Datacore Voyage Calc

JiveDutch

Some voyages "predictions" are really far off the mark.
I notice it happens quite often that a voyage that was calculated at 10.45mins will fail at 9.30 occasionally, and more often than my head would assume around the 10hr mark ... ( failing at 9.52, or waiting for dilemma with 21 AM remaining )..

It occured to me the datacore bot gives predictions on duration via discord , however the predictions of the bot seem alot different than prodiction through the site..

Datacore.app/voyage : 10hr 48mins

Then passing the results to the client, and getting the following numbers to feed the bot..

JiveToday at 4:15 PM
-d voytime 13809 13533 8169 4454 4248 4247 2850

DataCore BOT Today at 4:15 PM
Estimated voyage length of 10h 21m (99% worst case 9h 35m). 90% chance to reach the 10hr dilemma; refill with 127 dil for a 100% chance to reach the 12hr dilemma.

Why does 1 calc say 10.48, and the other 10.21 ??

Another "observation" i made .. is that voyages with MED as pri/sec tend to fail alot more than others, which i observe as these are the ones i find stuck just before the dilemma most often... And almost all voyages are areoun 10.30m with my roster is (according to the site predictions)

TemporalAgent7

The voyage calculator on the website is a port of IAmPicard's C++ implementation, while the bot's is an adaptation of Chewable's JavaScript implementation. Hence the differences in estimates.

That said, those are just estimates and should be taken as such. I haven't heard of widespread mis-estimation so it could be selection bias, but if we get more data (good or bad) that could help inform it.

JiveDutch

I recall there was a google form in the past collecting the voyage results ..
Is there anything like it currently to share the results ??

Between the maths versions ... a difference of 27mins seems rather significant
Personally i would rather be pretty dang sure the voyage would reach the dilemma or not ... so i could choose to recall at 8 due to not having the time to babysit ... for example ..

JiveDutch

As i forgot to recall last night .. i let it run on ... surprisingly it went on for quite a run .... 11.28 , with 21AM in the end .. voyage was Sci/Dip
It went over 1hr7m over the bot, and 40m over the site estimate ... i guess this was a lucky RNG ??

=j= Zathras

An estimate is an estimate. RNG is always a factor.

guest_866948680826881

There are some things to consider when trying to compare the two algorithms: IAmPicard's one is taking into account proficiency stats of each character separately from its core stats using them to calculate a transition period (when both loosing and wining a hazard is possible) for each skill while bot's algorithm de facto does not (the input stats for each skill is the sum of all the characters' voyage stats, so proficiency stats are included but not specifically stated). So IamPicard's algorithm is de facto more accurate. Bot's algorithm as I recall is specifically stated that reduces its prediction outcome by 20 mins to be on the safe side (which is correct considering that the voyages prediction-to-outcome follows a normal distribution with a standard deviation of roughly 20 mins). The said s.d. largely varies depended on the voyage characters choise: if you send gauntlet heroes with large proficiency stats on your voyage then you should expect a higher than 20 mins standard deviation which means that your voyage has the potential to last for much shorter or much longer time than if you have had chosen more stable voyage characters (less proficient) of the same overall voyage stats.

Bound2Fate

I just noticed that the Datacore voyage calculator can now work on mobile devices.

Sorry if I'm straying from the OP's main topic, but the thread title seems an appropriate place for this bit of information.

JiveDutch

Had another "interesting" encounter with the calc ..Voyage calc was set at 10.41... voyage failed at 9.58
RNGesus is in a bad mood today.

Commodore Pack

JiveDutch wrote: »

Had another "interesting" encounter with the calc ..Voyage calc was set at 10.41... voyage failed at 9.58
RNGesus is in a bad mood today.

At times like that I prefer to think of it as RNGeelzebub.

Prime Lorca

Is there a way to input the percentage of proficiencies in the voyage score? That seems like something which may interest a friend of mine.

~peregrine~

Prime Lorca [10FH] wrote: »

Is there a way to input the percentage of proficiencies in the voyage score? That seems like something which may interest a friend of mine.

I don’t know if DataCore has that proficiencies input, but @Chewable C++ ’s voyage estimator does (https://codepen.io/somnivore/full/Nabyzw):

[VIP0]USS Dart

I use Datacore exclusively to set up my voyages. Today, when I did so, the past event’s crew were excluded. I changed the setting to include them, but it seems like Datacore doesn’t realize that the event is over.

=j= StephLovesVoyager

[VIP0]USS Dart wrote: »

I use Datacore exclusively to set up my voyages. Today, when I did so, the past event’s crew were excluded. I changed the setting to include them, but it seems like Datacore doesn’t realize that the event is over.

It's showing this weekend's event for me, is this what you're seeing?:

D@MN

[VIP0]USS Dart wrote: »

I use Datacore exclusively to set up my voyages. Today, when I did so, the past event’s crew were excluded. I changed the setting to include them, but it seems like Datacore doesn’t realize that the event is over.

Have you confirmed you refreshed your Player data?
I can't recreate this.

JiveDutch

Is that not the result of the new system displaying the event characters all week ?
Scoundrels are event bonus the whole mega...

[VIP0]USS Dart

Steph loves Voyager wrote: »

[VIP0]USS Dart wrote: »

I use Datacore exclusively to set up my voyages. Today, when I did so, the past event’s crew were excluded. I changed the setting to include them, but it seems like Datacore doesn’t realize that the event is over.

It's showing this weekend's event for me, is this what you're seeing?:

It’s now showing the current event data. I’m not sure if it always was, or if there was a lag. I’ll pay closer attention next week and report back. My apologies.

D@MN

[VIP0]USS Dart wrote: »

Steph loves Voyager wrote: »

[VIP0]USS Dart wrote: »

I use Datacore exclusively to set up my voyages. Today, when I did so, the past event’s crew were excluded. I changed the setting to include them, but it seems like Datacore doesn’t realize that the event is over.

It's showing this weekend's event for me, is this what you're seeing?:

It’s now showing the current event data. I’m not sure if it always was, or if there was a lag. I’ll pay closer attention next week and report back. My apologies.

Also check if you've brought in a fresh copy of your Player data since the event change over, could definitely be that.

=j= StephLovesVoyager

[VIP0]USS Dart wrote: »

Steph loves Voyager wrote: »

[VIP0]USS Dart wrote: »

I use Datacore exclusively to set up my voyages. Today, when I did so, the past event’s crew were excluded. I changed the setting to include them, but it seems like Datacore doesn’t realize that the event is over.

It's showing this weekend's event for me, is this what you're seeing?:

It’s now showing the current event data. I’m not sure if it always was, or if there was a lag. I’ll pay closer attention next week and report back. My apologies.

All good no need to apologise

JiveDutch

RNGsus numbers for the weekend ....
Volyage calculated at 10h 52m ... got lucky with 41AM left at the 10H dilemma instead of a crash just before it ..
Voyage was CMD/Dip , with 14166,13701,6971,5993,3559,4743 and 2875 AM

Datacore bot was more in line with reality
Estimated voyage length of 10h 28m (99% worst case 9h 41m). 95% chance to reach the 10hr dilemma

Is there a way we can get the datacore-bot estimate added to the website plz ....