Lost leads and inherited wins of the last 10 years

[bhall]

My suggestion: If you can't stomach peer review, you might not want to present your views on an open forum.

[SectorOne]

Concluding who comes out on top is no explanation of an issue. Is the FIA championship biased because Vettel came out on top for the last 3 years? Are the team-bosses biased for Vettel's rankings? Or is it basic sense that the top drivers come out on top of most stats?

It is the same 4 or 5 people on here whose only interest is disruptive behavior towards people that are interested in a range of current, relevant topics. If you think it is "completely random" to look at lost leads after a race where two leaders had a mechanical failure, I can only facepalm your reasoning skills. However, I am not sharing this for the disruptive trolls who ridicule the people who are interested, but all the people that find it interesting, and all the people saying "thanks, I was just thinking about that".

If you are too narrow-minded to understand that people have an interest in something that has just happened, I can only advice you to try and open up a little and try fewer self-centered thoughts.

Trolls, delusional, arrogant, narrow-minded.. I applaud you for saying them but not directly saying them.
Putting in the word "if" releases you from any consequences of it which is clever.

Is the FIA championship biased because Vettel came out on top for the last 3 years?"

No obviously not. But it also has nothing to do with the obvious here, which you deny and which is fine but just so you know, it´s quite obvious, "all you need to do is look at the stats".
Or are my stats not valid but yours are?

It is the same 4 or 5 people on here whose only interest is disruptive behavior towards people that are interested in a range of current, relevant topics.

disruptive? some of us just pointed out some things. If you can´t take any critique and instead assume people are arrogant, narrow-minded, delusional or trolls then maybe that´s something to work on for next time.

If you think it is "completely random" to look at lost leads after a race where two leaders had a mechanical failure, I can only facepalm your reasoning skills.

yes i think this one was probably the most random one ever. It´s not exactly the first time a driver (or Vettel) lost the lead due to failures.
Maybe the stats wasn´t in your favor back then, who knows.

However, I am not sharing this for the disruptive trolls who ridicule the people who are interested, but all the people that find it interesting, and all the people saying "thanks, I was just thinking about that".

So basically anyone that does not use critical thinking when reading stuff.
Just the "way-to-go-sayers"

Imagine if scientists worked like that.

[Blanchimont]

Why not simply look at the retirement/race starts ratio?

Code: Select all

Vettel        18/109 = 0,165
Alonso        33/205 = 0,161
Hamilton      16/118 = 0,136
Raikonnen     42/184 = 0,228

Hakkinen      60/161 = 0,373
Schumacher    65/306 = 0,212
Senna         53/161 = 0,329
Prost         55/198 = 0,278

[SectorOne]

Because it´s the wrong results

Are the team-bosses biased for Vettel's rankings? Or is it basic sense that the top drivers come out on top of most stats?

This one is interesting,

Thread is called, "Alonso voted driver of the year by team bosses"
And that´s for 2012.

If you had only included what the thread title said, Alonso would have been on top. But you didn´t.
Instead you compiled the last 5 years so that Vettel would come out on top in two more lists.
(actually it´s three)

And other people noticed this too as the very first comment goes like this,

It was a nice post,until statistics came forth to bring Vettel back to #1 yet again...

Maybe we are all delusional, arrogant, narrow-minded, trolls etc but you my friend are in denial of the obvious.

[beelsebob]

Mark Twain is not a scientist, engineer etc.
Quite a bit of the fundamental physics theories have been developed or proved using statistical methods. It is a tool as long as people use it correctly it is very useful one for that matter. People who tries to mislead other with statistical data will be called for by other smart, knowledgeable people.

I may have missed your point, but I think that you've missed Mark Twain's point. It's not that stats are useless. It's that stats, when abused, can be used to "prove" any "fact" you like. They're incredibly easy to lie with, because your methodology must be scrupulous to get a result out that means anything worth a damn. The reason he's being quoted here is not because there are stats, therefore they're wrong. He's being quoted because the stats have a methodology with significant flaws, and hence show nothing useful at all.

It may be true that Vettel is on average less lucky than other drivers with mechanical failures... It may not. The stat doesn't say, because it's methodology is wrong.

And for those going on about "zomg, let people have opinions". This thread isn't about opinion. A stat is well founded or it isn't. The methodology here is flawed if you want to get a useful conclusion out. You're welcome to your opinion about Vettel as a driver, but that doesn't make this thread make any more sense.

[LionKing]

It may be true that Vettel is on average less lucky than other drivers with mechanical failures... It may not. The stat doesn't say, because it's methodology is wrong.

And for those going on about "zomg, let people have opinions". This thread isn't about opinion. A stat is well founded or it isn't. The methodology here is flawed if you want to get a useful conclusion out. You're welcome to your opinion about Vettel as a driver, but that doesn't make this thread make any more sense.

Please point out the significant flaws of his methodology then....
Maybe it is not perfect I do not see any major problem with it

The thread's title is "Lost leads and wins of last 10 years" Nothing more nothing less. It is you guys that is trying to make assumption about his intent, or basically complain about why this stat does not answer other specific questions or is not a good measure for another etc.

Did he say that his list and aggregate results are an estimate for the reliability problems each driver has faced or any other particular thing such as the estimate of the luck of driver. He has just tried to find out answer to a PARTICULAR question and worked towards that.

There is this nonsense things like bias, or "sampling bias" are thrown without thinking much about it. Bias will be present if there is a difference between what metric he wants to estimate and what he is actually measuring. If he had just compiled the wins lost due to mechanical problems and has given this number as an estimate for the overall reliability of the cars, or the reliability problems faced by the drivers then he would have a major problem. Because then reliability of cars that are usually at the front will be underestimated and for others it will be overestimated. I don't think he has said something along this line at all.

Once again, I don't think there is a problem with his stats, more but more with misinterpretations of people.

[beelsebob]

It may be true that Vettel is on average less lucky than other drivers with mechanical failures... It may not. The stat doesn't say, because it's methodology is wrong.

And for those going on about "zomg, let people have opinions". This thread isn't about opinion. A stat is well founded or it isn't. The methodology here is flawed if you want to get a useful conclusion out. You're welcome to your opinion about Vettel as a driver, but that doesn't make this thread make any more sense.

Please point out the significant flaws of his methodology then....
Maybe it is not perfect I do not see any major problem with it

The thread's title is "Lost leads and wins of last 10 years" Nothing more nothing less. It is you guys that is trying to make assumption about his intent, or basically complain about why this stat does not answer other specific questions or is not a good measure for another etc.

Did he say that his list and aggregate results are an estimate for the reliability problems each driver has faced or any other particular thing such as the estimate of the luck of driver. He has just tried to find out answer to a PARTICULAR question and worked towards that.

There is this nonsense things like bias, or "sampling bias" are thrown without thinking much about it. Bias will be present if there is a difference between what metric he wants to estimate and what he is actually measuring. If he had just compiled the wins lost due to mechanical problems and has given this number as an estimate for the overall reliability of the cars, or the reliability problems faced by the drivers then he would have a major problem. Because then reliability of cars that are usually at the front will be underestimated and for others it will be overestimated. I don't think he has said something along this line at all.

Once again, I don't think there is a problem with his stats, more but more with misinterpretations of people.

Okay, well, now we get to the opinion part... I don't think that this is an interesting stat, and relevant to a technical forum? I don't think it's an interesting stat for the same reason as "Who has won the most times in a yellow and blue car" is not an interesting or relevant stat. I don't think any reasonable, informative, or technical conclusion can be drawn from this stat.

Which informative, technical conclusion do you think can be drawn from it, or do you think it's interesting and relevant to a technical forum for a different reason?

[LionKing]

One more thing.... statistics is powerful tool if used correctly:

There is this guy named Nate Silver. He was spot on at the 2012 presidential elections. He did not do any separate polls but just used the results of published public polls to aggregate with his model. He predicted 50 of the 50 states correctly. On the other hand big firms like Gallup, Rasmussen screwed up...

[SectorOne]

Presidential election is easy. The guy with the most royal ancestors wins every time.
And that all presidents but one can be traced back to one English King, John Lackland Plantaganet.

I think that something like 25 presidents also are linked to each other.
Obama for example is related to Bush.

[Stradivarius]

I have often been confronted with Mark Twain's famous quote about statistics and I understand why many people share his opinion about statistics, but I must agree with LionKing, as statistics is nothing more than a tool that, when used correctly, is very useful at describing almost anything. Statistics never lie, but they must be interpreted by qualified people in order to add any useful information. The most common problem with statistics is that the people interpreting the data don't know what they are doing. In this specific case, there is nothing wrong about stating that Vettel has retired from the lead more often than other drivers. But if someone interprets this and concludes that Vettel is cursed with bad luck, the interpreter is making false claims. But that is not the statistics lying, it is simply the interpretation of the data which is faulty. In fact, I think it is fair to distinguish between statistics and data, where data is simply some raw material (numbers), while statistics involves a correct treatment/interpretation of the data in question in order to throw light on the questions of the study.

My statistic's professor at university once said that almost 10% of all people ever born in the world, are still alive today. He then made it very clear that this does not mean that there is only a probability of 90% that any given human being will ever die, although some would say that it is statistically true. I wouldn't say that, because in my opinion the word "statistically" involves more than just looking at data, it also involves interpreting the data in a reasonable way.

There is one very common mistake in statistics which is committed both intentionally and unintentionally, and that has to do with sampling errors. The text book example of this is how "The Literary Digest" tried to predict the outcome of the presidential election in USA in 1936, but won't discuss this further here, other than saying that this was an unintentional sampling error (combined with nonresponse bias).

The intentional sampling error involves first studying data and then finding a trend that support your view, and then present this data as if it was arbitrarily chosen. I am not claiming that this is the case here and I do not intend to accuse anyone, but accusations have already been made here, which I think makes a very illustrative example: Let's say you want to make it look like Vettel is very unlucky in general, after seeing him loosing the lead at Silverstone. In the hunt for proof, the first option is then probably to look at the number of retirements and compare it to other drivers. But if you do that, you will find that Vettel isn't particularly unlucky with respect to retirements. But then you discover that Vettel has lost the lead many more times than anyone else due to retirements. And then you choose to present this data as if you that is what you were looking for, while what you in fact was looking for was anything that would support your predetermined conclusion and then you simply presented that piece of information, while excluding everything else.

This is how many journalists work. They want to write about something spectacular, for example that carrots are not healthy. Then they look at many studies and discover that most of them conclude that carrots are healthy. But then they stumble over one report that arrives at the opposite conclusion. And then they simply refer to this report in their article, without mentioning that most studies disagree with this one. After all, the level of significance in research is often 5% or higher, so if sufficiently many studies are performed, you will sooner or later find one that arrives at the opposite conclusion to the "truth". Another well known method used in the media is to ask people on the street a particular question. Maybe the first 3 people you ask express an opinion that the journalist doesn't want to come through. But the fourth one says exactly what you want to show. And then you present this person's response on the news, as if it was just the opinion of some random guy on the street.

I do not want to accuse mnmracer of having done any of the above mentioned "sins". I actually agree that it might be interesting to some to see who have lost the race lead as a result of a retirement through history. But I also tend to agree with beelsbob that this is hardly a technical topic, (beyond the fact that it involves technical failures). Whether or not the topic is justified here is not for me to decide, but I would at least find it more interesting to look at the complete picture. If you are looking at technical failures, why limit the scope only to race leaders?

[raymondu999]

Yeesh. You guys are taking this way too seriously. This stat is what you call a "fun fact" stat - something not meant for conclusion-drawing, nor for any inferences.

It's just a way to pull on our heart strings because naturally, as humans, we will oftentimes have in our hearts a "moral winner" that isn't the "real" winner of the day.

[Steven]

I have sadly had to remove some posts as this had become a pissing match

Statistics are what they are, and it is often said that you can use it to prove anything you want. So far, this thread has only been a tirade about this being a pro-vettel statistic, and while it may be, I have not seen anyone find any mistake in the stat, or at least nobody mentioned one. That makes it a decent stat, doesn't it? Or is it just more fun to bash being for being suppotive to some driver?

There is this other rule: "what you do yourself, you usually do better". Anyone is free to post statistics they want and get feedback on them, but whining about a post from someone else is of course a bit easier than compiling a stat yourself...

Match done.

PS: I have also removed ALL votes in this topic, as they're all I agree / I disagree votes, rather than the result of using the system to its purpose