Feature Request: Record order of Daily Dungeon rooms

[VeeArr]

The order in which Daily Dungeon rooms are encountered is turning out to be a useful piece of data for a spading effort. It would be nice if there was a preference that recorded what encounter was in each room. (Only the type of encounter--monster, door, or trap--matters, not the specific monster fought or trap element.)

Ideally, the preference would be something like ????_????_???? at the start of an ascension and fill in each ? as that room's identity is found, finally becoming (for example) MDTD_MTMD_MDTT after the room order is fully-known. This preference resets on ascension.

 

[fronobulax]

Is the order identical for every day and only resets after ascension? (I think you said yes, but..)

Could this be satisfied by a Daily Dungeon script that you run and it creates or adds to or validates a user preference?

 

[VeeArr]

Yes, the order of room types is identical every day until you ascend.

In an ideal world the solution wouldn't require using a script to collect the data because the idea is to use information that you've naturally seen over the course of an ascension to predict other information.

 

[ckb]

Is this necessary?
That is, if you just want ascension info and details about the Daily Dungeon rooms, you could probably write a log parser and have people parse they logs from the past few years and generate thousands of data points. I know I have 10+ years of session logs.

 

[fronobulax]

In an ideal world the solution wouldn't require using a script to collect the data because the idea is to use information that you've naturally seen over the course of an ascension to predict other information.

We have different definitions of ideal world In my ideal world KoLmafia would not have features that could be implemented by scripts. If it is not in the code base then it does not have to be tested, maintained, or fixed and it does not create a possible security hole or something that could be maliciously exploited.

I was proposing using a script to run your DD turns. A side effect of running the script would be that it collects the data. I could have been clearer that I was thinking of a turn burning script and not something that acquired data.

If you are only concerned about using the data for the current ascension then a log parser might be the way to go, provided you had a decent tool set and were comfortable using it.

If you are thinking about historical data and some kind of predictive tool then log parsing is the way to go. My earliest session log is from January 1, 2014 which suggests to me that somewhere there is a disk with even older session logs that I just forgot to copy.

 

[VeeArr]

We have different definitions of ideal world In my ideal world KoLmafia would not have features that could be implemented by scripts. If it is not in the code base then it does not have to be tested, maintained, or fixed and it does not create a possible security hole or something that could be maliciously exploited.

I'm sympathetic to this position, but it doesn't match the reality of what KoLmafia currently is (in my opinion). It tracks all kinds of things that could theoretically instead be determined by going back and parsing the run's logs, because having that information already tracked enables scripts to leverage that information more readily.

I'll acknowledge that this particular case is maybe too niche to warrant tracking within KoLmafia, but let's not pretend that the vast majority of things it tracks already are only tracked because doing so is needed for KoLmafia to function.

 

[VeeArr]

Maybe for a little more context about what we're doing: It has been discovered that you can look at some bits of information about the current ascension and use that determine other information about the ascension. As an example, currently, we're using things like the effects of bang potions to predict the answers to the Mer-kin dreadscroll without actually gathering the various clues, and this is easy because KoLmafia automatically tracks the bang potion effects.

It turns out the daily dungeon room order also has predictive value in the same way, so it would be convenient to have ready access to it, rather than having to go back and figure out what it was after the fact.

 

[fronobulax]

I'm sympathetic to this position, but it doesn't match the reality of what KoLmafia currently is (in my opinion). It tracks all kinds of things that could theoretically instead be determined by going back and parsing the run's logs, because having that information already tracked enables scripts to leverage that information more readily.

I'll acknowledge that this particular case is maybe too niche to warrant tracking within KoLmafia, but let's not pretend that the vast majority of things it tracks already are only tracked because doing so is needed for KoLmafia to function.

I have spent almost two decades arguing against features that were unnecessary because they could be implemented by scripts. Same for niche features, such as things that only work for, or have relevance for, KoLmafia users who have access to the dev server. Obviously there are a lot of battles that I have not won I could probably justify some of your "all kinds of things" by arbitrarily saying a change that exposes data that KoLmafia was already tracking is somehow different from things KoLmafia never needed to track before and does not use internally. I could also keep this going by asking "who is really pretending?" and "what numbers justify "vast majority"?" but I recognize this is a philosophy disagreement and I'm more interested in explaining and am fine with agreeing to disagree. See next post

 

[fronobulax]

Maybe for a little more context about what we're doing: It has been discovered that you can look at some bits of information about the current ascension and use that determine other information about the ascension. As an example, currently, we're using things like the effects of bang potions to predict the answers to the Mer-kin dreadscroll without actually gathering the various clues, and this is easy because KoLmafia automatically tracks the bang potion effects.

It turns out the daily dungeon room order also has predictive value in the same way, so it would be convenient to have ready access to it, rather than having to go back and figure out what it was after the fact.

Context always helps to win me over and often other devs as well.

Who is the "we" who is using things like bang potion effects to predict other things? Where are results being shared or is this more like the beachcombing work which wasn't really shared until a second effort started publishing data? A pet peeve is people who ask for and get changes for spading and then never come back and share results or even thank people. But I have a problem with kids on my lawn too so I can be safely ignored.

If mafia is not going to use the data but it is going to be stored in a preference then the ability to use it to make predictions exists whether mafia collects it or a script does. But if the data could be used by mafia to provide predictive spoilers then mafia is using it and the case for collecting it is much simpler.

When you submit the PR that does this make sure you have some HTML driven tests

 

[VeeArr]

Where are results being shared or is this more like the beachcombing work which wasn't really shared until a second effort started publishing data?

No offense, but this comes across as unnecessarily negative. The effort is public in the ASS Discord.

When you submit the PR that does this make sure you have some HTML driven tests

I'll include at least as many HTML-driven tests as currently exist for tracking Daily Dungeon progress

 

[Ryo_Sangnoir]

I'll acknowledge that this particular case is maybe too niche to warrant tracking within KoLmafia, but let's not pretend that the vast majority of things it tracks already are only tracked because doing so is needed for KoLmafia to function.

Your tone is a bit harsh here. Cool it.

I hold the opposite philosophy to frono, and I prefer things go into Mafia than stay out. So certainly many of the things it does not need to have were added by me

No offense, but this comes across as unnecessarily negative. The effort is public in the ASS Discord.

Come on friend, I too am on the ASS Discord, so I know slightly more than frono, and I can see the spading threads and pick out your https://github.com/VeeArrKoL/seedfinder script, from which I can read the source code and see how the seed relates to various things. Still, I can't see the data -- the linked spreadsheet is practically empty. If it is somewhere in the thread, this is a definition of "public" that means "out there, somewhere, but not easy to find". A wiki page along the lines of https://wiki.kingdomofloathing.com/Deterministic_RNG for "ascension seed" would be appreciated. Also, for the interests of Mafia, I would like it if said page included the algorithms (simple in this case).

 

[VeeArr]

I can see the spading threads and pick out your https://github.com/VeeArrKoL/seedfinder script, from which I can read the source code and see how the seed relates to various things. Still, I can't see the data -- the linked spreadsheet is practically empty. If it is somewhere in the thread, this is a definition of "public" that means "out there, somewhere, but not easy to find".

I don't understand what sort of data you're looking to find. The sum total of the knowledge so far is contained among (1) the chat in the spading channel, (2) the brainstorming area of the data sheet (what few data points were actually collected for determining the algorithms for known seed uses are in the hidden sheets in that workbook--with the exception of the sea data that started the whole project which are instead in the dreadscroll spading spreadsheet), and (3) the seedfinder script (which has a README explaining the general concept and code with the algorithms for all of the currently-known ascension seed uses).

A wiki page along the lines of https://wiki.kingdomofloathing.com/Deterministic_RNG for "ascension seed" would be appreciated. Also, for the interests of Mafia, I would like it if said page included the algorithms (simple in this case).

I agree that this will be a good output at some point but, broadly-speaking, work on the ascension seed has only been going for ~2 days, so no one has distilled it down to this yet. The focus has been more on brainstorming ways to (1) determine the ascension seed and (2) leverage the ascension seed for useful things. (That said, I'm not sure a wiki page is a good way to document the specifics of algorithms.)

 

[fronobulax]

First I have shifted my concerns from strategy to tactics and so am actively trying to improve the implementation as published in a PR.

I'll stand by my comments and note I tired to moderate my tone with humor.

I have had "history" with some individuals concerning KoLmafia. No point in naming names and much of it is probably my fault but the experience does color how I view more recent events.

One example of not sharing is in https://kolmafia.us/threads/beachcomber-fast-and-efficient-beach-combing.23993/

I've been using combo happily for about a year and a half now, but I recently learned something interesting: it doesn't actually have data on all the rare tiles. There is at least one user I could name (and how many others?) who has a personal set of known-to-them but unshared rare tiles.

Since Discord does not support viewing without an account I do not consider the LASS discord channel a public or shared source in the sense that the official KOL forums, the KoL wiki or this site are. Yes I could make an account but I should not be forced to if the goal is to truly share. My attitude is certainly colored by the days when some spading and script writing and distribution took place on protected sites that belonged to clans that were deliberately restricting membership.

Finally in the very early days of KoLmafia there was tension between the developers and users (including script writers). The development team considered working on KoLmafia a labor of love for for the KoL community. As such there was a hope that users and scriptwriters also shared things that benefited the entire community. Unrealistically optimistic given human nature but the hope persists. It is certainly possible that my negativity and snarkiness can be attributed to "The world has changed and I don't like it".