 0.140s and a median of 0.235s in 2006. That would imply Vettel is close to half a second faster than Rosberg, and therefore about 4 tenths faster than Hamilton.

Of course, that Rosberg was a rookie is considered a valid excuse, but the fact that Kimi (who has a reputation for being a very sensitive driver) hated the handling of the F14T is not a valid excuse.

Rookie drivers improve, the Hulk got beat easily by Barrichello in his rookie year.

Hulkenberg was 0.140s slower than Barrichello in qualifying on average in 2010.

When you do a quick Hukenberg-Perez-Button-Barrichello cross comparison, you’ll find that gap is actually fairly representative. Hulkenberg and Perez aren’t quicker than Barrichello and Button.

Regarding Vettel vs Rosberg, Vettel was only a year older in 2009 than Rosberg was in 2006.

In 2006, Webber outqualified Rosberg 12-6 and by 2 tenths on average.
In 2009, Vettel outqualified Webber 15-2 and by 3 tenths on average.

If we use Webber as a constant, we come to the conclusion that a 22 year old Vettel was half a second faster than a 21 year old Rosberg.

Rookie drivers do improve, but by half a second?

[pokerman]

Ignoring available information is called a flawed sample.

Webber beat Rosberg by an average of 0.140s and a median of 0.235s in 2006. That would imply Vettel is close to half a second faster than Rosberg, and therefore about 4 tenths faster than Hamilton.

Of course, that Rosberg was a rookie is considered a valid excuse, but the fact that Kimi (who has a reputation for being a very sensitive driver) hated the handling of the F14T is not a valid excuse.

Rookie drivers improve, the Hulk got beat easily by Barrichello in his rookie year.

Hulkenberg was 0.140s slower than Barrichello in qualifying on average in 2010.

When you do a quick Hukenberg-Perez-Button-Barrichello cross comparison, you’ll find that gap is actually fairly representative. Hulkenberg and Perez aren’t quicker than Barrichello and Button.

Regarding Vettel vs Rosberg, Vettel was only a year older in 2009 than Rosberg was in 2006.

In 2006, Webber outqualified Rosberg 12-6 and by 2 tenths on average.
In 2009, Vettel outqualified Webber 15-2 and by 3 tenths on average.

If we use Webber as a constant, we come to the conclusion that a 22 year old Vettel was half a second faster than a 21 year old Rosberg.

Rookie drivers do improve, but by half a second?

Again using your methodology, in 2009 I had Vettel out qualifying Webber by 0.14s, also age has nothing to do with being a F1 rookie, even Vettel said it took him a year to get up to speed in F1, on his debut he was out qualified by Heidfled by 0.5s and his debut half season against the talent that was Liuzzi he out qualified him by 0.06s but lost the head to head 4-3, this being dry qualifying.

Regarding the Hulk I had Barrichello beating him by 0.26s, Button beat Perez by 0.06s, whilst the Hulk and Perez during their 3 years together were quite evenly matched..

[Invade]

How much can be read into any of this stuff?

Alonso-Hamilton and Alonso-Button were close enough, but Hamilton had a significantly bigger edge over Button. It doesn't seem like much of a consistent barometer for anything. So what's the deal?

[Zoue]

Ignoring available information is called a flawed sample.

depends whether you want the results to be honest or not. It also depends on what you are looking to measure. So if you want to measure Vettel against Kimi, then it's right to compare their whole time together. But if you want to compare how both Vettel and Alonso fared against Kimi, then it's only fare to make the comparison as equal as possible. Which means taking the most equal sample which is a close in size to each other as possible

How you deal with numbers, ignoring data, is not a honest way with dealing with numbers at all, plus I would say that you're not really interested in crunching numbers yourself just in disproving data that give results that you don't like.

I dare say if you swapped 2015 and 2016 around then you would come up that 2016 was more representative because Vettel was new to the team in 2015.

Actually it's the only way to make a truly honest comparison, to ensure that the date for each is as similar as possible.

My interest in crunching numbers is irrelevant. This isn't a contest to see who is more interested in data. But if you want to make a fair comparison then you don't need to be a numbers geek to see what's there.

I don't see why you'd swap the two years around. That would be playing with figures, whereas taking the same volume of data from the same period - ie 1st year together - is the only way to ensure the figures are as fair to both as possible.

[Zoue]

Ignoring available information is called a flawed sample.

depends whether you want the results to be honest or not. It also depends on what you are looking to measure. So if you want to measure Vettel against Kimi, then it's right to compare their whole time together. But if you want to compare how both Vettel and Alonso fared against Kimi, then it's only fare to make the comparison as equal as possible. Which means taking the most equal sample which is a close in size to each other as possible

How you deal with numbers, ignoring data, is not a honest way with dealing with numbers at all, plus I would say that you're not really interested in crunching numbers yourself just in disproving data that give results that you don't like.

I dare say if you swapped 2015 and 2016 around then you would come up that 2016 was more representative because Vettel was new to the team in 2015.

Why is it that with Vettel the representative year is always the abnormal one?

Ask Zoue I don't deal in representative years, I use all the years available.

That may be the problem right there. You don't deal in years that are equally representative to both drivers. I would argue that you are guilty of the same accusation you levelled at me, that you are only keen on the data that shows what is your own preference.

[pokerman]

Ignoring available information is called a flawed sample.

depends whether you want the results to be honest or not. It also depends on what you are looking to measure. So if you want to measure Vettel against Kimi, then it's right to compare their whole time together. But if you want to compare how both Vettel and Alonso fared against Kimi, then it's only fare to make the comparison as equal as possible. Which means taking the most equal sample which is a close in size to each other as possible

How you deal with numbers, ignoring data, is not a honest way with dealing with numbers at all, plus I would say that you're not really interested in crunching numbers yourself just in disproving data that give results that you don't like.

I dare say if you swapped 2015 and 2016 around then you would come up that 2016 was more representative because Vettel was new to the team in 2015.

Actually it's the only way to make a truly honest comparison, to ensure that the date for each is as similar as possible.

My interest in crunching numbers is irrelevant. This isn't a contest to see who is more interested in data. But if you want to make a fair comparison then you don't need to be a numbers geek to see what's there.

I don't see why you'd swap the two years around. That would be playing with figures, whereas taking the same volume of data from the same period - ie 1st year together - is the only way to ensure the figures are as fair to both as possible.

That makes no sense but I guess yet again that we have to agree to disagree.

[pokerman]

depends whether you want the results to be honest or not. It also depends on what you are looking to measure. So if you want to measure Vettel against Kimi, then it's right to compare their whole time together. But if you want to compare how both Vettel and Alonso fared against Kimi, then it's only fare to make the comparison as equal as possible. Which means taking the most equal sample which is a close in size to each other as possible

How you deal with numbers, ignoring data, is not a honest way with dealing with numbers at all, plus I would say that you're not really interested in crunching numbers yourself just in disproving data that give results that you don't like.

I dare say if you swapped 2015 and 2016 around then you would come up that 2016 was more representative because Vettel was new to the team in 2015.

Why is it that with Vettel the representative year is always the abnormal one?

Ask Zoue I don't deal in representative years, I use all the years available.

That may be the problem right there. You don't deal in years that are equally representative to both drivers. I would argue that you are guilty of the same accusation you levelled at me, that you are only keen on the data that shows what is your own preference.

Using all the data for all the drivers is not showing personal preference it's simple how you do this kind of thing, however wanting to use selective years that would be indicative of a personal preference.

[Lotus49]

Ignoring available information is called a flawed sample.

depends whether you want the results to be honest or not. It also depends on what you are looking to measure. So if you want to measure Vettel against Kimi, then it's right to compare their whole time together. But if you want to compare how both Vettel and Alonso fared against Kimi, then it's only fare to make the comparison as equal as possible. Which means taking the most equal sample which is a close in size to each other as possible

How you deal with numbers, ignoring data, is not a honest way with dealing with numbers at all, plus I would say that you're not really interested in crunching numbers yourself just in disproving data that give results that you don't like.

I dare say if you swapped 2015 and 2016 around then you would come up that 2016 was more representative because Vettel was new to the team in 2015.

Actually it's the only way to make a truly honest comparison, to ensure that the date for each is as similar as possible.

My interest in crunching numbers is irrelevant. This isn't a contest to see who is more interested in data. But if you want to make a fair comparison then you don't need to be a numbers geek to see what's there.

I don't see why you'd swap the two years around. That would be playing with figures, whereas taking the same volume of data from the same period - ie 1st year together - is the only way to ensure the figures are as fair to both as possible.

That makes no sense but I guess yet again that we have to agree to disagree.

What doesn't make sense? If one guy gets 3 seasons to come to grips with his team mate and another guy gets just one then that doesn't sound a very like for like comparison as the first guy gets so many more races to assert his authority and the odd race can make a big difference when you use the average gap.

Taking the first year for both sounds fair enough to me, they're both in the same position with the same amount of time to get to grips with their team mate. You're basically just imagining the contest ended at the same time for both, no advantages to either one. (Apart from who's new to the team and who isn't if that's the case)

[pokerman]

Seeing has it's very much on topic and in respect to recent posts got me curious with driver comparisons and cross references:-

Vettel > Webber 0.19s
Vettel > Kimi 0.2s

Therefore Webber = Kimi = Massa


Then:-

Alonso > Massa 0.27s
Alonso > Kimi 0.28s

Therefore Alonso > Vettel 0.08s

That's my twopenneth.

[Lotus49]

How you deal with numbers, ignoring data, is not a honest way with dealing with numbers at all, plus I would say that you're not really interested in crunching numbers yourself just in disproving data that give results that you don't like.

I dare say if you swapped 2015 and 2016 around then you would come up that 2016 was more representative because Vettel was new to the team in 2015.

Why is it that with Vettel the representative year is always the abnormal one?

Ask Zoue I don't deal in representative years, I use all the years available.

That may be the problem right there. You don't deal in years that are equally representative to both drivers. I would argue that you are guilty of the same accusation you levelled at me, that you are only keen on the data that shows what is your own preference.

Using all the data for all the drivers is not showing personal preference it's simple how you do this kind of thing, however wanting to use selective years that would be indicative of a personal preference.

No it isn't, it's using the same conditions rather than favouring or even hindering the guy who had multiple years and many more chances to improve or hurt his average.

Using the same amount seems fair to me. In a case of one battle being just one year and comparing it to a 3yr battle, If you were just randomly picking what year to use rather than using the same first year for both I could see the issue but that's not being put forward I don't believe.

[pokerman]

depends whether you want the results to be honest or not. It also depends on what you are looking to measure. So if you want to measure Vettel against Kimi, then it's right to compare their whole time together. But if you want to compare how both Vettel and Alonso fared against Kimi, then it's only fare to make the comparison as equal as possible. Which means taking the most equal sample which is a close in size to each other as possible

How you deal with numbers, ignoring data, is not a honest way with dealing with numbers at all, plus I would say that you're not really interested in crunching numbers yourself just in disproving data that give results that you don't like.

I dare say if you swapped 2015 and 2016 around then you would come up that 2016 was more representative because Vettel was new to the team in 2015.

Actually it's the only way to make a truly honest comparison, to ensure that the date for each is as similar as possible.

My interest in crunching numbers is irrelevant. This isn't a contest to see who is more interested in data. But if you want to make a fair comparison then you don't need to be a numbers geek to see what's there.

I don't see why you'd swap the two years around. That would be playing with figures, whereas taking the same volume of data from the same period - ie 1st year together - is the only way to ensure the figures are as fair to both as possible.

That makes no sense but I guess yet again that we have to agree to disagree.

What doesn't make sense? If one guy gets 3 seasons to come to grips with his team mate and another guy gets just one then that doesn't sound a very like for like comparison as the first guy gets so many more races to assert his authority and the odd race can make a big difference when you use the average gap.

Taking the first year for both sounds fair enough to me, they're both in the same position with the same amount of time to get to grips with their team mate. You're basically just imagining the contest ended at the same time for both, no advantages to either one. (Apart from who's new to the team and who isn't if that's the case)

There's actually good cross references to all of this as well but we want 1 year to be defining and all this so Vettel can be that bit closer to Alonso, a 1 year reference point = 0.04s, years of reference = 0.08s.

At the end of the day I don't really care which one is quicker and if anything on a personal level I prefer Vettel, this is just numbers driven to find an answer to who is quicker as some have asked in this thread and then guessed what the difference between Alonso and Vettel might be and it seems that most think that Alonso is a little bit quicker and so do the numbers.

[Lotus49]

How you deal with numbers, ignoring data, is not a honest way with dealing with numbers at all, plus I would say that you're not really interested in crunching numbers yourself just in disproving data that give results that you don't like.

I dare say if you swapped 2015 and 2016 around then you would come up that 2016 was more representative because Vettel was new to the team in 2015.

Actually it's the only way to make a truly honest comparison, to ensure that the date for each is as similar as possible.

My interest in crunching numbers is irrelevant. This isn't a contest to see who is more interested in data. But if you want to make a fair comparison then you don't need to be a numbers geek to see what's there.

I don't see why you'd swap the two years around. That would be playing with figures, whereas taking the same volume of data from the same period - ie 1st year together - is the only way to ensure the figures are as fair to both as possible.

That makes no sense but I guess yet again that we have to agree to disagree.

What doesn't make sense? If one guy gets 3 seasons to come to grips with his team mate and another guy gets just one then that doesn't sound a very like for like comparison as the first guy gets so many more races to assert his authority and the odd race can make a big difference when you use the average gap.

Taking the first year for both sounds fair enough to me, they're both in the same position with the same amount of time to get to grips with their team mate. You're basically just imagining the contest ended at the same time for both, no advantages to either one. (Apart from who's new to the team and who isn't if that's the case)

There's actually good cross references to all of this as well but we want 1 year to be defining and all this so Vettel can be that bit closer to Alonso, a 1 year reference point = 0.04s, years of reference = 0.08s.

At the end of the day I don't really care which one is quicker and if anything on a personal level I prefer Vettel, this is just numbers driven to find an answer to who is quicker as some have asked in this thread and then guessed what the difference between Alonso and Vettel might be and it seems that most think that Alonso is a little bit quicker and so do the numbers.

I think it's just a reasonable way, taking the battle in the same circumstances but I agree it's not that important. There are so many different ways of doing it, as seen in this thread, that most of the actual numbers and cross comparison numbers seem to get ignored in preference of your gut feeling.(Not yours personally to be clear)

[Paolo_Lasardi]

Ignoring available information is called a flawed sample.

depends whether you want the results to be honest or not. It also depends on what you are looking to measure. So if you want to measure Vettel against Kimi, then it's right to compare their whole time together. But if you want to compare how both Vettel and Alonso fared against Kimi, then it's only fare to make the comparison as equal as possible. Which means taking the most equal sample which is a close in size to each other as possible

First: But that does not imply to ignore part of Alonso's results against Räikkönen - incidentally those with the biggest margin of FA over KR - without doing the same with SV vs KR. Which is what KingVoid is doing ....

Second: only if the different size systematically biases the result (in contrast to: I don't like the results then), i.e. that length of the dataset is a causal factor for the result, i.e. the number of races as teammates causally influences the relative performance to each other.

[Zoue]

Ignoring available information is called a flawed sample.

depends whether you want the results to be honest or not. It also depends on what you are looking to measure. So if you want to measure Vettel against Kimi, then it's right to compare their whole time together. But if you want to compare how both Vettel and Alonso fared against Kimi, then it's only fare to make the comparison as equal as possible. Which means taking the most equal sample which is a close in size to each other as possible

First: But that does not imply to ignore part of Alonso's results against Räikkönen - incidentally those with the biggest margin of FA over KR - without doing the same with SV vs KR. Which is what KingVoid is doing ....

Second: only if the different size systematically biases the result (in contrast to: I don't like the results then), i.e. that length of the dataset is a causal factor for the result, i.e. the number of races as teammates causally influences the relative performance to each other.

If you take the same dataset for each, you are much less likely to have anything unduly influencing in favour of either driver

[Paolo_Lasardi]

Ignoring available information is called a flawed sample.

depends whether you want the results to be honest or not. It also depends on what you are looking to measure. So if you want to measure Vettel against Kimi, then it's right to compare their whole time together. But if you want to compare how both Vettel and Alonso fared against Kimi, then it's only fare to make the comparison as equal as possible. Which means taking the most equal sample which is a close in size to each other as possible

First: But that does not imply to ignore part of Alonso's results against Räikkönen - incidentally those with the biggest margin of FA over KR - without doing the same with SV vs KR. Which is what KingVoid is doing ....

Second: only if the different size systematically biases the result (in contrast to: I don't like the results then), i.e. that length of the dataset is a causal factor for the result, i.e. the number of races as teammates causally influences the relative performance to each other.

If you take the same dataset for each, you are much less likely to have anything unduly influencing in favour of either driver

Only if these "influences" are a function of the length of the datset, i.e. of the number of datapoints (because that's the only "same" thing here).