cache rating - request for comments [closed]

Discussion about the Geocaching Australia web site
rhinogeo
2700 or more caches found
2700 or more caches found
Posts: 1213
Joined: 31 October 03 11:45 am
Twitter: rhinogeo
Location: Benalla, VIC

Post by rhinogeo » 14 September 04 4:46 pm

Rabbito seems to have hit the nail on the head - If a subjective rating system is to be implemented, the simpler the better.

However I do agree with some of the previous posts regarding getting the Terrain and Difficulty ratings consistent rather than implementing something new. If more cachers used Clayjar's Rating System a lot of angst could be avoided.

If I'm visiting a new area, I do a search for the nearest caches, read the logs, and take it from there. It's worked for me so far :)

User avatar
Aushiker
350 ? I am the lizard queen
350 ? I am the lizard queen
Posts: 1397
Joined: 30 July 04 2:35 pm
Twitter: Aushiker
Location: Fremantle, WA
Contact:

Post by Aushiker » 14 September 04 8:00 pm

rhinogeo wrote: If more cachers used Clayjar's Rating System a lot of angst could be avoided.
<p>
Is this the same rating system linked from Geocache.com when one logs a new cache or something different?
<p>
Andrew

User avatar
caughtatwork
Posts: 16300
Joined: 17 May 04 12:11 pm
Location: Melbourne
Contact:

Post by caughtatwork » 14 September 04 10:14 pm

Yep.

User avatar
TEAM LANDCRUISER
Posts: 476
Joined: 04 February 04 9:28 pm
Location: Port Kennedy WA
Contact:

Post by TEAM LANDCRUISER » 14 September 04 10:17 pm

<font color="blue"><font face="verdana"><b>Yes Andrew it's the same rating system that's linked from GC.com

User avatar
ideology
Posts: 2763
Joined: 28 March 03 4:01 pm
Location: Sydney
Contact:

Post by ideology » 15 September 04 2:05 am

whew, thanks for all the discussion - there were a lot more sides to this than we originally thought

we probably led the discussion off in the wrong direction right at the start by calling it a "rating" system. as facitman and caughtatwork pointed out, we need to focus on the purpose, which for us is to highlight the "don't miss" caches. we should probably have called it a "don't miss" system or a "wow" system rather than a rating system because once you start thinking of rating things you get tangled up in saying that one thing is "better" than another. which is not necessarily the case - they can be just different. apples and oranges, meat and chewy and all that. dingbats and mind socket highlighted a second purpose, which was to provide feedback for the hider, which is a good idea, but as we saw, could lead to negative feedback.

rabbittos' proposal seems to achieve the primary purpose quite well. if we saw a nearby cache which was on the top 10 list, we'd be grabbing our gps and heading off! yippee! it partially achieves the secondary purpose of providing feedback to the hider, but not as much because it is designed not to provide feedback unless the cache is very good.

based on the loon's comments, we'd like to get away from asserting that the cache with the most number of top-ten listings is some kind of winner. so rather than put up a top-ten list, we'd prefer to simply highlight the cache in cache listings to say that it's a favourite and leave it at that. once you know it's a favourite, you could click through to the cache page to find out who listed it as a favourite.

as caught at workmentioned, you need to know why someone liked a particular cache before you can determine whether you will like it. that's true. however, we would not blindly do the cache with the highest number of favourites. we'd use the favourites to narrow down the choices. the identity of the nominator can convey a lot of information. for example, we know leek, so we know that if he's got a favourite cache, we will probably enjoy it too. we don't know facitman, but if he thought a cache was good, it would probably be enough for us to at least scan the logs. if we saw that we and facitman shared some of the same favourites, we'd take his opinion even more highly. yes, there is a possibility that noone shares our preference and thus we'll miss a particularly quirky cache, however hopefully this is the exception not the rule. we would still scan the other caches (particularly those by hiders whose previous caches we have enjoyed) because virgin caches wouldn't be on anyone's favourites.

all cached up and snuva, thanks for offering to subscribe to the site. yes, it does take time, effort and money to do this, but we enjoy doing it! this way we also get to make the occasional boo-boo and noone gets upset! if you'd like to help at the moment, the main help we need is testing the sytem and getting good bug reports. we also ask for additional help from people. for example swampy and leek have done some alpha testing, the moderators did some as well. mind socket is trying to work out a cache recommendation algorithm, pyrenees kid and the moderators are keeping the forum ticking over. hey, here's a job: if there is anyone with good html/css skills who would like to volunteer to fix that pesky header overlap issue, we'd be more than happy to throw the problem their way!

alex, loon and swampy: what do you think of rabbitto's proposal? comments from others?

User avatar
riblit
It's the journey.
It's the journey.
Posts: 3444
Joined: 04 April 03 6:30 pm
Location: Land Grant of John Campbell

Post by riblit » 15 September 04 11:33 am

ideology wrote: hey, here's a job: if there is anyone with good html/css skills who would like to volunteer to fix that pesky header overlap issue, we'd be more than happy to throw the problem their way!
I just had a quick look at this. I can reproduce it at will, the pages render differently on IE6 than they do on Firefox.
on IE the colums will stretch as the browser is resized wider so the margins remain a constant size. the gap between the 'tools' tab and the end of the
logo varies.
With Firefox the distance between the end of the logo and the 'tools' tab is constant. As the browser is widened the margins change size, giving the effect of a fixed page size with a varying left and right margin.

I have some notes on this somewhere, I'll see if I can find them.
You have to detect the browser and send different css.

The tabs disappearing under the logo happens on both browsers as the width is reduced.

On IE the tab bar and search box appear to be right aligned and move with any change in width, eventually sliding under the logo as the width is reduced.

On Firefox the page margins decrease as the width is reduced, then the tab and search box slide under the logo the same as with IE.

I'll grab the css files and see if I can come up with anything..

swampgecko
It's all in how you get there....
It's all in how you get there....
Posts: 2181
Joined: 28 March 03 6:00 pm

Post by swampgecko » 15 September 04 8:30 pm

The feral animal ( sorry rabbitto :wink: I grew up on a farm and have seen first hand the damage caused by domestic and feral animals to our bush) has come up with some good ideas there, but my concern is that they are being called the cacher's top ten.... I prefer the term "recommended caches"...... with no set number to be in their list.

It would be nice for the list not be numbered as such, as even putting a number against them is still giving it a rating as such.

"voting" should be left for the annual cache awards. or that up coming circus......

my two cents.

User avatar
The Ginger Loon
450 or more roots tripped over
450 or more roots tripped over
Posts: 824
Joined: 28 March 03 9:09 pm
Location: Tamworth
Contact:

Post by The Ginger Loon » 16 September 04 1:43 am

Rabbittos suggestions are quite good as would remove my reservations.

On reflection however I'm wondering on the usefullness of allowing 10 nominations per cacher per state. Assuming that the majority of cachers predominately cache in their own state (yes, I know there are exceptions), I feel nominating interstate could be a wasted resource. Better to let the locals reflect on the local scene and let visitors enjoy the benefit of their knowledge.

Calling the list "Recommended Caches" (as Swampy sugests) would also help remove any misunderstanding about what such a list is. You could also achieve this by listing the caches in Alphabetical order rather than by Highest Nomination.

Someone else suggested having monthly nominations. I don't know about anyone else but I would get bored with that VERY quickly. We already have too many elections in Australia as it is. Nominate ten caches and remove one to add one works for me.

The Most Popular caches for the month can be handled by the state pages which rate caches by number of finds. Maybe this data can be displayed on a separate page month-by-month as an historical record of activity (slightly off-topic to this tread I know, sorry)

Swampy said "voting should be left for the annual cache awards".
Maybe that process could be co-ordinated nationwide (but still run on a State basis) through these forums or the GCA Website (also off-topic)

Another 2 cents into the ring...

User avatar
The Ginger Loon
450 or more roots tripped over
450 or more roots tripped over
Posts: 824
Joined: 28 March 03 9:09 pm
Location: Tamworth
Contact:

Post by The Ginger Loon » 17 September 04 1:27 am

I must do all my best thinking at night, which doesn't give me much hope for my day job... :oops: :roll:

Another thing that I've thought of is the question of caching regions. Here in NSW we have three distinct major caching regions; Greater Sydney, Newcastle/Hunter & Dubbo/ed Ochre. If we have a State based Recommended Cache system, it doesn't really give an indication of the 10 must-do caches for each area.

To put it another way if I'm visiting Dubbo or Newcastle from Sydney how will I know what the must-do caches are if they are all in Sydney or only 1 or 2 regional caches make the cut?

I'm guess the other states have caching hotspots besides their respective Capital Cities.

So should we have a must-do/don't miss/recommended cache/whatever list for regional caching centres as well as capital cities?

2 cents cache cash more in the ring...

User avatar
Papa Bear_Left
800 or more hollow logs searched
800 or more hollow logs searched
Posts: 2573
Joined: 03 April 03 12:28 am
Location: Kalamunda, WA
Contact:

Post by Papa Bear_Left » 17 September 04 8:43 am

swampgecko wrote:The feral animal ( sorry rabbitto :wink: I grew up on a farm and have seen first hand the damage caused by domestic and feral animals to our bush)
I hate the vermin as much as you do (it became personal when they ate the vines I'd spent days breaking my back planting), but a Rabbitto (more usually Rabbitoh) is someone who traps and sells the buggers. As long as he's not farming them, I'm all for the noble occupation!

Over here it's possums that cause the damage. It's strange to think of what I've always known as protected natives as vermin!

User avatar
ideology
Posts: 2763
Joined: 28 March 03 4:01 pm
Location: Sydney
Contact:

Post by ideology » 17 September 04 9:23 am

yes, areas and hotspots make things a tricky. we think it could work without any area limitations because the recommendations could be viewed within the context of existing searches, eg nearby caches or "caches found by these finders"

we have been experimenting with how to implement recommendations. we've set up an example on
our cacher page. there, we recommend three different caches. (interestingly, they are quite different caches - there's obviously no formula for "wow!") when you click on the cache, you go to the normal cache page, where you can see the recommendation marked against our logs.

at the moment these are the only places you can see the recommendation, but if people like this type of thing, then we can put the recommendation on the results of all searches. this means that if you search for all caches in dubbo, you'll see not only the caches, but something next to the cache which say they have been recommended.

there is currently no user interface to record recommendations. the reason is that we aren't sure how to store them in the database. at the moment they are a flag within an existing found log. we might implement it this way, but we will probably implement it as a separate log type because of synchronisation issues with the secondary site. this would be quite nice, because then once we write the recommendation user interface, we've written the logging interface as well!

your thoughts?

on another note, we've added some little icons to the cache and cacher pages. we have been trying to keep the pages as clean as possible, but we also wanted to show more than just the finds, so we neeed a way to denote them. do you think that the pages are getting too messy? if so, any ideas on a cleaner layout? for example, dropping the suburb makes it cleaner, but it's really handy to see where people have been going. for example, ecoteam must have got out their passports and headed for the northen beaches recently! we've tried dropping the brackets, doing italics, smaller font, etc. we would appreciate feedback on what works best for you.

User avatar
caughtatwork
Posts: 16300
Joined: 17 May 04 12:11 pm
Location: Melbourne
Contact:

Post by caughtatwork » 17 September 04 10:36 am

This is interesting ideaology.

On the cache Panorama Cache you have provided a Recommendation (I appreciate that you may have randomly assigned the recommendation, but for today, let's assume you actually meant to put it there).

When I go to read you log, I find nothing in your log that allows me to identify why you recommended it.

Reading the other logs, I see it's an 18k walk and that a couple have included logs that suggest it's worthwhile, but I really can't tell whether it's worth an 18k walk.
cache I recommend for all to do.....
Good one, Gren
A great cache
Great walk with excellent views
This is obviously where subjectivity comes in. You have a personal recollection of the cache that is not reflected in the cache logs, so you can recommend the cache, but I can't see what the recommendation is really for (I assume it's the views).

Is it possible to consider this in a slightly different manner.

Make a recommendation and then on the 'make your recommendation' page, add some radio buttons (or check boxes for multiple criteria) that allow selection of (say).
Views, Walk, Container, etc.
Then when you make a recommendation, you select a radio button of group of tick boxes and the final result would be that the cache would be:

Recommended (Views) or Recommended (Views, Container) or Recommended (Puzzle, Views, Hiking) or something.

This may be more effort but I wonder if it would help determine the reason for the recommendation, especially if the log does not highlight the reasoning.

Just a thought.

Mind Socket
Posts: 1329
Joined: 29 March 03 6:04 pm
Location: Gladesville, Sydney
Contact:

Post by Mind Socket » 17 September 04 11:16 am

If someone recommends a cache, there's a good chance it won't suck for anybody (taking the cache description into account). Isn't that enough?

Let's not take the elements of mystery and surprise out of it. The caches I enjoyed most were attempted without knowing much about them.

If you want to know why someone recommends a cache, go log it. :wink: I can only think of a handful of particularly sucky caches out of nearly 200 finds, and I'm sure no one else would put recommendations against them.

- R

User avatar
caughtatwork
Posts: 16300
Joined: 17 May 04 12:11 pm
Location: Melbourne
Contact:

Post by caughtatwork » 17 September 04 11:34 am

I agree that when taking the cache description and all of the logs (not necessarily just the log for the cacher who did the recommendation) into account you get a much better opinion about what the cache is all about.

This comes into play if you're on a trip and need to find recommended caches to play with in a short amount of time.

Just taking this another bit further (and I have no statistics to back me up on this one and maybe only time will tell).

Say there are 50 active cachers in a state. With 10 recommendations each that's a potential for 500 individual caches to be recommended. This situation is very unlikely.

So say each cacher happened to nominate 2 unique caches and 8 that someone else also recommends (80/20 rule).

That's still 150 caches that have a recommendation. In Victoria that's 1/3rd to 1/4tr of all caches in the state. Given the distribution is likely to be in the Metro areas where most people do their caching the percentage goes up even higher.

Taking into account that there is no 'leaderboard' for recommended caches, until you seek the cache page you won't know how many people made a recommendation for a cache. It's simple to presume the more recommendations the higher the liklihood of wanting to find it, but with no leaderboard how do you determine which of the 150 have the most recommendations? Much as I dislike leaderboards (for personal reasons) the recommended list may need to have one so people can determine the highest recommended caches and target those first.

I just don't know about the end result. We may end up with 1/3rd of all Metro caches being recommended (for very different reasons) and that doesn't necessarily help anyone who's in town for a quick hunt.

Again, not trying to be overly negative and maybe the solution may be 5 recommendations rather than 10 which would reduce the distribution to fewer caches (possibly).

Just thoughts, not criticisms.

Mind Socket
Posts: 1329
Joined: 29 March 03 6:04 pm
Location: Gladesville, Sydney
Contact:

Post by Mind Socket » 17 September 04 11:56 am

I think the design could be to show on individual logs who recommends a cache, but in the cache listings (nearest etc), a threshold of recommendations would have to be reached before it is highlighted. Perhaps a weighted ratio of finds to recommendations. For example, if more than 10% of finders recommend then it shows as a must do. That means only 1 recommendation is needed for the first 10 finds, 2 for up to 20 finds and so on. Picking a sweet spot might depend on the density of caches, but would more likely be just a function of how many recommendations each cacher may nominate. That is, if you can only recommend 3 caches, then a "must do" might show at a 5% recommendation rate. If you can nominate 100 caches, then "must do" caches might be at a 50% recommendation rate.

The latter would be a bad choice.

- R

Post Reply