Mitch's Poll - Wood Coaster Results

And 14 riders of 629 (a little over 2%) unanimously choosing a given coaster as the best they've ever ridden is a pretty solid indicator.

That's not how I see it, especially if the people comprising that 2% happened to ride the coaster at the same time, perhaps a TPR trip. Those people experienced the coaster around the same time, when it could have been running very well. Coasters ridden by a lot more than 2% of voters likely saw riders over a very long period of time- some of them rode the coasters on good days and some of them rode the coasters on bad days. The way those coasters are ranked indicates the average opinion when positive and negative experiences are factored into the equation. It's probably the reason why Phoenix sits in the six spot.

Is T-Express the best wood roller coaster in the world? I'm sure it's possible, but also unlikely. We'll visit this conversation in a few years, around the time we see how The Voyage, El Toro and Troy have aged. Remember the time when Shivering Timbers was the ultimate wood coaster?

I still feel that Mitch's poll is far and away the best indicator of the opinions of enthusiasts in general and a great starting point for finding the 'good' rides out there.

Definitely. The point of my post is to show that the poll is flawed, but still the best way to see how well-traveled people view all the available options.

T Express may very well be the best woodie in the world. From all reports it is a very good ride. But the fact that it took the #1 spot on Mitch's poll is clear evidence that the poll is not a valid sampling of the entire enthusiast community.

The poll is supposed to be based on personal experience, if I understand correctly, and there is no way that if you were to survey the entire community that even 5% of the enthusiasts have made it to Korea. In fact, it's probably less than 1%.

Even if every single enthusiast who visits Korea gives T Express the top ranking, it shouldn't have the statistical weighting to compete with much more highly ridden coasters like Voyage and El Toro.

As has been said, polls are no more than popularity contests, snapshots frozen in time. But these results are the clearest proof that Mitch's poll isn't even an accurate popularity contest.

Re Rob: But Shivering Timbers has gotten noticably worse over time (from what I understand, I have never ridden it). It's not that people's opinion of the same experience changed. It's that the expereince changed.

Re: Ensign

I think you're missing the point of Mitch's poll. It's not supposed to be a popularity contest, and so you are exactly right - it isn't an accurate one. People who have not ridden t-ex cannot have an opinion of it. Therefore the only opinions that matter when it comes to T-express...are the people who rode it. You could say the sample size is too small, but you can't say that El Toro or Voyage should be on top BECAUSE they have more riders.

Ensign -
I mean no offense whatsoever, but your statements seem to suggest that either you don't understand the poll, or you just don't agree with what it's "supposed" to be.

The whole point of Mitch's poll is that it's not a popularity contest. It's not supposed to matter how many people have ridden one ride vs another; ridership numbers should have no effect on the outcome of the poll. If 15 people ride T Express, and 100% of them rank it above both Voyage and El Toro, that's not significantly significant enough for you? What would be statistically significant?

We can argue whether 15 (out of 600 or so ballots) is enough for a ride to compete/qualify in the poll, or whether it's accurate for a single group (TPR in this case) to have the power to decide the number one ride, but it's not a problem with the methods he uses to calculate the rankings.

-Nate

But the bias moves from popularity to the relative "seasoning" of these nuts who travel all over the world to ride. In past years, rides with two or three people riding would propel a ride to the top. Two people are not ever statistically relevant.

The theory behind the poll is absolutely solid, but it still requires a statistically relevant sample.

I can't seem to access the site right now, but I believe Mitch requires that a ride has 10 riders in order to be assigned a spot. I think that number is still probably too low, but I think it's a matter of finding a compromise between accuracy and not excluding less-ridden coasters.

-Nate

I was never on {ride x}, there is no way it could be that good! It's not like Mean Streak is #1!

Right. And I am arguing that 2% of the voting population does not qualify as being statistically relevant, especially when the ride is brand new and it's likely the majority of those people got to experience the ride at the same time.

But Shivering Timbers has gotten noticably worse over time (from what I understand, I have never ridden it).

Totally an aside, but comparing the Timbers I rode this past summer to that of a few years prior, it's gotten better again---I'm not sure why, if I just got a good day, or what. Still might not be up to its onetime reputation, but a darn good ride, for sure.

matt. said:
Gonch took care of this very well above, but I think he kinda buried his best point - if you feel like something has been skewed by a small sample size, then just ignore that part of the results. It doesn't really invalidate the rest of the poll.

Yeah, I did kinda bury that and it's a great point.

So if I eliminate T Express in my mind, I'm forced to look at Voyage and Toro as the 1 and 2 spots.

What I immediately see is that they both have a winning percentage that's within 1/10,000th of each other and that El Toro's is actually the higher of the two, but in the head-to-head Voyage eeeked out a few more wins over Toro. 4 to be exact...and I officially declare 4 to be an statistically insiginificant number. So these two cancel each other.

#4 is Boulder Dash, but I'm calling it #1. ;)

I dunno. Still looks like far and away the best way to do a poll that tries to capture the general opinions of coaster enthusiasts. How many people have to have ridden a coaster before their opinion matters?

I think it's good to have even some people to ask for comparisons rather than none. The only thing less significant than too small of a sample is no sample at all.

We have this discussion/argument every year. It's a poll, people. It's not law. So what if T-Express is #1? That doesn't negate what any one else thinks is #1.

Time for some people to get over themselves....

Martin Valt said:Er.....no offence intended, but which bit of El Toro could possibly be described as pointless meandering? :)

LOL, none taken. Ever. Hehe, but anyhow, the ride overall is absolutely stunning, a "tour de forces", so to speak. But the infield section really needed (IMO) a little more elevation change. When the ride is slower, that infield portion reminds me too much of the ending on Cheetah - a small *pop* of air coming out of that tangled trackage would have been a major improvement. Regardless - would I take Toro over virtually every other coaster out there? Heck yeah! There's only one wooden coaster that I *rank* over Toro...but even Voyage could be enhanced. Perfectionism means never having to say "as good as it can be". ;)

This drama makes our drama look downright reasonable. :)

One of the great things about this "poll" is so many coasters which have gone unnoticed before now find thmeselves being recognized for the "international" excellence. Who ever would have known about Eagle's Fortress and all of it's awesomeness.

There's more to life than Cedar Point, more to life than what's happening in the US, and this poll finally reveals it. That's really the BIG elephant in the room that no one wants to talk about. Finally there is a respected poll out there that focuses on more than just the same 3 or 4 parks every year.

Mamoosh said:
We have this discussion/argument every year. It's a poll, people. It's not law. So what if T-Express is #1? That doesn't negate what any one else thinks is #1.

As best as I can tell, that's not what is being discussed. So who exactly needs to "get over themselves?"

Polls of any kind to me are fun distractions, sure, but my personal fascination is how to make one "accurate" or "better" in some objective way. That's fun when you consider the subjective nature of any poll.

I never took a statistics class in college but I've married/dated enough scientists to take away some general guidelines. The bottom line is that a small sample, no matter how you slice it up, creates a bias in the survey that reduces its relevance.

Comparing a ride that 14 people like to a ride that hundreds like ignores the environmental biases that can skew the smaller group. If it just so happens that those people went together or were just so euphoric about going to Southeast Asia that they could only see rainbows and puppies, then that doesn't represent a broad spectrum of experiences the way that, say, that hundreds of people have had going to Cedar Point or even Holiday World.

So if you wanted to be scientific about it, you have to adjust and compensate the smaller opinion based on some much bigger statistics like demographics or preferences toward whatever. I can't even begin to suggest what useful attributes would be worth adjusting on, and even then, the likelihood that enough people out of 14 can match the adjustment factors is not very high.

I like to bring this up just because of the annual suggestion that it's a "pretty good way" to conduct a poll. And yes, if the sample were bigger, it would be. But with these sample sizes, it's simply not.

Two years from now, Voyage will be right back where it belongs. TPR is the same group that got served Beer for free at one park and ranked it over voyage even though it craws thru the turns.No Offense to T Express. I need a much bigger diverse sample than 6 of 9 verses over 250 people..The good ones and well maintained ones keep their status. Even though on this list. 50th or in some cases 100th could be someones personal favorites.

Jeff said:
I never took a statistics class in college but I've married/dated enough scientists to take away some general guidelines. The bottom line is that a small sample, no matter how you slice it up, creates a bias in the survey that reduces its relevance.

So at what point does that bias no longer exist?

There's has to be an absolute answer. I can't believe we don't have a certified numbers dork around here who can give us an actual scientific answer.

I like to bring this up just because of the annual suggestion that it's a "pretty good way" to conduct a poll. And yes, if the sample were bigger, it would be. But with these sample sizes, it's simply not.

Just for the sake of clarity, do you feel that the sample size of certain rides is too small (like the 14 T Express riders) compared to the overall number of voters or that the overall number of voters (629) is too small to accurately represent the opinions of the enthusiast community?

As far as it being a "pretty good way" - given the importance of the subject (or lack thereof), the amount of data combed through and the number of people sampled, I'd say it's a friggin outstanding way to judge which coasters people enjoy the most during any given calendar year.

Steel results are online now too.

Let the bitching continue! :)

Jeff said:So if you wanted to be scientific about it, you have to adjust and compensate the smaller opinion based on some much bigger statistics like demographics or preferences toward whatever. I can't even begin to suggest what useful attributes would be worth adjusting on, and even then, the likelihood that enough people out of 14 can match the adjustment factors is not very high.

Since Mitch collects data on "personal preferences" like where "extreme airtime" rates for you....seems like a statistician with some free time could actually analyze the data by specific taste preferences and compare them to yours in roder to give you a personalized top ten list - eHarmony or whatever, only for coasters. Then you might decide to go coaster-tripping to Scandinavia or Japan or Germany or even Korea, based on the results. ;)

Ensign Smith said:
Even if every single enthusiast who visits Korea gives T Express the top ranking, it shouldn't have the statistical weighting to compete with much more highly ridden coasters like Voyage and El Toro.

This tells me you fundamentally don't understand the process and purpose of the poll. You'd probably be better off checking out the Golden Tickets or something like that if you'd prefer coasters to rank higher simply because they're in OH, PA, or FL.