Advanced

Re: GC9XA74Bigger better challenge

[Resolved] GC9XA74Bigger better challenge
July 17, 2022 05:14PM
I tried to design something similar before but it didn't fit within the time restricted rules like this challenge does.

Like the childhood trading game of 'bigger better', you are trying to find a group of at least geocaches that are bigger or better than each other. Date found is not a factor so this is not time limited in any way. We are using difficulty, terrain, type, and size as our categories. To add to your qualifying group a geocache has to increase in at least one of the four categories, but cannot decrease in any of them. You need a group of at least 20, but up to 26 is possible.

For difficulty and terrain, it is easy to grasp, 1 is the lowest and 5 is the highest.
For size we need to exclude the virtual size caches (all events, locationless, lab cache, earth, virtual, and webcam)

not chosen
micro
small
other
regular
large

And for type, in order of getting 'better' and excluding ones without 'size' we have:

traditional cache
multi cache
mystery cache
letterbox cache
wherigo cache

Here is an example of a qualifying group. There are a few places another geocache could be added, such as a Difficulty 1.5 terrain 1.5 size not chosen traditional. I highlighted in green each time one of the categories increases. You'll notice none of the categories get smaller as you go down the chart. The chart is only organized smallest to biggest for visual clarity, these geocaches could have been found in any order.
Re: GC9XA74Bigger better challenge
July 17, 2022 06:57PM
I expect this request requires a special script
Re: GC9XA74Bigger better challenge
July 18, 2022 08:38AM
When I tried to do a time sensitive version it had a fifth column of data that needed to join the 'don't decrease and at least one category has to increase'. That challenge was against the rules due to time limiting, but I knew the coding would use a fourier series after talking with the mathematics community. This is less columns so maybe it doesn't need a fourier series.
Re: GC9XA74Bigger better challenge
September 26, 2022 05:01AM
I have written a challenge checker for this challenge:
https://project-gc.com/Challenges/GC9XA74/71643

Unfortunately, according to the script I wrote, you don't pass the challenge (you have 18, not 20 - just like me)? Thus, according to the rules, you would not be able to publish it. As long as I understood it correctly, this is a very hard challenge, and very few users I've tried pass. If you'd like to check it out against your finds, and see if there's some error, I can fix it, or if you'd like to lower the threshold, that's also no problem.
Re: GC9XA74Bigger better challenge
September 26, 2022 04:12PM
Hello Bmuzzin.

Thank you for helping with this script. I tried to open the link to check it out against my finds and for some reason I couldn't. The page opened up and said the checker was disabled with this message beneath:

"If you found this link on a cache page you could inform the cache owner of the issue if it persists for more than 24 hours."

Is it something I'm doing wrong on my end? Either way, I'm super excited about this.
Re: GC9XA74Bigger better challenge
September 28, 2022 07:49PM
Nope - that was totally my fault. I forgot to enable the checker, which apparently if you don't, non-developers can't see it. Sorry, I'm new at this!
Re: GC9XA74Bigger better challenge
September 28, 2022 09:49PM

Here are my qualifying finds that give me a list 22 long. This sounds like a challenging checker to build. Thank you for your help with this.
Re: GC9XA74Bigger better challenge
September 29, 2022 04:14AM
I hate to be the bearer of bad news but this checker request isn't resolved just yet. The checker is displaying that I have only 18 out of 20 but I found a list of 22 qualifying finds. You can see the image in the last post. I don't know what is causing the false result but I'm guessing it takes the closest path instead of the longest path.

As a non developer, the inner workings of this exciting checker are a mystery to me, but I'm guessing even in LUA coding the solution would be a directed acyclic graph seeking the longest path in a topological sort. The path subsequence can be non contiguous since we don't need to maintain temporal order (date found).
Re: GC9XA74Bigger better challenge
October 03, 2022 04:45AM
Hi there - just an update - I'm still working through this one. You are correct, I was making a bad assumption in the original search code. I see the mistake now, and have corrected it. Unfortunately, finding the longest path on the 4 sort criteria with an exhaustive search takes considerably longer than 60 seconds, so I'm going to have to optimize it. I'll probably have to use A* search (or similar) to get a reasonable completion time. That will take a little time to implement, so please be patient.
Re: GC9XA74Bigger better challenge
October 03, 2022 05:15AM
You are doing something that is so far beyond me. You have my awe, appreciation, and as much patience as you need.
Re: GC9XA74Bigger better challenge
October 04, 2022 04:02PM
Okay, I think I'm finally happy with it! I optimized the search, so now it completes very quickly (about 1 second). I updated the output columns, so it looks very similar to the qualifier table you posted. It gives 22 for your longest path. It's not the same path as you posted above, but it's just looking for any path that would be the longest, so I think it's still valid. I might use this for a challenge cache of my own, since it's a unique kind of challenge.

Please test it out, and respond:
https://project-gc.com/Challenges/GC9XA74/71643

Cheers,
bmuzzin
Re: GC9XA74Bigger better challenge
October 05, 2022 12:44AM
bmuzzin, you are an absolute BEAST! this checker is fantastic. I don't know what black magic wizardry you did behind the scenes but that works so fast and it worked for both positive and negative tests. AMAZING!
The output is also just incredible with the green highlight on each increase.

Please do use this challenge checker for your own challenge, especially with how much work went into it.

Cheers,
CheekyBrit
Re: GC9XA74Bigger better challenge
October 05, 2022 03:52AM
@CheekyBrit can I steal this idea and attempt to get 1 published in Australia?

:)

Cheers
Adriaan
Re: GC9XA74Bigger better challenge
October 05, 2022 04:43AM
YES! Let's honor the fine work bmuzzin put into this and spread the idea.
If be nice if you referenced bmuzzin and I, but it's not required. Same for anyone else wanting to do a similar challenge. Reference this one so your checker builder can just adjust the gc code it is linked to.
Re: GC9XA74Bigger better challenge
October 05, 2022 05:47AM
Sorry, used a friend's device to respond while out in the field doing a CITO. CheekyBrit here, Yes, you can absolutely adapt this checker for a challenge of your own. Make sure it gets adjusted to reflect your GC code.
Re: GC9XA74Bigger better challenge
October 05, 2022 06:28AM
LOL, was totally confused by the request to be credited by a random LOL.

Thanks for clarifying. :)

On it now
Re: GC9XA74Bigger better challenge
November 09, 2022 06:44PM
Awesome coding. I would like to use this challenge code for a challenge in Pennsylvania. Someone posted about it in the geocaching Facebook page and said it couldn't be published.

I was thinking of dropping the threshold for the challenge to 12 after dropping the "not chosen" and "wherigo" types. Is that possible and reasonable?

I really should have learned Lua years ago.

Thank you.
Sorry, only registered users may post in this forum.

Click here to login