Bioware, do you know the PVP map cycling is an issue right now?

[stringcat]

That's not unlikely with a random generator. If you read the description of the investigation, it sounds pretty solid.

 

With a large population, some will get streaks. They'll be the ones who post in this thread. So posts about it don't confirm it.

 

You don't get it. I am invariably getting streaks. Every single night. I'm sure I was at the beginning of a huttball streak (except that I logged). Last night was Novare. The night before that it was Civil War (I hate Civil War. Felt like slitting my wrists.).

 

If something should happen 4% of the time and is happening every single night to every hardcore PVPer I know, there's an issue. (Did I mock the friend getting Voidstar after Voidstar, while I was getting huttball after huttball? Yes. Yes I did.)

 

Also -- I apologize for not grabbing this information myself, but I haven't encountered this issue. Might be that I don't queue up for WZs non-stop -- so I don't really get to see it.

 

It's true that it only affects people who don't play many warzones, won't have the issue of getting the same warzone many warzones in a row.

Edited April 24, 2013 by stringcat

[RyanReagan]

Because a sample of 5 would be much better than the sample of 250,000 they menioned.

 

take my car to the shop. say "hey dude, my car won't start."

shop comes back and says "no way, we did some tests and only .01% of cars in the world don't start."

 

i know for a fact there is a problem. and maybe it is only happening at certain times or to certain people, which is why their numbers don't add up. but instead of looking to prove that the problem DOESN'T exist, they should be trying to prove the problem DOES exist. (because we do know it does). And then looking at the parameters as to WHY it exists.

 

One theory I've seen is that they pre-provision the maps onto servers. So when you queue back up it looks for an available instance, sees the one you just cleared and puts you right back in. And the timing works out that you get those you left most frequently....

Edited April 24, 2013 by RyanReagan

[ML_DoubleTap]

take my car to the shop. say "hey dude, my car won't start."

shop comes back and says "no way, we did some tests and only .01% of cars in the world don't start."

 

That's probably the worst attempt at an analogy you could have come up with.

[To-le]

Just to add my experience from last night.

 

I played about 5 matches last night. The first 3 matches were all void stars. After the second VS pop no big deal, it has happened before. As I was watching my team load in, I noticed some people I that was in my last VS match. It ended and we all exited the instance. Next pop was another VS! again I had most of the same peeps that was in the last one. We all mentioned in chat how this is very strange and wished we had a different map. After that match we did get a hypergate and I had new teammates. When that match ended I got another hypergate. Then I quit pvping for the night.

 

Even though its a small sample size, I believe there are issues.

[Twin]

Hey everyone,

Text.

 

Orly.

 

This is my WZ entries so far tonight....and it looks like any other night has done for some time.

Sure, it can be different WZ's -but it's allways 1-2 that completly dominates with popups over the rest...

 

Civil War: 5

Novar Coast: 3

Huttball: 1

Voidstar: 1

Hypergates: 0

[ArdoNorrin]

I think there's a huge misunderstanding about how map assignment works that is reflected in the name of this thread.

 

The maps don't CYCLE. That is, there is no pattern that the first match generated is, say, Voidstar, the next Civil War, then Huttball, then Novare, then Hypergates, then repeat. If that was the case, your chances of getting repeats would be significantly higher, especially if the number of simultaneous matches on your server was a multiple of 5 (essentially guaranteeing repeats). The generation of a match is random, and has ABSOLUTELY NO RELATION to the match you, or anyone else, did before.

 

The thing is, if you flip coins long enough, you will eventually get any number of consecutive "heads". It's a logical fallacy to believe that the next flip depends on any of the previous ones. The same thing holds for WZs: your chance of having the next pop be a Civil War has no relation to whether the previous was a Civil War. And, if you queue for enough warzones (though it may take billions), you'll eventually have a streak of 100 consecutive Civil Wars.

[TUXs]

The thing is, if you flip coins long enough, you will eventually get any number of consecutive "heads". It's a logical fallacy to believe that the next flip depends on any of the previous ones. The same thing holds for WZs: your chance of having the next pop be a Civil War has no relation to whether the previous was a Civil War. And, if you queue for enough warzones (though it may take billions), you'll eventually have a streak of 100 consecutive Civil Wars.

 

Oh whatever lol. Read the freaking posts from players - this is an OBVIOUS issue, don't pretend we're all ******* who don't understand chance.

[Elear]

Oh whatever lol. Read the freaking posts from players - this is an OBVIOUS issue, don't pretend we're all ******* who don't understand chance.

 

Let me guess, you also complain about not getting schematic from20 RE attempts and blame it on bug?

 

Random is random.

God of RNG may hate you, but there is nothing BW can do about that.

 

JonCrow did point out that there were 35 cases of streaks of 6, but that's just it, random bad streak. For all we know, only people complaining in this thread could belong to that unlucky group of 35, while all other players had repeats within 2-3.

 

If you don't unterstand that this fact that you and 10 other people in this thread have bad streaks is insignificant, then there is no hope for you.

[Twin]

I think there's a huge misunderstanding about how map assignment works that is reflected in the name of this thread.

 

The maps don't CYCLE. That is, there is no pattern that the first match generated is, say, Voidstar, the next Civil War, then Huttball, then Novare, then Hypergates, then repeat. If that was the case, your chances of getting repeats would be significantly higher, especially if the number of simultaneous matches on your server was a multiple of 5 (essentially guaranteeing repeats). The generation of a match is random, and has ABSOLUTELY NO RELATION to the match you, or anyone else, did before.

 

The thing is, if you flip coins long enough, you will eventually get any number of consecutive "heads". It's a logical fallacy to believe that the next flip depends on any of the previous ones. The same thing holds for WZs: your chance of having the next pop be a Civil War has no relation to whether the previous was a Civil War. And, if you queue for enough warzones (though it may take billions), you'll eventually have a streak of 100 consecutive Civil Wars.

 

Yeah, that might be so. Either way, each pvp session i see ALOT more of 1-2 warzones, compared to all the rest.

[Alizaro]

Let me guess, you also complain about not getting schematic from20 RE attempts and blame it on bug?

 

Random is random.

God of RNG may hate you, but there is nothing BW can do about that.

 

JonCrow did point out that there were 35 cases of streaks of 6, but that's just it, random bad streak. For all we know, only people complaining in this thread could belong to that unlucky group of 35, while all other players had repeats within 2-3.

 

If you don't unterstand that this fact that you and 10 other people in this thread have bad streaks is insignificant, then there is no hope for you.

 

Obviously you just think this is rng but something is wrong. If bioware had the ability to just look at my warzones and how many of those were duplicate and then figure the odds of me getting that many duplicates they'd find the chances of that randomly happening are almost impossible. Their metrics may show little variance but it's also possible this is a bug that affects mainly people that queue non stop which isn't the majority by any means.

[znihilist]

Obviously you just think this is rng but something is wrong. If bioware had the ability to just look at my warzones and how many of those were duplicate and then figure the odds of me getting that many duplicates they'd find the chances of that randomly happening are almost impossible. Their metrics may show little variance but it's also possible this is a bug that affects mainly people that queue non stop which isn't the majority by any means.

 

You keep forgetting that even if the probability is doesn't mean it will not happen. Seeing how large the number of WZs going on, law of large numbers will tell us that those improbable events will most probably happen.

It is one in a hundred million chance that you win the lottery (maybe even less probable ?), yet every month or so you hear about someone who won it. Do you get it now ?

Edited April 24, 2013 by znihilist

[ML_DoubleTap]

I have the day off tomorrow and will likely be playing most of that time. This thread has me interested enough to track the WZs I get. The small sample doesn't mean much, but I'm assuming I'll find it's not as bad as everyone is making it sound. My guess is that because this thread is here people are more aware when they get a couple in a row and exaggerate the issue.

[EricMusco]

So what you're saying is. You didn't bother to have someone queue up for an hour to tell you that yes, it is broken.

 

reminds me of a joke:

A mathematician, a physicist and a statistician went hunting for deer. When they chanced upon one buck lounging about, the mathematician fired first, missing the buck's nose by a few inches. The physicist then tried his hand, and missed the tail by a wee bit. The statistician started jumping up and down saying "We got him! We got him!"

 

We are definitely aware that this kind of "streakiness" is possible in the live game. Interestingly, my personal account is one of the accounts we used as a test bed to help gather data. Last weekend I experienced Ancient Hypergate 6 times in a row! In fact, Jon himself didn't dispute that players can and will see streaks of Warzones, it is totally possible, just improbable.

 

We are aware of these concerns and are actively tracking metrics to make sure this isn't a bigger issue. This is the kind of thing we will always keep an eye on and look to improve.

 

-eric

[znihilist]

We are definitely aware that this kind of "streakiness" is possible in the live game. Interestingly, my personal account is one of the accounts we used as a test bed to help gather data. Last weekend I experienced Ancient Hypergate 6 times in a row! In fact, Jon himself didn't dispute that players can and will see streaks of Warzones, it is totally possible, just improbable.

 

We are aware of these concerns and are actively tracking metrics to make sure this isn't a bigger issue. This is the kind of thing we will always keep an eye on and look to improve.

 

-eric

 

Why don't you switch to a bayesian implementation ? Shouldn't be very hard to do.

The more you get a WZ the lower the probability of getting it again.

Edited April 24, 2013 by znihilist

[jibboo]

All I know is that getting two of the same warzones in a row very rarely happened prior to 2.0 and now it happens on a regular basis.

 

Just a coincidence though, I'm sure...

[WorldSecurities]

Hey everyone,

 

We reviewed the data behind the Warzone queuing system in order to identify the recurrence of streaks of identical Warzones popping for players entering the system. If you assume a basic random Warzone assignment model you would have a 1 in 5 chance of repeating the same Warzone after queuing for any given Warzone. The probability of a streak of 3 identical Warzones using random assignment would be 0.2^2=0.04 or 4%. Again the probability of having a streak of 4 or more goes down precipitously.

 

We queried the data an examined the current streakiness of a random sample of players participating in Warzones currently and plotted the probability of certain streaks occurring against the random probabilistic model. The actual data is on par with basic random chance. Recent Warzone players have had a little over 15% chance of repeating the same Warzone compared to the assumed 20%, and subsequent streaks probabilities are also on par with random chance.

 

Does this mean that you won’t get a streak of the same Warzones? No! In fact, of the sample of 250,000 Warzones we examined, there were 35 streaks of 6, but this is truly the exception rather than the rule and it falls within the confines of the design model.

 

Also we examined the current trends of number of Warzone instances being generated by the system pre and post ROTHC launch and we have almost the identical number of Warzone instances being generated across Warzone types throughout the whole time frame. This helped confirm that, overall, there isn’t any one Warzone being “favored” over time.

 

However, we are still looking at ways to improve the design of the warzone matchmaking system in order to maximize the end user experience. Our designers are hard at work looking in details at ways to optimize the system in order to reduce further the likelihood of streaks occurring.

 

Thank you all for the comments on this topic. We appreciate your passion for the game and are glad to get constructive feedback. This allows us to investigate potential issues that we might not have caught otherwise. Keep the feedback coming.

 

Here are a two examples that fits the above explanation and player observations.

 

Extreme example ("there were 35 streaks of 6"):

 

Player 1: WZ 1, 1, 1, 1, 1, 2, 2, 2, 2, 2

Player 2: WZ 2, 2, 2, 2, 2, 3, 3, 3, 3, 3

Player 3: WZ 3, 3, 3, 3, 3, 4, 4, 4, 4, 4

Player 4: WZ 4, 4, 4, 4, 4, 5, 5, 5, 5, 5

Player 5: WZ 5, 5, 5, 5, 5, 1, 1, 1, 1, 1

 

Slightly more balanced example:

 

Player 1: WZ 1, 1, 2, 2, 1, 2, 2, 2, 1, 1

Player 2: WZ 2, 2, 3, 3, 2, 3, 3, 3, 2, 2

Player 3: WZ 3, 3, 4, 4, 3, 4, 4, 4, 3, 3

Player 4: WZ 4, 4, 5, 5, 4, 5, 5, 5, 4, 4

Player 5: WZ 5, 5, 1, 1, 5, 1, 1, 1, 5, 5

 

In both cases, for the total number of WZs played and the total population, there's a 1 in 5 chance of a specific WZ popping up. However in both cases the only random aspect is the 1 in 5 chance of being stuck in a pattern.

 

HOWEVER the SWTOR random number generator might not be broken: "This means that a good random number generator will also produce sequences that look nonrandom to the human eye (e.g., a series of ten rolls of six on our die) and which also fail any statistical tests that we might expose it to."

 

Meaning BioWare could have the very best random number generator, and it will look non-random.

 

BioWare should contract a real Mathematician to certify the random number generator for it's customers. Or perhaps replace the RND function, just in case.

 

/*

 

"So, if it is impossible to definitively prove randomness, what can we do instead? The pragmatic approach is to take many sequences of random numbers from a given generator and subject them to a battery of statistical tests. As the sequences pass more of the tests, the confidence in the randomness of the numbers increases and so does the confidence in the generator. However, because we expect some sequences to appear nonrandom (like the ten rolls of six on our die), we should expect some of the sequences to fail at least some of the tests. However, if many sequences fail the tests, we should be suspicious. This is also the way you would intuitively test a die to see if it is loaded: Roll it many times, and if you see too many sequences of the same value coming up, you should be suspicious."

[Alizaro]

You keep forgetting that even if the probability is doesn't mean it will not happen. Seeing how large the number of WZs going on, law of large numbers will tell us that those improbable events will most probably happen.

It is one in a hundred million chance that you win the lottery (maybe even less probable ?), yet every month or so you hear about someone who won it. Do you get it now ?

 

I get it. Lots of people hitting the lottery, yay us.

[znihilist]

I get it. Lots of people hitting the lottery, yay us.

 

Actually quite the opposite, seeing the WZ sample and the streaks lottery winners are falling within the expected probability. You are forgetting all those people who are not posting here because they are not seeing these streaks.

[noobzor]

Oh whatever lol. Read the freaking posts from players - this is an OBVIOUS issue, don't pretend we're all ******* who don't understand chance.

 

But that is exactly what is going on. A bunch of people have observed sequences that appear non-random, and they assume it is because the entire system is non-random. But, in any random system, especially one with so many samples, it is inevitable that many people will experience this phenomenon.

 

I'm not saying that there isn't a problem, I'm just saying that no player's experience, or even a handful of players' experiences, is sufficient evidence to prove that there is an issue. It just doesn't work that way.

[stringcat]

We are definitely aware that this kind of "streakiness" is possible in the live game. Interestingly, my personal account is one of the accounts we used as a test bed to help gather data. Last weekend I experienced Ancient Hypergate 6 times in a row! In fact, Jon himself didn't dispute that players can and will see streaks of Warzones, it is totally possible, just improbable.

 

We are aware of these concerns and are actively tracking metrics to make sure this isn't a bigger issue. This is the kind of thing we will always keep an eye on and look to improve.

 

-eric

 

This is an "improbability" that has occurred nightly on Begeren Colony since 2.0 -- since you "made adjustments to balance the Warzone types that players will see when queuing." It may not be 6 Ancient Hypergates in a row, no. But a player is very likely to see 3 Hypergates, then a Voidstar, then another Hypergate....Etc. Pre 2.0, I almost never saw the same warzone twice in a row.

 

There is something wrong with the way you are sampling or reading your metrics. It doesn't help to "track" them if you are tracking the wrong data. If you don't see the issue in your current data samples, you are tracking the wrong data. I'm not reassured.

 

I suggest that you look at the data for the smaller servers (like mine say!), collecting data on players who have played at least 8 consecutive warzones. This is the experience of every PVPer on Begeren Colony.

Edited April 24, 2013 by stringcat

[TUXs]

We are definitely aware that this kind of "streakiness" is possible in the live game. Interestingly, my personal account is one of the accounts we used as a test bed to help gather data. Last weekend I experienced Ancient Hypergate 6 times in a row! In fact, Jon himself didn't dispute that players can and will see streaks of Warzones, it is totally possible, just improbable.

 

We are aware of these concerns and are actively tracking metrics to make sure this isn't a bigger issue. This is the kind of thing we will always keep an eye on and look to improve.

 

-eric

 

Eric, it's a bigger issue. This isn't some chance happening. I've been doing WZ's nearly every night I've played since beta, the frequency of the repeats is far beyond "chance". The simple fact that it's an issue that so many players are acknowledging they themselves have experienced, should be the biggest indicator that something isn't working correctly.

 

Having "streaks" is fine - we've probably all experienced the duplicate Huttball matches or the return to a WZ after just one other map - but this isn't that. This is something very different. Last night in a 30-54 AH, all of us bugged out on the final load screen (the scoreboard) - it never popped up so we all exited the WZ using the mini map. This is the first time I've seen that as well and I'm guessing that something that is supposed to happen at the end of a WZ, isn't happening correctly.

 

This isn't about "chance" or RNG's - there's something very noticeably wrong with the WZ rotations since 2.0. Please encourage Jon to seek help from someone else on this. The faith Bioware puts in their "metrics" vs. player experiences is mind boggling to me. Metrics only tell PART of the story...your fix won't be found there.

[SCHodge]

We are aware of these concerns and are actively tracking metrics to make sure this isn't a bigger issue. This is the kind of thing we will always keep an eye on and look to improve.

 

-eric

 

A very simple solution, at least in theory, is to track your last 2 warzones. If they're the same, make it so internally you aren't queuing for that warzone a third time.

 

More generally, set some threshold - given that 20% is average, say 40% - and if a given warzone in your last n matches exceeds that, take it out of your queue. (And perhaps, ahem, bolster the chances of warzones not showing up).

 

Or allow choices of what warzones to queue for? I'll actually spend cartel coins for an "All Huttbal, All the Time" button.

 

Clarin @ The Harbinger

[Mhorlok]

Extreme example ("there were 35 streaks of 6"):

 

Player 1: WZ 1, 1, 1, 1, 1, 2, 2, 2, 2, 2

Player 2: WZ 2, 2, 2, 2, 2, 3, 3, 3, 3, 3

Player 3: WZ 3, 3, 3, 3, 3, 4, 4, 4, 4, 4

Player 4: WZ 4, 4, 4, 4, 4, 5, 5, 5, 5, 5

Player 5: WZ 5, 5, 5, 5, 5, 1, 1, 1, 1, 1

 

Slightly more balanced example:

 

Player 1: WZ 1, 1, 2, 2, 1, 2, 2, 2, 1, 1

Player 2: WZ 2, 2, 3, 3, 2, 3, 3, 3, 2, 2

Player 3: WZ 3, 3, 4, 4, 3, 4, 4, 4, 3, 3

Player 4: WZ 4, 4, 5, 5, 4, 5, 5, 5, 4, 4

Player 5: WZ 5, 5, 1, 1, 5, 1, 1, 1, 5, 5

 

From the poster above sums up exactly what I think is happening

 

I do Business Intelligence for a livign - thats turning data into information - to me this looks like bioware has a bad BI team. They are seeing the data but not the information.

If 4 in a row is 4% and 6 in a row is extremely unlikely how were you able to produce those results in a simple test - because you see the raw data but you fail to understand everything it is telling you.

[toomsuba]

/signed. Huttball 8 times in a row isn't exactly my definition of fun.

[znihilist]

If 4 in a row is 4% and 6 in a row is extremely unlikely how were you able to produce those results in a simple test - because you see the raw data but you fail to understand everything it is telling you.

 

What ??