Timed weekly needs to be renamed to GSF weekly

[vicadin]

Please do not tell me that Ben "RNGesus" Irving took a break from Anthem to infect the timed weekly mission? I just don't get how for weeks that the timed weekly seemed to be a rotating mission to suddenly RNG based.

 

Musco, you guys seriously need to look into putting some kind of "bad luck protection" into anything that has to do with RNG. I get that it's a huge part of the game, but there should be some limit. In this example GSF should still have a chance to show up HOWEVER, it's chance of showing up should not be exactly the same as the others. In fact, I think it's been 9-10 weeks since it was the PVP, 7-8 weeks since the SM Ops, and 4 weeks since the MM FPs. That would mean that PVP should have the best chance of getting picked. Sure, GSF should still have a chance...but it should be less than the others.

[sharkfishman]

Aww. I was hoping we'd go 4 for 4. GSF is life Eric. =(

 

After one week of PVP it will go back to 3 more GSF, lol.

[Nemmar]

258 Rating Gear is the main purpose of the Masterwork Data Crystals, and the only content in the game you need it for is Nightmare Mode Gods of the Machine, so how come you earn nightmare progression gear from an auxiliary game mode? This is hilarious and makes it obvious that the Dev Team doesn't even bother to play the game, or you would notice how ridiculous this is.

 

Except it's not. Gear matters in PvP and honestly the pursuit of power is the point of an RPG. So, they have to give everyone a path towards getting the gear. It should not be reserved for a select few who then can ravage through other game modes with it.

Challenging content should allow you to aquire the best gear faster. It does not entitle you to it's exclusivity.

Edited May 21, 2019 by Nemmar

[sdom]

This thread has had me thinking about the problem that your facing.

I lack an intimate knowledge of your games source code which is I'm sure. what you go for.

But what I do know is the RNG is a script.

And that the matches that can be "potentially" called - would be variables so they exist in the script in question in a way convenient to refer to within the code.

 

Simply piping the result of the decision by the RNG into another variable, and adding a section at the bottom of your script that takes the decision of the RNG and looks for repetitiveness within past weeks - should make it possible with something like an if loop to cap any repetitiveness within the RNG, a process that it clearly needs in order to make it truly random it seems.

 

If eric could be so kind to pass that information along to the devs, there's no reason that solution can't be experimented with right away and potentially rolled out within a day or two.

It's a real simple job man.

I'd do it for you but you don't want me in at your code lol

[sdom]

A child could code their first script to selectively choose a number between 1 - 5.

The answer to executing this script 5 times could have been: 1.2.3.4.5

This in no way indicates the choice made by the script is random.

As it is a script called once with no memory or recognition of its own output.

The truth of the matter is the probability to get each result is an even split.

 

20% for 1

20% for 2 and so on.

 

Its when the child becomes an adult and returns to this script with enough perspective to recognise this philosophy, that allows them to develop a method of achieving the same result with more effective code.

 

Your devs say this is random?

Well I say fire them because it doesn't seem like that the result is the same as the implimentation, and if they're still trying to tell you otherwise, then they're incompotent.

Edited May 21, 2019 by sdom

[xJediRyanx]

A child could code their first script to selectively choose a number between 1 - 5.

The answer to executing this script 5 times could have been: 1.2.3.4.5

This in no way indicates the choice made by the script is random.

As it is a script called once with no memory or recognition of its own output.

The truth of the matter is the probability to get each result is an even split.

 

20% for 1

20% for 2 and so on.

 

Its when the child becomes an adult and returns to this script with enough perspective to recognise this philosophy, that allows them to develop a method of achieving the same result with more effective code.

 

Your devs say this is random?

Well I say fire them because it doesn't seem like that the result is the same as the implimentation, and if they're still trying to tell you otherwise, then they're incompotent.

 

I asked my 3 year-old to write out a line of code to choose a number between 1 and 5. He couldn't do it. I'm thinking of giving him up for adoption.

[DarthSpuds]

Hey all,

 

I double checked with the team and it is not a bug, it was just that the 1 in 4 chance pulled GSF 3 weeks in a row (not the best if you aren't a GSF fan). We did verify though that next week is not GSF.

 

Thanks all!

 

-eric

 

How did you verify that if it is a genuinely random 1 in 4 chance?

 

Sounds to me like Bioware choose what the weekly is and someone just forgot to change it, because you guys don't play the game we do,

 

All The Best

[sdom]

I asked my 3 year-old to write out a line of code to choose a number between 1 and 5. He couldn't do it. I'm thinking of giving him up for adoption.

 

Obviously the problems not your three year old its you.

Technology work is one of the most profitable job industries in the world and also paramount to know for digital security.

There are plenty of toys, for example:

https://techcrunch.com/2017/08/31/the-littlebits-droid-inventor-kit-lets-you-build-an-r2-d2-of-your-very-own/?guccounter=1&guce_referrer_us=aHR0cHM6Ly93d3cuZ29vZ2xlLmNvbS8&guce_referrer_cs=R0xpiB-Q1CSloGH1-nKfSw

 

Keeping in with the star wars theme, the little bits programmable R2-D2. Something you could build together, and instead of raising a completely average joe only special to you, you can be parent to a child earning 150k USD a year + for AI development, who is critical to the worlds infrastructure, with the right inspiration.

 

Of course all this would need you to care about your child and the face of the threat landscape that will most certainly impact their adult life, but obviously thats out the question since you're considering adoption.

Edited May 21, 2019 by sdom

[xJediRyanx]

Obviously the problems not your three year old its you.

Technology work is one of the most profitable job industries in the world and also paramount to know for digital security.

There are plenty of toys, for example:

https://techcrunch.com/2017/08/31/the-littlebits-droid-inventor-kit-lets-you-build-an-r2-d2-of-your-very-own/?guccounter=1&guce_referrer_us=aHR0cHM6Ly93d3cuZ29vZ2xlLmNvbS8&guce_referrer_cs=R0xpiB-Q1CSloGH1-nKfSw .

You must be really fun at parties.

[TUXs]

Hey all,

 

I double checked with the team and it is not a bug, it was just that the 1 in 4 chance pulled GSF 3 weeks in a row (not the best if you aren't a GSF fan). We did verify though that next week is not GSF.

 

Thanks all!

 

-eric

GSF needs to be removed from this. This was ridiculous Eric.

[phalczen]

Sure, except for the fact not everyone, like u it seems, has been around since ossus gear launch. Anything i miss out on is sorely missed cause ive been back a few weeks personally. 4 months is a long time, might as well have said there is no reason to gear ever, cause there will be another expansion with a new lvl cap, eventually.

 

Well, you only need gear high enough for you to complete content you like at a difficulty level comfortable for you. For some people, myself included, achieving the highest possible gear rating, best-in-slot configuration for my toons is part of the power progression. In as much as I'm always trying to improve to the best of my ability, reaching BiS is part of the game's content, for me. I'm sure it is for others, too. Just as for many others its not, and 258 gear is way, way overkill for KOTFE/KOTET chapters (even MM), Eternal Championship, planetary heroics, and other solo-focused content.

 

If you are an "exclusive" pvp'er, perhaps you are concerned that your lack of effectiveness is due to the gear difference, and not your rustiness from your break from the game. I would submit to you that getting rating 240 augments is a far more useful investment in terms of your pvp effectiveness than the difference between bolstered 252 and 258, except for the mainhand and offhand. Mathematically, the increase in stat budget relative to units of the prized Charged Matter Transubstantiator, the crafting material used in making both 236/240 augments and 258 gear, is in favor of the 236 augment by like ten times. In other words, you'll see far more difference in dps/healing from upgrading 228 augments to at least 236, than you will from the difference between bolster 252 and 258. You don't even need to set foot on Ossus to get the stuff you need for 236 augments. Then Bolster according to Hottie's guides and get the 252 mainhand from the One-time Flesh and Steel world boss mission on Ossus. You will be at like 97-98% of the stat budget of someone in 258 gear.

 

If you're not an exclusive pvp'er, and you don't do Master Mode Gods from the Machine, then you are in luck. You can easily complete the content on Ossus with a gear rating around 240. I know this because the one thing galactic command is still good for is getting set bonus gear for alternate disciplines. So, in the time since 5.10, I've been able to build not only the primary set for the main discipline but an off set. This is from a combination of twinking from alts, saving item modifications instead of replacing them, and 246-rated crafted left side stuff. I've tanked the world bosses with 244 average item rating, no problem. Furthermore, even if you want to do hard mode content or MM of anything less than GotM, all that stuff was balanced around 242 gear.

 

Hey all,

 

I double checked with the team and it is not a bug, it was just that the 1 in 4 chance pulled GSF 3 weeks in a row (not the best if you aren't a GSF fan). We did verify though that next week is not GSF.

 

Thanks all!

 

-eric

 

The probability of that happening is 0.25^3, which is about a 1.5% chance of occurring.

 

GSF needs to be removed from this. This was ridiculous Eric.

 

Removing it isn't ridiculous. Would you, or anyone, be saying this if the outcome had been "Earn 8 medals in a single unranked warzone that you don't even need to stay till the end for" three weeks in a row? You and I both know the answer to that, TUX.

 

Keeping some sort of log so that you don't get duplicates two weeks in a row isn't unreasonable, and changing 3 wins to something else like GSF medals or completed matches isn't unreasonable either, but removing it simply because you don't like it is what's ridiculous TUX

[spatnatz]

Before I get into all this, let me say that I am just as disgruntled as the rest of you... not only about getting 3 GSFs in a row, but about the bad design of the weekly quests in general. The time required for some of them is definitely not reasonable.

 

Consider this: There are 4 weekly quests to pick from. Each week, a quest has a 1 in 4 chance of being selected. Therefore the chance of a given week being GSF is 1/4 or 25%.

 

Suppose the week immediately after was also selected as GSF. So that becomes a probability of (1/4) * (1/4) = 1/16, or 0.0625%.

 

And you guessed it... the third week? That's (1/4) * (1/4) * (1/4) = 1/64 , or a 1.56% chance. Let's interpret this statistically... This means that, given 64 3-week "sets" of timed weeklies, we can expect one of those to have all 3 GSF.

 

This seems like a very low chance, to be sure. However, what about the probability of the first week being GSF, the second week being PVP, and the 3rd week being operations? The calculations for this would be exactly the same, resulting in a 1/64 chance for this particular event as well. So all that really means is that there are 64 possible configurations for any 3-week set of events, each with an equal probability of occurring (if we assume they are in a specific order).

 

So, while this event may seem unique, it is no less unique than any other event. It just happens to stick out like a sore thumb because it's repeating something three times. (A simpler example: Same probability for rolling three 1's in a row, as it is for rolling 1, 2, 3 in order on a standard 6-face die)

 

All that said... it does seem a bit fishy that we had (to my recollection) 2 PvP weeks in a row, followed by 2 MM weeks in a row, followed by 3 GSF weeks in a row. Let's ignore the 3rd gsf week for just a moment. I can say for a fact (you're welcome to check my work on this) that the chances of getting any given event twice in a row, and to then have that repetition happen *three* times in a row, is also 1/64. And this is for each 6-week period... meaning that the probability of this happening more than once in a 384-week period is even smaller. So we'd better not see something like 2 pvp, 2 MM, and 2 GSF anytime soon, or I'm calling bananas on this.

 

TL;DR:

 

Yes, the GSF weekly 3x happened by random chance, but it's a pretty small one and not any less likely than any other 3 combinations of events.

 

Getting two events in a row three times in succession (ignoring the third repetition of GSF) is definitely more of a statistical anomaly, and we'd hope not to see something like this anytime soon if the system is truly random, as BW claims it is.

 

If I got anything wrong in my analysis, feel free to correct. Thanks for reading...

Edited May 22, 2019 by spatnatz

[phalczen]

I bet the system is "rolling" the timed weeklies at the beginning of a patch cycle, and not at the start of every week. i.e. the timed weeklies were determined when 5.10.2 launched and for the intervening weeks between then and when they expected to launch 5.10.3. I bet that's why when patches are delayed things like conquest get screwed up. It would also explain why the dev team was able to "look up" what next week's timed weekly will be.

[JediBoadicea]

Hey all,

 

I double checked with the team and it is not a bug, it was just that the 1 in 4 chance pulled GSF 3 weeks in a row (not the best if you aren't a GSF fan). We did verify though that next week is not GSF.

 

Thanks all!

 

-eric

 

I actually really do like GSF, but this feels a little silly. More specifically, I think it's a little silly to have made an option of only 4 activities random. That's a low number of activities, not like rolling against a loot table of 10-30+ items. I can't help but feel that it would make more sense to just set the four missions on a fixed rotating schedule, one after the other. Not only do you avoid a fluke RNG instance like this one that makes some people unhappy, but it would also allow guilds to plan more carefully for activities around it. That keeps the difficulty barrier evened out as well, preventing multiple super easy 8 medals in a row just as much as it prevents multiple Ops in a row.

[spatnatz]

Hey all,

 

I double checked with the team and it is not a bug, it was just that the 1 in 4 chance pulled GSF 3 weeks in a row (not the best if you aren't a GSF fan). We did verify though that next week is not GSF.

 

Thanks all!

 

-eric

 

 

I actually really do like GSF, but this feels a little silly. More specifically, I think it's a little silly to have made an option of only 4 activities random. That's a low number of activities, not like rolling against a loot table of 10-30+ items. I can't help but feel that it would make more sense to just set the four missions on a fixed rotating schedule, one after the other. Not only do you avoid a fluke RNG instance like this one that makes some people unhappy, but it would also allow guilds to plan more carefully for activities around it. That keeps the difficulty barrier evened out as well, preventing multiple super easy 8 medals in a row just as much as it prevents multiple Ops in a row.

 

I agree.. it's not "the best", regardless of whether you are a GSF fan or not. When even the hardcore GSFers are unhappy, we've got a problem.

 

Clearly no one would complain if it was 3 pvp weeks in a row... so the root of the issue really has to do with the time required on the quests. If all quests took reasonable (and hopefully somewhat similar) amounts of time , then this thread would not exist even if GSF was 8 weeks in a row.

 

Devs should also keep in mind that GSF only pops around certain times of the day/week, while pvp pops more or less 24/7.

Edited May 22, 2019 by spatnatz

[Lakemine]

Well, you only need gear high enough for you to complete content you like at a difficulty level comfortable for you. For some people, myself included, achieving the highest possible gear rating, best-in-slot configuration for my toons is part of the power progression. In as much as I'm always trying to improve to the best of my ability, reaching BiS is part of the game's content, for me. I'm sure it is for others, too. Just as for many others its not, and 258 gear is way, way overkill for KOTFE/KOTET chapters (even MM), Eternal Championship, planetary heroics, and other solo-focused content.

 

If you are an "exclusive" pvp'er, perhaps you are concerned that your lack of effectiveness is due to the gear difference, and not your rustiness from your break from the game. I would submit to you that getting rating 240 augments is a far more useful investment in terms of your pvp effectiveness than the difference between bolstered 252 and 258, except for the mainhand and offhand. Mathematically, the increase in stat budget relative to units of the prized Charged Matter Transubstantiator, the crafting material used in making both 236/240 augments and 258 gear, is in favor of the 236 augment by like ten times. In other words, you'll see far more difference in dps/healing from upgrading 228 augments to at least 236, than you will from the difference between bolster 252 and 258. You don't even need to set foot on Ossus to get the stuff you need for 236 augments. Then Bolster according to Hottie's guides and get the 252 mainhand from the One-time Flesh and Steel world boss mission on Ossus. You will be at like 97-98% of the stat budget of someone in 258 gear.

 

If you're not an exclusive pvp'er, and you don't do Master Mode Gods from the Machine, then you are in luck. You can easily complete the content on Ossus with a gear rating around 240. I know this because the one thing galactic command is still good for is getting set bonus gear for alternate disciplines. So, in the time since 5.10, I've been able to build not only the primary set for the main discipline but an off set. This is from a combination of twinking from alts, saving item modifications instead of replacing them, and 246-rated crafted left side stuff. I've tanked the world bosses with 244 average item rating, no problem. Furthermore, even if you want to do hard mode content or MM of anything less than GotM, all that stuff was balanced around 242 gear.

 

 

 

The probability of that happening is 0.25^3, which is about a 1.5% chance of occurring.

 

 

 

Removing it isn't ridiculous. Would you, or anyone, be saying this if the outcome had been "Earn 8 medals in a single unranked warzone that you don't even need to stay till the end for" three weeks in a row? You and I both know the answer to that, TUX.

 

Keeping some sort of log so that you don't get duplicates two weeks in a row isn't unreasonable, and changing 3 wins to something else like GSF medals or completed matches isn't unreasonable either, but removing it simply because you don't like it is what's ridiculous TUX

 

Well to be honest, I don't think anyone would be complaining if the pvp one was 3 weeks in a row. The reason why is because look at the difficulty that is required for the pvp one vs. the GSF one. The pvp one is just earn 8 medals, the GSF one is WIN 3 matches. There is a massive difference between those. Same with do one GF OPs vs 4 MM fps. If they made it to do 2 GSF matches or earn 5 medals in GSF, people would not be complaining as much.

[Lakemine]

I agree.. it's not "the best", regardless of whether you are a GSF fan or not. When even the hardcore GSFers are unhappy, we've got a problem.

 

Clearly no one would complain if it was 3 pvp weeks in a row... so the root of the issue really has to do with the time required on the quests. If all quests took reasonable (and hopefully somewhat similar) amounts of time , then this thread would not exist even if GSF was 8 weeks in a row.

 

Devs should also keep in mind that GSF only pops around certain times of the day/week, while pvp pops more or less 24/7.

 

^ This

[Batwer]

I find the effort ridiculous. If you start with the Pvp Quest. You only have to once in a WZ for a stone. To do it fair. Would the 2x GSF without winning and 1x MM FP....

[JattaGin]

I double checked with the team and it is not a bug, it was just that the 1 in 4 chance pulled GSF 3 weeks in a row (not the best if you aren't a GSF fan). We did verify though that next week is not GSF.

 

Just to confirm: You have a random 1-4 script going on, but you can manually see (and manipulate?) what the next weeks will bring? Am I the only one irritated by this method?

 

Hopefully this doesn't come across as rude Eric, but if it can be checked manually, why was it not checked to make sure this issue doesn't happen? Did it just not come to mind for the devs? Just asking for the reason, and hopefully things can be looked at to prevent this issue from happening again? Thanks

 

OK, I'm apparently not the only one.

 

258 Rating Gear is the main purpose of the Masterwork Data Crystals, and the only content in the game you need it for is Nightmare Mode Gods of the Machine, so how come you earn nightmare progression gear from an auxiliary game mode? This is hilarious and makes it obvious that the Dev Team doesn't even bother to play the game, or you would notice how ridiculous this is.

 

This is the change that Mr. Kanneg brought to this game. The complaints after 5.0 were that BW only focused on story. Mr. Kanneg's task was to shift this focus back to... well, everything. Since there are different reasons why certain parts of the game are not well appreciated by different players, they use different methods to force these different players to play these different things nonetheless. Look at how GSF buzzes with life these days (I assume it does as I ignore the 258 gear progression)! The metrics don't lie. You can think about Mr. Kanneg's methods whatever you want (I personally disapprove), but he properly does what (I assume) he was hired for, I give him that.

 

I asked my 3 year-old to write out a line of code to choose a number between 1 and 5. He couldn't do it. I'm thinking of giving him up for adoption.

 

Don't give up on him yet. He could still end up becoming a professional footballer.

 

Last party I was at I put my friend through a table because we were with someone we wanted to ditch and had to get kicked out. I then proceeded to plan a GTA style heist of the local chip shop, apparently retained the cognitive abilities to enact said plan when I blacked out. Here was me in the middle of the knife crime capital of europe trying to steal chips over the counter.

 

Woke up the next day in a thong with a black eye and no phone.

 

Ah, that was you. Small world.

[Morniel]

You could actually, you know, try to learn GSF and enjoy it for the great game that it is. Too many people complain about GSF and don't give it a fair chance or appreciate how deep and well-designed it really is.

 

If you seriously hate GSF that much, don't do the weekly. Get your crystals elsewhere. You lose out on one stinking crystal, big deal.

 

Before Ossus I had never played a single GSF match (despite playing since beta and rank 95 valor on my main) and was not a big fan of this aspect, but admittedly the timed weekly seems to have had the desired effect on me. I just returned to the game a few weeks ago so, in order to "catch up" and get as many crystals as I could, I put in a serious effort to do GSF and am actually enjoying it. The 25,000 fleet requisition from the starter quest helps a great deal.

 

That being said, there are typically multiple means of obtaining most other currencies in the game like UC. Even if you enjoy GSF, the queues pop very rarely and one of the other quests to kill the Geonosian Queen on VM is out of the question for all except hardcore raiders with a regular group. That leaves only the 2 world boss quest for a single crystal a week. Maybe this is Bioware's intent. If you are not engaged in various aspects of the game then we have to accept that gear progression will be slow.

Edited May 22, 2019 by Morniel

[Joonbeams]

...

Consider this: There are 4 weekly quests to pick from. Each week, a quest has a 1 in 4 chance of being selected. Therefore the chance of a given week being GSF is 1/4 or 25%.

 

Suppose the week immediately after was also selected as GSF. So that becomes a probability of (1/4) * (1/4) = 1/16, or 0.0625%.

 

And you guessed it... the third week? That's (1/4) * (1/4) * (1/4) = 1/64 , or a 1.56% chance. Let's interpret this statistically... This means that, given 64 3-week "sets" of timed weeklies, we can expect one of those to have all 3 GSF.

 

...

 

This is the right math if the question is "what's the probability of an event with 0.25 chance occurring 3x in a row?"* The question for me though is how do you achieve a 25% probability in code? Now, it's been a while since I've had to think about this, so others should weigh in where I'm off, but I see at least two ways (probably several, but for conceptual purposes and simplicity I'm focusing on two).

 

One is, create a "randomizer" function and run that function against the 4 variables (a,b,c,d) you want to return. The randomizer picks any of the variables at random. In theory, "a" should return 25% of the time, "b" 25% and so forth...

 

But another approach, and IMO the better approach, is to write a function that returns "a" "b" "c" and "d" with equal probability (i.e. 25%). In other words, write the function to force the 25% probability.

 

You might end up in the same place as the first option. But the first option is more error prone in that it depends on this "randomizer" actually doing it's job properly and not getting bugged, etc. The second is also "cleaner" because you're not really looking for random chance - what you actually want is 25% chance for each possibility (I hope I'm being clear on this subtle distinction). If nothing else, the second approach would be easier to debug.

 

All said, there could be some small bug somewhere in the code (I know Eric said there isn't), but also I think (in the footnote below) it may be more likely to see an event like this one time after the probability of occurrence in a single attempt is factored over multiple attempts. For example, 1.5% seems low, but if that was the rate of airline crashes daily (and for the sake of argument, the probability of crash was the same for all planes), we'd see crashes daily based on traffic volume (and no one would fly!)...

 

*The right question might actually be: "what's the probability of an event with 0.25 chance occurring 3x in a row, out of, say, 24 tries (i.e. once a week, for 6 months)? So where X = (0.25)^3, I believe this would be (1 - X)^24 to get the probability of this not happening once in 24 attempts, or ~ 39% (or 1 - "result") chance of happening. In other words, over the amount of times this actually could happen, the more tries there are, the higher the probability that this could happen once. But this means at some point, there is >1 chance of this happening, which seems odd. Anyone else have thoughts?

[spatnatz]

This is the right math if the question is "what's the probability of an event with 0.25 chance occurring 3x in a row?"* The question for me though is how do you achieve a 25% probability in code? Now, it's been a while since I've had to think about this, so others should weigh in where I'm off, but I see at least two ways (probably several, but for conceptual purposes and simplicity I'm focusing on two).

 

One is, create a "randomizer" function and run that function against the 4 variables (a,b,c,d) you want to return. The randomizer picks any of the variables at random. In theory, "a" should return 25% of the time, "b" 25% and so forth...

 

But another approach, and IMO the better approach, is to write a function that returns "a" "b" "c" and "d" with equal probability (i.e. 25%). In other words, write the function to force the 25% probability.

 

You might end up in the same place as the first option. But the first option is more error prone in that it depends on this "randomizer" actually doing it's job properly and not getting bugged, etc. The second is also "cleaner" because you're not really looking for random chance - what you actually want is 25% chance for each possibility (I hope I'm being clear on this subtle distinction). If nothing else, the second approach would be easier to debug.

 

All said, there could be some small bug somewhere in the code (I know Eric said there isn't), but also I think (in the footnote below) it may be more likely to see an event like this one time after the probability of occurrence in a single attempt is factored over multiple attempts. For example, 1.5% seems low, but if that was the rate of airline crashes daily (and for the sake of argument, the probability of crash was the same for all planes), we'd see crashes daily based on traffic volume (and no one would fly!)...

 

*The right question might actually be: "what's the probability of an event with 0.25 chance occurring 3x in a row, out of, say, 24 tries (i.e. once a week, for 6 months)? So where X = (0.25)^3, I believe this would be (1 - X)^24 to get the probability of this not happening once in 24 attempts, or ~ 39% (or 1 - "result") chance of happening. In other words, over the amount of times this actually could happen, the more tries there are, the higher the probability that this could happen once. But this means at some point, there is >1 chance of this happening, which seems odd. Anyone else have thoughts?

 

Perhaps what you might be getting at (???) is that most code-based random number generators are almost never truly random--the code starts with some kind of "seed." (See MIT's view on the issue)

 

I suppose that means that my assumption, and that of phalczen, are based on a rather large presumption of randomness. But then, this low probability not only of 3 in a row, but having repeats 3 times in a row, would seem to imply otherwise. I am inclined to believe what others are saying, that there is some level of deliberation in this facade of randomness.

 

(See my post before last, for context...)

 

(My big long boring post)

Edited May 22, 2019 by spatnatz

[Toraak]

Even as someone who likes GSF, I think the rotating weeklies were done incorrectly. It should NEVER have had any randomness to it. They should have been on a set rotating schedule. 1st week: PvP, 2nd week: SM Ops, 3rd Week: GSF, and 4th week: MM FP's. Then repeat in the same order.

 

No they shouldn't take out the GSF timed weekly anymore then they should remove any of the others. Various people like different content and should have access to the mission at least once ever 4 weeks.

[Joonbeams]

Perhaps what you might be getting at (???) is that most code-based random number generators are almost never truly random--the code starts with some kind of "seed." (See MIT's view on the issue)

 

I suppose that means that my assumption, and that of phalczen, are based on a rather large presumption of randomness. But then, this low probability not only of 3 in a row, but having repeats 3 times in a row, would seem to imply otherwise. I am inclined to believe what others are saying, that there is some level of deliberation in this facade of randomness.

 

(See my post before last, for context...)

 

This is partially what I'm getting at. "Randomizers," like RNGs aren't all created equal. I'd suspect the devs are using an open source ones for all their work (nothing wrong with that - often done) rather than wasting any time trying to perfectly optimize a custom ones for SWTOR. Still even "pseudo-randomness" is good enough for most applications (e.g. even gear drops, as painful as RNG can be). Technically though, what the randomizer is programmed to be is "unpredictable" rather than truly "random."

 

But if your goal was to take a dataset (a, b, c, d) and have each variable in the dataset return with a 25% probability, you wouldn't want to use code that's programmed to be "unrepredictable" and then apply that pseudo-random/unpredictable code to a dataset of 4 variables and assume the output will be 25% for each. It almost certainly will not be.

 

You'd instead want code that's programmed to be truly random (which is hard to get and which certainly isn't in SWTOR for reasons beyond here) and apply the truly random code to the 4 variable set and say "pick one" OR take the simpler and "cleaner" approach of just forcing the 4 variables to return 1/4 of the time each (this is a nuanced distinction btw). For a small dataset like this, IMO that's the preferred approach. (Again, this is conceptual, not the only way, etc...)

 

So what I was really saying was that if the approach was more the latter, then we'd expect the probability of hitting "c" three straight times to be 1.5%. However, that probability is for just one attempt of 3 consecutive tries. If we actually took this probability (and I realize now that I write this that I made a slight error before, but the concept is the same) and applied over 6 months of attempts (so 8 sets (24 weeks/3), rather than 24 like I had in original - just too lazy to go correct it:o), I think the probability of seeing this happen once in that run goes up using: (1 - X)^8, where X = (0.25)^3.

 

This means that after six months, the chance of seeing GSF hit 3x in a row (assuming 25% chance, properly programmed) goes up to ~15% (1 - result). (I think this is right, but I admit I'm not putting lots of effort into it)...

[spatnatz]

So what I was really saying was that if the approach was more the latter, then we'd expect the probability of hitting "c" three straight times to be 1.5%. However, that probability is for just one attempt of 3 consecutive tries. If we actually took this probability (and I realize now that I write this that I made a slight error before, but the concept is the same) and applied over 6 months of attempts (so 8 sets (24 weeks/3), rather than 24 like I had in original - just too lazy to go correct it:o), I think the probability of seeing this happen once in that run goes up using: (1 - X)^8, where X = (0.25)^3.

 

This means that after six months, the chance of seeing GSF hit 3x in a row (assuming 25% chance, properly programmed) goes up to ~15% (1 - result). (I think this is right, but I admit I'm not putting lots of effort into it)...

 

(1 - X)^8, where X = (0.25)^3.

 

 

You may have been thinking of the geometric distribution, which is close to what you have--you'd have to multiply what you had by X again. That is probably what would make this theory work. It has to do with counting the number of failures until the first success (or vice versa).

 

Furthermore, according to a certain statistical rule of thumb (won't bore you with details), where np>= 10, and n(1-p) >= 10, with such a low probability, you'd need a very large value of n before you start to see results that converge to the expected value. In particular, we'd be talking about 642 3-week periods, or roughly 5.3 years. (!!!!)

 

So this means that, in the short term, we might see some surprising and unlikely results. As I mentioned previously, as time goes on, we would expect to see these unlikely events occur less and less frequently in a "perfectly random world."

Edited May 22, 2019 by spatnatz