Talk:Adventure Queue

Jick revealed some information about the adventure selector algorithm here, and I'd interpret that as saying that the chances of getting an adventure that's already in the queue is not a fixed percentage. I'm editing the article to reflect this interpretation; feel free to change it back if it's disproven. --Poochy 04:36, 26 December 2006 (CST)

The post by Jick you linked to is from December 2005. All of the HCO spading, and the move to a 5-adv queue, are subsequent to this. Please read this thread thoroughly before updating. --DirkDiggler 04:49, 26 December 2006 (CST)

Anyone know exactly how the queue works? We need a page on this.--Snarles2 08:56, 8 October 2006 (CDT)

September 06, 2006 - The adventure and monster queues are now 5 units long rather than 3. If you don't know what this means, don't worry about it... :)

Are there two queues? I always though there was just one adventure queue (with noncombat and combat adventures in it), are the noncombat and combat/monster adbventures separately queued? And how can one use the queues to their advantage? For example, adventuring between two different locations, switching after you get a desired adventure, and switching back, hoping the queue was reset therefore having a "non-reduced" chance of getting it again. (Hope that was comprehendable) How many locations are there queues for? just the current one? last 2? 3? I do know the queue keeps track of recent adventures, and for you next adventure if one of the adventure in the queue is the one randomly chosen, a "coin is flipped" (50/50) chance of giving you that adventure or not and re-randomly piking an adventure for you (forcing "broader" random distribution). Also some adventures have a greater chance of showing up anyway, for example if an area has 4 different adventures, but if in the adventure list for that location #4 is put in twice, so there are "5" adventures being chosen from (1, 2, 3, 4, 4 ), in the queue, are the 4th and 5th considered as different adventures (when "coin flipping"), even though they end up appearing as the same adventure? --JRSiebz (☎|§|‡) 19:51, 8 October 2006 (CDT)

• There are two queues -- combats and non-combats are rolled separately, and once the flavor is chosen only that queue and pool are used.

• Here is how the formula for queue-adjusted encounter frequency is found. With

• E = the number of encounters in an area (e.g., 7 in the Valley)
• C = (1-j) = the chance of retaining an encounter that is selected if it is in the queue
• j = the chance of rejecting an encouter that is selected if it is in the queue
• u = the number of unique types of encounters in the queue (can be 1, 2, 3, 4, or 5)

we get the recursive function

                 1
 inq(E,j,u)  =  --- (1-j)       +      (adventure is chosen, not rejected)
                 E
                 1
                --- j inq(E,j,u)  +    (adventure is chosen, rejected, but chosen again)  
                 E
                u-1
                --- j inq(E,j,u)       (some other q'd adventure chosen, but it's rejected
                 E                      and our adventure is taken

which gives the equation

                       inf
                1-j   \---  (  u j  )^i
 inq(E,j,u)  = -----   >    ( ----- ) 
                 E    /___. (   E   )
                       i=0

But Sum( (a/b)^i ) for i 0 to infinity is just b/(b-a), so

                1-j       E         1 - j
 inq(E,j,u)  = ----- ( ------- ) = -------
                 E     E - u j     E - u j

Similarly we can find that for an adventure not in the queue,

                  1   
 outq(E,j,N) = -------
               e - u j

and you can check that

  u*inq(E,j,N)) + (e-u)*outq(E,j,u) = 1

That is, the chances of drawing any of the u encounters in the queue plus the chances of drawing any of the (e-u) encounters not in the queue totals to 100%. --DirkDiggler 22:59, 9 November 2006 (CST)

I created a better derivation, including the case of unequal encounter rates. --DirkDiggler 14:43, 20 November 2006 (CST)

The probabilities in the last table appeared wrong by inspection (e.g., 3 enc in zone, 1 in queue would have a 11.11 % chance of an adventure in the queue and a 44.44 % chance of an adventure not in the queue, which does not sum to 1.) I think that is the normal way to parse the 'an' in that sentence. If you were to ask people what the probability of rolling an even number on a die is, the common response would be 1/2 rather than 1/6- an defined as 'any' rather than as 'each'. --Foobar 19:51, 16 March 2007 (CDT)

Data

Here is raw data taken with the batrat/ratbats Not that it really matters, but the - is a ratbat, and the | a batrat. There was 1 noncombat in all of this, but that presumably still uses a seperate queue. :)

| | - - | - - | -| - - - | - - | - - | - | | - | - | | | - | - | | - - - - - | | - | | - - | - | | | - | | - - | - - | | - - - - - - | - - | | | | | - - | | - - | - | | | | - - | - | - | - - - | | - | | | | - | | | | | - | - - - - | | | | | - | - | - | - - | | - | - | - - - | - - | - | - - - - | | | - | | | - | |

This is a total of 160 combats. --Starwed 17:40, 4 July 2007 (CDT)

Separate Queues for Each Zone

I do not know if this is the case, but I think I know of a way to find out. if you look at my On The Trail data for Knob Laboratory trailing Very Mad Scientists, I actually make note of this, that immediately after coming off of the trail of Very Mad Scientists, the queue is now "supersaturated" with them (i.e. They are the less likely encounter, yet (on average) 4/5 of the queue slots will have Very Mad Scientists, so The Queue is probably going to toss you 4 or 5 Mad Scientists before it'll toss you another Very Mad). So what I'm thinking, is that if you trail Very Mads, (and get 4 or better yet 5 in a row), then remove On The Trail, go to another area and spend 5 adventures (which would clear out the queue if there is only one per player), then the next 5 adventures should be distributed with the normal rate of the Laboratory. Granted this will take a long time, to get any meaningful data you'd have to repeat this procedure about 200 times to get 1000 adventures in the Knob Laboratory after you've attempted to flush the queue. However, if you get data like I got (80/20 encounter rate rather than 66/33), then that's a pretty strong indication that each zone has its own queue. I might try doing this, but it sounds very, very tedious.--MindlessGames 20:27, 23 March 2008 (CDT)

• The NS13 queue has been confirmed to be 5 units long, as well as having a new queue per zone. Numbers -QuantumNightmare 19:51, 28 October 2008 (UTC)

"special rejection rate"?

The page mentions that some encounters (both c and nc) have a "special rejection rate." I found three other references to this concept in the wiki: 1) The olfaction page mentions divine poppers etc. drive away monsters by setting rejection rate to 100% (in the context that said rejection is overridden by olfaction) 2) Astronomers have a rejection rate of 25% 3) Izchak's store has a rejection rate of 50%. To me, this information seems both relevant and short enough to be included on the page.--Stannius 21:28, 14 January 2011 (UTC)

Rollover

The results at http://kolspading.com/forums/viewtopic.php?f=2&t=175 indicate that the queue does not reset on rollover. --Starwed 20:37, 29 January 2011 (UTC)

URs

Not sure where to put this, but cannon (about a week or so ago) managed to get a nuge's crossbow by using several ten-thousands of clovers, so URs are above clovers. Now, who's up to test SRs? — Cool12309 (talk) 00:59, 5 April 2014 (UTC)