New Feature: Cache Survey/ratings

[+ReadyOrNot]

There's been a lot of talk in the forums about the numbers game and those evil micro caches (personally, I love micro's that are well hidden and bring me somewhere interesting). The survey/rating system would create social change by rewarding those that take the time to hide well thought-out, interesting caches. But it doesn't hurt those that still like to hunt micros (like me!).

 

Whenever I use words, I always seem to screw up, so I'll let the pictures do the explanation for me. Let me know what you think????

 

The cacher is not forced to review the cache, but is given the option

 

If they choose to, they are put here after logging their find to complete the survey.

 

This is one of many possibilities for interpreting the results of the survey.

[+StarBrand]

Probably too easy to manipulate. If you, your sockpuppet and 5 friends are all first to rate the cache, it will be a while before the true rating comes to light. A group could aslo manipulate things negatively. Also, your sense of what is good is far different than mine so I firmly believe that given enough time and logs that all but the very very best and worst would all have a rating of 5 - +/- 1. So that won't help anybody. Also the nature of many hides shifts a bit over time due to a number of factors so caches rated high a year ago just might not be the same anymore. Etc..........

[+ReadyOrNot]

Probably too easy to manipulate. If you, your sockpuppet and 5 friends are all first to rate the cache, it will be a while before the true rating comes to light. A group could aslo manipulate things negatively. Also, your sense of what is good is far different than mine so I firmly believe that given enough time and logs that all but the very very best and worst would all have a rating of 5 - +/- 1. So that won't help anybody. Also the nature of many hides shifts a bit over time due to a number of factors so caches rated high a year ago just might not be the same anymore. Etc..........

 

A high rating in one area is not necessarily good or bad. I wouldn't mind hunting a micro for instance, if there was a scenic view or scenic hike associated with it. Some people are looking for the quick park and grab.. It's been a while since I took statistics, but if a place has a nice view, it's not going to average to a 5 over-time, likewise if the view stinks. Over time, it's going to average out to a fairly accurate number.

 

I purposely called it a survey and not rating (well I guess I did write rating didn't I, well ignore that ) It's not a bad --> good rating, it's a survey of the cache.

Edited June 15, 2006 by ReadyOrNot

[+KC0GRN]

I'd have to disagree with a rating system.

 

My basic reason is this. There are a few cachers in my area that strongly dislike certain cachers and specific caches. Giving them a tool like this would be like begging them to virtually trash a cache. While the cache might be fun for people to do, what if the neutral cachers look at the page and see it's had a bad review, and decide not to go to it? They could be depriving themselves of a cache just because of some vindictive cacher.

 

In a more perfect world, that type system would be great, and people would learn to place nicer/more interesting caches. Unfortunately, we don't live in that kind of world.

[+ReadyOrNot]

I'd have to disagree with a rating system.

 

My basic reason is this. There are a few cachers in my area that strongly dislike certain cachers and specific caches. Giving them a tool like this would be like begging them to virtually trash a cache. While the cache might be fun for people to do, what if the neutral cachers look at the page and see it's had a bad review, and decide not to go to it? They could be depriving themselves of a cache just because of some vindictive cacher.

 

In a more perfect world, that type system would be great, and people would learn to place nicer/more interesting caches. Unfortunately, we don't live in that kind of world.

 

Referring back to statistics again... I forget what the term is, but there is a way to drop off the straglers from either side of the survey group and you are left with a nice average (10 people average 8 on scenic view, while 1 person votes 0 to be difficult, the 0 would be removed). Another possibility is to not show the results until a certain number of surveys are received. This along with dropping off the straglers in the group would eliminate a vindictive person from trashing a rating.

[+The Blind Acorn]

My theory is...if you dont like micros, don't look for them. If you dont like long hikes, don't go. It's not fair to bring criticism to a cache when someone at least goes through the effort of hiding it.

 

Groundspeak provided the ignore function awhile back, try using that to eliminate caches you don't intend on searching for.

[+ReadyOrNot]

My theory is...if you dont like micros, don't look for them. If you dont like long hikes, don't go. It's not fair to bring criticism to a cache when someone at least goes through the effort of hiding it.

 

Groundspeak provided the ignore function awhile back, try using that to eliminate caches you don't intend on searching for.

 

That sounds great on the surface, but it's not that simple. I'm sure the vast majority of cachers out there don't ignore every single micro on their list. There are micros with beautiful views and there are some micros with views of Walmart. For someone wanting to do a quick grab, they may not care about the view. And it's not just the view, there are many different criteria that a cache can be judged by. And the criteria is not good versus bad.

 

Some days I feel like doing a micro and some days I don't. If I ignored every micro out there, what good is that going to do me?

[+StarBrand]

 

Referring back to statistics again... I forget what the term is, but there is a way to drop off the straglers from either side of the survey group and you are left with a nice average (10 people average 8 on scenic view, while 1 person votes 0 to be difficult, the 0 would be removed). Another possibility is to not show the results until a certain number of surveys are received. This along with dropping off the straglers in the group would eliminate a vindictive person from trashing a rating.

 

So again - sooner or later most all caches would be hovering around 5 in all categories. And the first x number of finders would not have a rating to go by so might choose to ignore the cache - thus taking longer to see a rating. etc.....

[+ReadyOrNot]

 

Referring back to statistics again... I forget what the term is, but there is a way to drop off the straglers from either side of the survey group and you are left with a nice average (10 people average 8 on scenic view, while 1 person votes 0 to be difficult, the 0 would be removed). Another possibility is to not show the results until a certain number of surveys are received. This along with dropping off the straglers in the group would eliminate a vindictive person from trashing a rating.

 

So again - sooner or later most all caches would be hovering around 5 in all categories. And the first x number of finders would not have a rating to go by so might choose to ignore the cache - thus taking longer to see a rating. etc.....

 

Currently we don't even have a rating, so having to wait to get one seems like a moot point. You say again that most caches would be around 5 in all categories after a while. This makes absolutely no mathematical sense. If a cache has a nice view, the average would be high. If a cache has no view, the average would be extremely low. If a cache has a mediocre view, it would then hover around 5. Isn't this the desired result? The more people that rate the cache, the more accurate it is going to be...

 

Maybe giving people the option of a "N/A" instead of forcing to vote on every item?

 

EDIT: Regarding people ignoring a cache with no rating -- Typically the caches with no ratings are going to be fairly new caches. Being new caches, they are going to get visited more often than an old cache anyways.

Edited June 15, 2006 by ReadyOrNot

[+StarBrand]

 

Referring back to statistics again... I forget what the term is, but there is a way to drop off the straglers from either side of the survey group and you are left with a nice average (10 people average 8 on scenic view, while 1 person votes 0 to be difficult, the 0 would be removed). Another possibility is to not show the results until a certain number of surveys are received. This along with dropping off the straglers in the group would eliminate a vindictive person from trashing a rating.

 

So again - sooner or later most all caches would be hovering around 5 in all categories. And the first x number of finders would not have a rating to go by so might choose to ignore the cache - thus taking longer to see a rating. etc.....

 

Currently we don't even have a rating, so having to wait to get one seems like a moot point. You say again that most caches would be around 5 in all categories after a while. This makes absolutely no mathematical sense. If a cache has a nice view, the average would be high. If a cache has no view, the average would be extremely low. If a cache has a mediocre view, it would then hover around 5. Isn't this the desired result? The more people that rate the cache, the more accurate it is going to be...

 

Maybe giving people the option of a "N/A" instead of forcing to vote on every item?

 

EDIT: Regarding people ignoring a cache with no rating -- Typically the caches with no ratings are going to be fairly new caches. Being new caches, they are going to get visited more often than an old cache anyways.

I love the view - 10

I hate the view - 0

I think view is ok - 5

I love canyons - 10

I have been to too many canyons - 5

Just another desert - ho hum - 0

 

Average =5 (or thereabouts maybe +/- 1) I see this happening on many many caches.....

[+JDandDD]

Starbrand, you are absolutely correct about every cache would be expected to hover towards the middle. That's well understood in statistics. Take a large enough sample about half will be neutral to positive, and about half would be neutral to negative.

 

Add to that the ability to manipulate and its easy to see that any rating system would not be useful for long. I dislike micros in lampposts but others like them. Neither one of us is wrong. So how do I know that what you like I will and vice versa.

 

JDandDD

[+tozainamboku]

Starbrand, you are absolutely correct about every cache would be expected to hover towards the middle. That's well understood in statistics. Take a large enough sample about half will be neutral to positive, and about half would be neutral to negative.

 

Add to that the ability to manipulate and its easy to see that any rating system would not be useful for long. I dislike micros in lampposts but others like them. Neither one of us is wrong. So how do I know that what you like I will and vice versa.

 

JDandDD

 

Thats a poor understanding of statistics. If it were true there would be no winner on American Idol and President Bush's approval rating would be 50%. With a large enough sample, the mean of the sample will tend to the mean of the population. So in fact a cache that most people like would likely score higher on a survey and a cache that most people don't like would score lower. The problem is that very few people would vote on most caches - so the survey might not reflect what most people would feel. In fact, the people participating in the survey would not be a random sample. I would suspect that most caches (including lamppost hides) would get a positive score because people who enjoyed the cache would be the ones most likely to participate in the survey. By having ratings on several categories, as ReadyOrNot suggests, you at least get an idea of which of several factors people enjoyed more about a cache and depending on what kind of caches you enjoy it may be useful.

[+ReadyOrNot]

Starbrand, you are absolutely correct about every cache would be expected to hover towards the middle. That's well understood in statistics. Take a large enough sample about half will be neutral to positive, and about half would be neutral to negative.

 

Add to that the ability to manipulate and its easy to see that any rating system would not be useful for long. I dislike micros in lampposts but others like them. Neither one of us is wrong. So how do I know that what you like I will and vice versa.

 

JDandDD

 

Thats a poor understanding of statistics. If it were true there would be no winner on American Idol and President Bush's approval rating would be 50%. With a large enough sample, the mean of the sample will tend to the mean of the population. So in fact a cache that most people like would likely score higher on a survey and a cache that most people don't like would score lower. The problem is that very few people would vote on most caches - so the survey might not reflect what most people would feel. In fact, the people participating in the survey would not be a random sample. I would suspect that most caches (including lamppost hides) would get a positive score because people who enjoyed the cache would be the ones most likely to participate in the survey. By having ratings on several categories, as ReadyOrNot suggests, you at least get an idea of which of several factors people enjoyed more about a cache and depending on what kind of caches you enjoy it may be useful.

 

That sums up probably the biggest obstacle -- Having a large enough sample. What about hiding the results except in search results. In other words, I could search for "Micros with a view" and caches would be returned matching that criteria. So there would be no place to "VIEW" the rating on a cache.. The ratings would only be used behind the scenes on either Pocket Queries or search results. Similar to when you type in "Geocaching" in google. You don't know the algorithm and math behind why the results came up the way they did.. What do you think?

Edited June 16, 2006 by ReadyOrNot

[+Wacka]

Easy solution:

Read the cache page and logs.

If the logs are generic and short- probably means it is a lame cache.

[markandlynn]

The easy answer is come over to the UK to do your caching as we have This List to help us find the best and worst rated caches in the country.

[+Yellow ants]

The easy answer is come over to the UK to do your caching as we have This List to help us find the best and worst rated caches in the country.

 

We also have cache reviews on the Danish geocaching site. Each cache can only be reviewed when you've logged a find on gc.com, and you can then see the ratings for Danish caches when three people or more have rated.

 

Not surprisingly, the cache ratings do not tend towards the middle.

 

(Yes, I'm aware the top cache only has 3 ratings. But since the number of ratings is shown, you can also pick out those with more ratings as being more likely to be rated properly.)

Edited June 16, 2006 by Yellow ants

[+StarBrand]

The easy answer is come over to the UK to do your caching as we have This List to help us find the best and worst rated caches in the country.

 

We also have cache reviews on the Danish geocaching site. Each cache can only be reviewed when you've logged a find on gc.com, and you can then see the ratings for Danish caches when three people or more have rated.

 

Not surprisingly, the cache ratings do not tend towards the middle.

 

(Yes, I'm aware the top cache only has 3 ratings. But since the number of ratings is shown, you can also pick out those with more ratings as being more likely to be rated properly.)

I never said the middle would happen on ALL caches. But I am sure it would on the majority. A few caches will always hold the distinction of being good and a few others will be universally seen as bad but the majority of all caches would tend towards the middle. My point being, such ratings will not have the desired effect of aiding me find caches that I enjoy. Reading the description and logs will give me a much better idea of what I will find at the site.

 

Sure you have a top 100 list but just how many total caches exist on the list? 2000? 3000? Graph out the scores and I'll bet it is heavy towards the middle.

 

Add to that, the fact that the most likely people to use a voluntary rating are those with strong feelings one way or another and you see what I mean.

 

Thats a poor understanding of statistics. If it were true there would be no winner on American Idol and President Bush's approval rating would be 50%.....

 

Did you notice the night on American Idol when just 3 were left and the difference in percentages was less than 1%?? Split votes. Strong feelings. Just a slight nudge towards the 2 that moved on.

[+ReadyOrNot]

Did you notice the night on American Idol when just 3 were left and the difference in percentages was less than 1%?? Split votes. Strong feelings. Just a slight nudge towards the 2 that moved on.

 

Let's use the example of polling. Rarely do you ever see a poll that shows 50/50, unless it truly is 50/50. Watching the news this week, I saw a poll regarding immigration that was "80/20" with a sampling of 5000 citizens. According to your logic, it would be 50/50. It also depends on how you ask the question. A simple question like "Did you like this cache?" would clearly end up being 50/50 with a large sample. A question like "Did this cache have a scenic view?". If the cache had a spectacular view of Mt. Hood from atop a 1500 ft. hill, not many are going to vote "0". In fact, I can imagine not many would vote 5,6 or 7 either. The vast majority of votes would be extremely high. If the cache was located next to a garbage dump, I would expect the results to be extremely low. If the view was "okay" I would expect them to be in the middle. This is exactly what one would expect.

 

As far as American Idol.... You have 2 people with an option of "YES" or "NO" and a voting population of several million. What do you think is going to happen? If they asked the public to rate specific qualities of each singer on a scale system, you wouldn't see everyone come in at 5 on the scale.

Edited June 16, 2006 by ReadyOrNot

[+ReadyOrNot]

Easy solution:

Read the cache page and logs.

If the logs are generic and short- probably means it is a lame cache.

 

I'm guessing you also live in a small cabin in Montana? That's such an archaic way to determine the cache's value. Tell the caches along a route crowd, "Just get out a map and manually figure it out. Yes, you could do that, but there is probably a better way don't you think?

 

That's not an easy solution, that's a lazy solution.

[+StarBrand]

Let me be clear before backing out of this thread. I actual wish you could have some kind of mult-question ratings on a cache. In a perfect world with honest participation and similar thinking - it would be great. Very helpful to seek quality caches. However, given the vast diversity of cachers and locations and styles, I have been overtaken by a dose of reality.

 

I believe this type of system could be easily abused. (for better or worse)

 

I believe the diversity of opinion would drive many (not all) ratings toward center.

 

I believe not all cachers would contribute. You would be left with a few strong opinions and not a majority.

 

That is why I must vote no.

 

What would work is something akin to Markwell's solution. Allow each cacher to create a top 10% list. They could list as many as 10% of the caches they have logged finds on. Then when a cache appears on (for example) 5 or more lists - it would garner some text that says "xx number of cachers have listed this cache as a favorite 10%". That way - there would be no penalties and much less chance of manipulation. No rating number to rank. We are almost there with the inclusion of bookmark lists on the listing pages. Not quite there but close.

 

Thanks for your thoughts on the subject.....

[+tozainamboku]

 

I never said the middle would happen on ALL caches. But I am sure it would on the majority. A few caches will always hold the distinction of being good and a few others will be universally seen as bad but the majority of all caches would tend towards the middle. My point being, such ratings will not have the desired effect of aiding me find caches that I enjoy. Reading the description and logs will give me a much better idea of what I will find at the site.

 

Sure you have a top 100 list but just how many total caches exist on the list? 2000? 3000? Graph out the scores and I'll bet it is heavy towards the middle.

 

Add to that, the fact that the most likely people to use a voluntary rating are those with strong feelings one way or another and you see what I mean.

 

Thats a poor understanding of statistics. If it were true there would be no winner on American Idol and President Bush's approval rating would be 50%.....

 

Did you notice the night on American Idol when just 3 were left and the difference in percentages was less than 1%?? Split votes. Strong feelings. Just a slight nudge towards the 2 that moved on.

 

I don't watch American Idol, but I suspected that when I wrote that that someone would point out that difference between #1 and #2 would be very small.

 

If a rating system did accurately reflect the general public's rating of caches then about half the caches would be above average and the other half would be below average. The problem is that that a rating system would not be a random sample. As I said before, I think people would more likely rate the cache they liked - so my suspicion is that all rated caches would be above average like the children in Lake Wobegone .

 

A rating system that simply asks "would you recommend this cache" would probably not be useful. Micro lovers would rate micros high, puzzle lovers would rate puzzles high, etc. ReadyOrNot's proposal is to rate several attributes: Scenic View, Scenic Hike, Difficulty of Hide, Muggles, Family Friendly, and so forth. The question is not "does cache A have a better view than cache B?" - but instead to asks you to indicate if "scenic view" applies to the cache. Cache that are particularly scenic will score higher than cache that have no scenic value. Even if you are correct that most people will respond with something in the middle, a more scenic cache will end up with a higher number than a less scenic one. You could make use of this in a Pocket Query to find scenic caches. Becasue of small sample sizes you may run the risk of missing a scenic cache that got a low score or visit an unscenic cache that has a higher score than expected, but I think most caches will get ratings that are useful in sorting out which cache you want to look for and which ones you want to avoid.

[+ReadyOrNot]

Let me be clear before backing out of this thread. I actual wish you could have some kind of mult-question ratings on a cache. In a perfect world with honest participation and similar thinking - it would be great. Very helpful to seek quality caches. However, given the vast diversity of cachers and locations and styles, I have been overtaken by a dose of reality.

 

I believe this type of system could be easily abused. (for better or worse)

 

I believe the diversity of opinion would drive many (not all) ratings toward center.

 

I believe not all cachers would contribute. You would be left with a few strong opinions and not a majority.

 

That is why I must vote no.

 

What would work is something akin to Markwell's solution. Allow each cacher to create a top 10% list. They could list as many as 10% of the caches they have logged finds on. Then when a cache appears on (for example) 5 or more lists - it would garner some text that says "xx number of cachers have listed this cache as a favorite 10%". That way - there would be no penalties and much less chance of manipulation. No rating number to rank. We are almost there with the inclusion of bookmark lists on the listing pages. Not quite there but close.

 

Thanks for your thoughts on the subject.....

 

Your reasoning is not mathematically sound. It's all in how you ask the questions. I agree, if you ask the wrong questions, you'll get the wrong answers.. But to assume everything will move to center is just not accurate. Everything will create a bell curve of some sort with the majority at the center of the curve. To say that the center of the bell curve will always be at the middle of the set is just... Well, I'm beating a dead horse now.. It's just not accurate.

[+StarBrand]

 

Your reasoning is not mathematically sound. It's all in how you ask the questions. I agree, if you ask the wrong questions, you'll get the wrong answers.. But to assume everything will move to center is just not accurate. Everything will create a bell curve of some sort with the majority at the center of the curve. To say that the center of the bell curve will always be at the middle of the set is just... Well, I'm beating a dead horse now.. It's just not accurate.

 

ARrrrrrrrrGGGGGGGggggg!!!!!!!!!

 

I never said "everything" would head to center. Just the majority - and that would not necessarily mean that I would not find it to be an exceptional cache. And thus of no help.

 

I also never posted that there was ANY mathamatical soundness to my theory. It is an opinion you are asking for - not many mathamatical principles apply.

Edited June 16, 2006 by StarBrand

[+ReadyOrNot]

 

Your reasoning is not mathematically sound. It's all in how you ask the questions. I agree, if you ask the wrong questions, you'll get the wrong answers.. But to assume everything will move to center is just not accurate. Everything will create a bell curve of some sort with the majority at the center of the curve. To say that the center of the bell curve will always be at the middle of the set is just... Well, I'm beating a dead horse now.. It's just not accurate.

 

ARrrrrrrrrGGGGGGGggggg!!!!!!!!!

 

I never said "everything" would head to center. Just the majority - and that would not necessarily mean that I would not find it to be an exceptional cache. And thus of no help.

 

Arrrrggg back at ya! You are saying the MAJORITY would lean towards center and that's not true. If the questions were stupid, I would agree with you, but let's assume the questions aren't going to be stupid. In that case, they will lean towards whatever they lean towards.

 

Can you please fill me in on what mathematical concept you are using as a basis for this argument? Your argument being that given time, things average out. All polls would be worthless if this were a true mathematical concept.

[+Prime Suspect]

It's been my observation that when the topic starter has posted nearly half the posts in a thread, he or she is fighting a losing battle. You're around 40% right now.

[+ReadyOrNot]

It's been my observation that when the topic starter has posted nearly half the posts in a thread, he or she is fighting a losing battle. You're around 40% right now.

 

I'm having a conversation with Starbrand. If you don't have anything positive to add or anything useful, why don't you go somewhere else? I'm trying to actively be a part of this topic because I find it interesting.

 

Honestly.. Just the fact that you are in here counting posts shows your motivation. Beat it.

Edited June 16, 2006 by ReadyOrNot

[+Sparrowhawk]

I don't geocache as often as I could simply because my motivation is excellent location more than anything... and it's too difficult to tell which caches are good for location uniqueness and quality.

 

I am discouraged from geocaching a lot because I cannot sort by quality of location from looking the cache pages. I still cache some... not a whole lot.

 

If I am lucky enough to see by the logs that folks really love the interesting location, that makes it better. But I don't have time to look at all the logs on lots of cache pages just to find the one "wow, didn't know THIS was here!" kind of cache that makes me glad to go geocaching.

[+ReadyOrNot]

I don't geocache as often as I could simply because my motivation is excellent location more than anything... and it's too difficult to tell which caches are good for location uniqueness and quality.

 

I am discouraged from geocaching a lot because I cannot sort by quality of location from looking the cache pages. I still cache some... not a whole lot.

 

If I am lucky enough to see by the logs that folks really love the interesting location, that makes it better. But I don't have time to look at all the logs on lots of cache pages just to find the one "wow, didn't know THIS was here!" kind of cache that makes me glad to go geocaching.

 

Where in Oregon are you? Maybe we could share some lists of interesting caches. I'm about 15 miles out of Portland in Clackamas. PM me if you are interested. Thanks!

 

EDIT: You just placed a new cache about 6 miles from my house, so you must be close I'll send you an email...

 

--> Usual Suspects: Am I at 50% yet?

Edited June 17, 2006 by ReadyOrNot

[+JDandDD]

 

Thats a poor understanding of statistics. If it were true there would be no winner on American Idol and President Bush's approval rating would be 50%. With a large enough sample, the mean of the sample will tend to the mean of the population. So in fact a cache that most people like would likely score higher on a survey and a cache that most people don't like would score lower. The problem is that very few people would vote on most caches - so the survey might not reflect what most people would feel. In fact, the people participating in the survey would not be a random sample. I would suspect that most caches (including lamppost hides) would get a positive score because people who enjoyed the cache would be the ones most likely to participate in the survey. By having ratings on several categories, as ReadyOrNot suggests, you at least get an idea of which of several factors people enjoyed more about a cache and depending on what kind of caches you enjoy it may be useful.

Actually, having studied statistics through grad school I understand stats. George Bush's approval is yes/no, binary votes. Its a dichotomy and so there is no regression to the mean just majority. American idol is also not a rating scale but really a series of yes/no binary votes. Yes or no on each contestant. If you vote Y for one you automatically vote not on the others. So there are no ranges of scaling so again no regression to the mean.

 

What we are talking about here is an ordinal rating scale that has a defined range of answers. It has a range so will have regression to the mean. With enough answers you will tend to the middle. A very well known and proven statistical concept.

 

Yes, its all more complicated. There are four types of scaling: nominal. ordinal. interval and ratio. Ok, lets stop there and keep it simple like above or we'll lose everybody.

 

JDandDD

[+Prime Suspect]

It's been my observation that when the topic starter has posted nearly half the posts in a thread, he or she is fighting a losing battle. You're around 40% right now.

 

I'm having a conversation with Starbrand. If you don't have anything positive to add or anything useful, why don't you go somewhere else? I'm trying to actively be a part of this topic because I find it interesting.

 

Honestly.. Just the fact that you are in here counting posts shows your motivation. Beat it.

 

43%

[+ReadyOrNot]

It's been my observation that when the topic starter has posted nearly half the posts in a thread, he or she is fighting a losing battle. You're around 40% right now.

 

I'm having a conversation with Starbrand. If you don't have anything positive to add or anything useful, why don't you go somewhere else? I'm trying to actively be a part of this topic because I find it interesting.

 

Honestly.. Just the fact that you are in here counting posts shows your motivation. Beat it.

 

43%

 

Keep it up, you're just ensuring that the thread stays at the top of the list. Thanks! And adjusting the percentage removing my responses to you, I'm well below 40%.

Edited June 17, 2006 by ReadyOrNot

[+KBI]

Referring back to statistics again...

That's well understood in statistics...

Thats a poor understanding of statistics....

Your reasoning is not mathematically sound. It's all in how you ask the questions ...

Actually, having studied statistics through grad school I understand stats ... binary votes ... dichotomy ... regression ... mean ... majority ... ranges of scaling ... regression to the mean ... Yes, its all more complicated. There are four types of scaling: nominal. ordinal. interval and ratio......

 

Following a flaming snowmobile crash, one statistician asked the other if he was OK. The second said "well, my hair's on fire and my toes are frostbitten, but ... overall I feel pretty good."

 

 

Q: What do you get if you cross an elephant with a grape?

A: Elephant grape cosine theta

 

 

79.48% of all statistics are made up on the spot.

 

 

One day there was a fire in a wastebasket in the dean's office and in rushed a physicist, a chemist, and a statistician. The physicist immediately starts to work on how much energy would have to be removed from the fire to stop the combustion. The chemist works on which reagent would have to be added to the fire to prevent oxidation. While they are doing this, the statistician is setting fires to all the other wastebaskets in the office. "What are you doing?" they demanded. "Well to solve the problem, obviously you need a large sample size" the statistician replies.

 

 

It is proven that the celebration of birthdays is healthy. Statistics show that those people who celebrate the most birthdays become the oldest.

 

 

Statistics are like a bikini. What they reveal is suggestive, but what they conceal is vital.

 

 

Three statisticians go deer hunting with bows and arrows. They spot a big buck and take aim. One shoots and his arrow flies off exactly ten feet wide to the right. The second shoots and his arrow flies off exactly ten feet wide to the left. The third statistician starts jumping up and down shouting "We got him! We got him!"

 

 

Statistics means never having to say you're certain.

 

 

Two statisticians were traveling in an airplane from LA to Honolulu. About an hour into the flight, the pilot announced that they had lost an engine, but don't worry, there are three left. However, instead of 5 hours it would take 7 hours to get to Honolulu. A little later, he announced that a second engine failed, and they still had two left, but it would take 10 hours to get to Honolulu. Somewhat later, the pilot again came on the intercom and announced that a third engine had died. Never fear, he announced, because the plane could fly on a single engine. However, it would now take 18 hours to get to Honolulu. At this point, one statistician turned to the other and said, "Gee, I hope we don't lose that last engine, or we'll be up here forever!"

 

 

 

In the film "The World According to Garp," Garp (Robin Williams) is looking at a house he's thinking of buying when, as if from nowhere, a small plane veers out of control and crashes into the home.

 

"We'll take the house," Williams says to the dumbstruck real estate agent. Then he says to his wife, "Honey, the odds on another plane hitting this house are astronomical. It's been pre-disastered. We'll be safe here!"

[+tozainamboku]

 

Thats a poor understanding of statistics. If it were true there would be no winner on American Idol and President Bush's approval rating would be 50%. With a large enough sample, the mean of the sample will tend to the mean of the population. So in fact a cache that most people like would likely score higher on a survey and a cache that most people don't like would score lower. The problem is that very few people would vote on most caches - so the survey might not reflect what most people would feel. In fact, the people participating in the survey would not be a random sample. I would suspect that most caches (including lamppost hides) would get a positive score because people who enjoyed the cache would be the ones most likely to participate in the survey. By having ratings on several categories, as ReadyOrNot suggests, you at least get an idea of which of several factors people enjoyed more about a cache and depending on what kind of caches you enjoy it may be useful.

Actually, having studied statistics through grad school I understand stats. George Bush's approval is yes/no, binary votes. Its a dichotomy and so there is no regression to the mean just majority. American idol is also not a rating scale but really a series of yes/no binary votes. Yes or no on each contestant. If you vote Y for one you automatically vote not on the others. So there are no ranges of scaling so again no regression to the mean.

 

What we are talking about here is an ordinal rating scale that has a defined range of answers. It has a range so will have regression to the mean. With enough answers you will tend to the middle. A very well known and proven statistical concept.

 

Yes, its all more complicated. There are four types of scaling: nominal. ordinal. interval and ratio. Ok, lets stop there and keep it simple like above or we'll lose everybody.

 

JDandDD

 

You are correct that what is proposed is an ordinal scale. As such it makes no sense to to talk about the mean response. (If one person rates a cache as 0 and another rates it 10 the cache isn't necessarily a 5) I'm not even sure why you bring up regression to the mean here. You seem to think that an average cache would get a ranking of five, and that if someone were to rank a cache significantly higher or lower than 5 the next person would rank it closer to 5. This is a misunderstanding of regression to the mean. Even if this were true it only means that no cache would get a average (or in the case of an ordinal scale, median) ranking close to 0 or 10. A scenic cache will most likely still have a higher median ranking than the non scenic cache. The proposed system would be useful in separating out caches that are scenic from ones that are not.

[+KBI]

But seriously;

 

What we are talking about here is an ordinal rating scale that has a defined range of answers. It has a range so will have regression to the mean. With enough answers you will tend to the middle. A very well known and proven statistical concept.

Eh. Maybe you're right, but my lack of education leaves me more inclined to agree with ReadyOrNot on that point. Think about it: suppose you applied the same survey/rating system to a restaurant. Do you really think the item labeled "The pizza with anchovies tasted GOOD" would trend toward 50%?

[+Wacka]

Easy solution:

Read the cache page and logs.

If the logs are generic and short- probably means it is a lame cache.

 

I'm guessing you also live in a small cabin in Montana? That's such an archaic way to determine the cache's value. Tell the caches along a route crowd, "Just get out a map and manually figure it out. Yes, you could do that, but there is probably a better way don't you think?

 

That's not an easy solution, that's a lazy solution.

 

Nope, I live in the SF Bay area wit ~8K caches within 100 miles of me. I've found about 2500 of them. I generally look at the cache listing on my palm when I'm dne with thenprevious cache and going to the next one.

[+ReadyOrNot]

Easy solution:

Read the cache page and logs.

If the logs are generic and short- probably means it is a lame cache.

 

I'm guessing you also live in a small cabin in Montana? That's such an archaic way to determine the cache's value. Tell the caches along a route crowd, "Just get out a map and manually figure it out. Yes, you could do that, but there is probably a better way don't you think?

 

That's not an easy solution, that's a lazy solution.

 

Nope, I live in the SF Bay area wit ~8K caches within 100 miles of me. I've found about 2500 of them. I generally look at the cache listing on my palm when I'm dne with thenprevious cache and going to the next one.

 

So you don't care about the quality of the caches then? How would having a rating system hinder the way you find caches? Out of serious curiousity. It just seems like the method you mentioned of looking at the logs is far less clear and concise then a rating system would be. I don't think a "Perfect" system exists, but that's why it's fun to discuss.

 

(Unusual Suspect --> 39% and falling.. What's that mean?)

Edited June 17, 2006 by ReadyOrNot

[+Vinny & Sue Team]

There's been a lot of talk in the forums about the numbers game and those evil micro caches (personally, I love micro's that are well hidden and bring me somewhere interesting). The survey/rating system would create social change by rewarding those that take the time to hide well thought-out, interesting caches. But it doesn't hurt those that still like to hunt micros (like me!).

 

Whenever I use words, I always seem to screw up, so I'll let the pictures do the explanation for me. Let me know what you think????

 

The cacher is not forced to review the cache, but is given the option

 

If they choose to, they are put here after logging their find to complete the survey.

 

This is one of many possibilities for interpreting the results of the survey.

Fun idea, but your sense of what is desirable about a cache is largely very different from mine, and thus your chosen scale items do not at all cover what is important to me nor to many folks whom I know. At least your review concept allows only finders to review a cache: one of the extant problems with the off-site cache review services such as that offered by Keenpople.com is that the stats may be manipulated by casual visitors to the cache listng page, i.e., people who have never found the cache.

[+KBI]

Fun idea, but your sense of what is desirable about a cache is largely very different from mine, and thus your chosen scale items do not at all cover what is important to me nor to many folks whom I know.

Which criteria would you add or remove from ReadyOrNot's proposal?

[+Wacka]

The rating system a you showed it would not work on a palm for one thing. I have about 3K caches on my palm, covering the Bay Area.

I go for all of them. The bad quality ones in developed areas usually get muggled quickly.

There is a 12 year old in the area that has all of his 6 caches missing. He thinks someone is out to get him, I think it is just poor placement/camo. I will avoid his caches until they last for a while. I saw this using GSAK.

 

Besides, you're beating a dead horse. This comes up almost monthly and Jeremy has stated he will not implement a rating system.

Edited June 17, 2006 by Wacka

[+ReadyOrNot]

This comes up almost monthly and Jeremy has stated he will not implement a rating system.

 

This shows why it is important that everyone should read the forums for themselves and not take anything at face value. Here's what Jeremy actually said: (Similar enough that it is worthy of debate)

 

What does Jeremy think?

 

EDIT: (Usual Suspects --> 38%)

Edited June 17, 2006 by ReadyOrNot

[+ReadyOrNot]

Fun idea, but your sense of what is desirable about a cache is largely very different from mine, and thus your chosen scale items do not at all cover what is important to me nor to many folks whom I know. At least your review concept allows only finders to review a cache: one of the extant problems with the off-site cache review services such as that offered by Keenpople.com is that the stats may be manipulated by casual visitors to the cache listng page, i.e., people who have never found the cache.

 

I didn't put a whole lot of thought into making the list in the graphics. Also, it's not necessarily what I find desirable. It's simply criteria that can be searched on. I'm sure that not every cache I go after is going to have a view, but on a nice sunny, clear day, maybe I want a view? Maybe I'm 2 caches from 500 and want a couple quick park and grab's to make it over the hump. Maybe I'm feeling anti-social and don't want to run into any muggles. Maybe I've got my family with me and want something they would enjoy (I've seen many posts in the forums "Does anyone know of any family friendly caches in Detroit?") What criteria would you use?

 

EDIT: Didn't see KBI's response.. Sorry! (Now my Usual Suspects % is up, dang!)

Edited June 17, 2006 by ReadyOrNot

[+JDandDD]

You are correct that what is proposed is an ordinal scale. As such it makes no sense to to talk about the mean response. (If one person rates a cache as 0 and another rates it 10 the cache isn't necessarily a 5) I'm not even sure why you bring up regression to the mean here. You seem to think that an average cache would get a ranking of five, and that if someone were to rank a cache significantly higher or lower than 5 the next person would rank it closer to 5. This is a misunderstanding of regression to the mean. Even if this were true it only means that no cache would get a average (or in the case of an ordinal scale, median) ranking close to 0 or 10. A scenic cache will most likely still have a higher median ranking than the non scenic cache. The proposed system would be useful in separating out caches that are scenic from ones that are not.

Ok, I should have done that as one statistician to another it looks. Sounds like we both know the statistics issues and I do know about the mean and ordinal.

 

However, as others have said, lets drop this and keep more to the topic. My point to you was simply that you shouldn't assume the other doesn't know about stats and neither should I have. Enough said!

 

I will say that I generally don't like rating systems for caches or waymarks. They are using a rating system on waymarks and my experience there is that it doesn't matter what others say. Some people really liked waymarks that I didn't and vice versa. I'm not sure that it accomplishes anything positive and am concernede about the possible negatives on what some people would see as good caches. I think there is the same possibility on Waymarking and it shouldn't be assumed that its necessarily a good thing.

 

Interesting ideas in this thread, but, there is no assurance that people or a group can't attack a cache with negative ratings if they have something against the owner or the cache type. You can create sock puppets faster than they can be found and removed so even with one vote per find you couldn't control fake voting. There are enough instances of cache thieves stealing caches or destroying them, fake finds (another source of inaccurate reviews), and other types of behaviour that I'm not convinced you could ever be sure that the ratings you are reading are really an accurate reflection of the cache (and similarly the waymark).

 

JDandDD

[+tozainamboku]

I agree that just ranking the overall quality of cache, like ranking the overall quality of a waymark, is not very useful. What one person likes, another may dislike. This is my objection to the suggestion of using favorites list to find recommended caches. It might help if there were a "cachers who liked this cache also liked" type of mechanism but I'm not sure how it would be implemented. There are a few cachers that never met a cache they didn't like and that could skew the results of that type of system. ReadyOrNot's system asks you to rank how well several statements apply to a cache. We could debate which statements should be included in his survey, but whatever they are the survey will allow people to see how other cachers perceive a cache. Will there be abuses? Probably, but I think most people who participate will be honest in their answers. Once you use the rankings to narrow down the caches you want to look for, it would still be the cachers responsibility to look at the cache page and logs to make a final decision if they want to do the cache.

 

Like Wacka, I use other ways to decide what cache I look for in my own area. If I feel like hiking I'll look for higher terrain caches; if I just want to grab a few, I will usually just load a bunch into my GPS and just drive to the next nearest one; I like puzzles, so I get instant notification of new puzzles and add the ones I want to work on to a bookmark list. Once I have solved the puzzle, the corrected coordinates are entered in GSAK and I treat them like traditional caches. But when I go out of town and will only have a few hours to cache in an unfamiliar setting it would be nice to have a few "best" caches. I also realize that most people don't cache as often a I do in their local area. I suspect these people would also like to have some kind of rating system to help select caches.