Bug - Fixed Outskirts combat rate incorrect

[rlbond86]

The numbers given in appearance_rates() do not appear to be correct. For example:

> ash appearance_rates($location[outskirts of the knob])

Returned: aggregate float [monster]
Knob Goblin Assistant Chef => 10.0
Knob Goblin Barbecue Team => 10.0
none => 60.0
Sleeping Knob Goblin Guard => 10.0
Sub-Assistant Knob Mad Scientist => 10.0

The wiki article says the combat% is 79.2, but this return value seems to imply it is only a 40% combat rate, a number which seems far too low to match the actual game parameters. It's not just this entry either -- nearly all of them seem to be incorrect. Sometimes, the appearance of certain monsters is higher, for example in the airship, the ISoRE is listed as 20.0, though I've seen no documentation that they are twice as likely to appear.

 

[Fluxxdog]

The Irritating Series of Random Encounters is more likely than the other combats. There are two instances of this monster in this area, with one of every other monster, leading to 22.5% of combats being the Irritating Series, and 15.5% of combats being other monsters. With no combat modifiers, this becomes 18.0% encounter rate for the Irritating Series, and 12.4% encounter rate for the other monsters.

 

[lostcalpolydude]

That number is modified by combat modifiers. At the time you did that, you were running +10% noncombat. Mafia thinks the zone has a 50% combat rate, as can be seen in http://kolmafia.svn.sourceforge.net/viewvc/kolmafia/src/data/combats.txt . There is nothing wrong with the appearance_rates() function.

 

[rlbond86]

That number is modified by combat modifiers. At the time you did that, you were running +10% noncombat. Mafia thinks the zone has a 50% combat rate, as can be seen in http://kolmafia.svn.sourceforge.net/viewvc/kolmafia/src/data/combats.txt . There is nothing wrong with the appearance_rates() function.

You're correct about the NC rate, and at least that explains why the numbers were a bit weird, but the percents are not accurate. For example, Fluxxdog's wiki link shows that the ISoRE is not twice as likely as other combats, due to the adventure queue. Additionally, a 50% rate is not correct for Cobb's Knob. If this were true, it would take 64 turns to complete the Cobb's Knob bounty.

 

[lostcalpolydude]

Yes, the data file is probably incorrect, and could be updated. If I was going to update that, I would probably check all zones rather than just updating that one zone. Thread title updated.

I don't expect mafia to account for the queue. To properly do that, it would have to track the last 5 monsters in every zone, and the math would be a mess. Even if it was misleading and pretended the queue is empty, there are still weird cases like astronomers with a rejection chance on top of there being 2 of them. Unless someone decides to be really ambitious and tackle that for a small gain, it's going to stay like that.

 

[Theraze]

If there are specific zones with rates that are off, provide links to the spading.. Since TPTB like to change things, that's not a bug... it's just New Content, and should be posted as such, individually, so it can be implemented. Don't make one big thread that says "Lots of areas are wrong and I'll keep adding in the information as I get it spaded" since that depresses the devs... they can't ever really say that it's 'done' and that you won't find another. Make devs happy and let them have threads with happy green colours.

The adventure queue is not implemented. There is a FReq to implement it, but since it would take several hundred more preferences, it's currently indefinitely paused.

I think this bug report can be closed as not a bug...

 

[Bale]

Make devs happy and let them have threads with happy green colours.

I think this bug report can be closed as not a bug...

I like the green color also, but I think that 50% instead of 21% is too drastic for me to just make it go green. As rlbond86 noted, our experience with the time it takes to complete the bounty is notable proof of that 50% is wrong.

 

[lostcalpolydude]

I like the green color also, but I think that 50% instead of 21% is too drastic for me to just make it go green. As rlbond86 noted, our experience with the time it takes to complete the bounty is notable proof of that 50% is wrong.

Bounty completion time seems useless for looking at the combat rate. There are a bunch of one-time adventures, which are probably superlikelies. Some noncombats can be skipped and don't make your bounty hunt longer. Others become a combat, and half of the time the combat will give you the bounty item.

It's pretty easy to spade actually, just use +20% combat and spend some turns in the zone. At 30 turns with all combats, it's only a 0.08% chance of 50% combat rate being right.

 

[Theraze]

At the time I said it, it was still saying that the appearance_rates() function was busted, not that Outskirts was incorrect.

 

[rlbond86]

The adventure queue is not implemented. There is a FReq to implement it, but since it would take several hundred more preferences, it's currently indefinitely paused.

I am not saying that the queue should be implemented. What I am saying is that the probability of a specific monster appearing is not linear with the number of copies present. For example, there are two ISoRE encounters (instead of one, like the other adventures), but due to the properties of the queue, ISoREs are not twice as likely.

 

[Veracity]

Which means that, unless we add in knowledge of the queue, KoLmafia will never give you a number you like.

 

[Catch-22]

I am not saying that the queue should be implemented. What I am saying is that the probability of a specific monster appearing is not linear with the number of copies present. For example, there are two ISoRE encounters (instead of one, like the other adventures), but due to the properties of the queue, ISoREs are not twice as likely.

I am not well versed in manners of the queue, because these days I play fairly casually. Correct me if I'm wrong, but if you're adventuring for the first time in that zone (no monsters encountered there yet), aren't your odds of encountering an ISoRE double that of all the others?

 

[rlbond86]

I am not well versed in manners of the queue, because these days I play fairly casually. Correct me if I'm wrong, but if you're adventuring for the first time in that zone (no monsters encountered there yet), aren't your odds of encountering an ISoRE double that of all the others?

This is correct, but in the steady-state, the probability is about 50% more. See On the trail for a more detailed analysis.

 

[lostcalpolydude]

I personally verified the combat rate for the zone while spading a Zombie skill, so I updated the data file in 11418.