Curse

Assyl Threng · Feb 16, 2011, 09:45 PM // 21:45

One day, I got fed up of just saying that I had lousy luck with Random Arenas. As you all know, it's pretty much Monk or bust (with too few of exceptions to even worth mentioning). So here... well, I'll just copy & paste the notepad I was using. This shows the data after 100 games in RA, split into four categories. Games where only my team had a monk, games where only their team had a monk, games where both of us had a monk, and games where none of us had a monk.

Ok, here's the copy paste:

- - - - - - - - - - - - - - - - - - -

02/10/2011
Starting the recording of monks on my team and monks on their team for RA, after two days of none on mine, and almost always one on theirs.

Recording pauses on win streaks to avoid uneven data.

A slash indicates 1 game.
A pound symbol at the end of 4 slashes serves as a 5th slash, for easier viewing in multiples of 5.

MONKS ON [ONLY MY TEAM]
-----------------------

////# ////# //

MONKS ON [ONLY THEIR TEAM]
--------------------------

////# ////# ////# ////# ////#

////# ////# /

MONKS ON [BOTH TEAMS]
---------------------

////# ////# ////# ////#

MONKS ON [NEITHER TEAM]
-----------------------

////# ////# ////# ////# ////#

////# //

[TOTAL GAMES]
-------------

////# ////# ////# ////# ////#

////# ////# ////# ////# ////#

////# ////# ////# ////# ////#

////# ////# ////# ////# ////#

- - - - - - - - - - - - - - - - - - -

So let's take a look at this...

Monks on only my team: 12
Monks on only their team: 36
Monks on both teams: 20
Monks on neither team: 32

The absolute most likely thing to happen is for a Monk to only appear on the enemy team, and is 3 times more likely than a Monk only appearing on my team.

Including when both teams have a Monk, I only have one 32% of the time. And of that 32, only 12 of them can be taken advantage of as a Monk vs No Monk Game.

The opponents, however, have a monk 56% of the time. And of those 56, they have the majority of them (36) to take full advantage of as a Monk vs No Monk Game.

And, of course, 32% of the time nobody has a Monk anyway.

These odds are so heavily stacked against me, it makes me wonder if Random Arena really is random at all...

Maybe sometime, I'll make a Monk and start a new documentation about how many games I get 2+ Monks on the same team instead of actual damage dealers. Because, that's the reason I don't play a Monk in RA anyway. Perhaps then I can start proof of that, too.

Del · Feb 16, 2011, 09:52 PM // 21:52

Cool thread.

*lemming* · Feb 16, 2011, 09:57 PM // 21:57

You should include time of day in your data collection, too.

Assyl Threng · Feb 16, 2011, 09:59 PM // 21:59

Quote:

Originally Posted by lemming

You should include time of day in your data collection, too.

Good idea.

Too late for this bit of data, but I'll note that next time.

As for this bunch, I did it during all different times. So, it wasn't time-biased, but, it would be good to know when the Monks come online

I Jonas I · Feb 16, 2011, 10:36 PM // 22:36

http://en.wikipedia.org/wiki/Sample_error
http://en.wikipedia.org/wiki/Confidence_interval
http://en.wikipedia.org/wiki/Confirmation_bias
etc

LifeInfusion · Feb 16, 2011, 10:36 PM // 22:36

This doesn't count the amount of monks that don't know how to play GW right? Because people sit in their 40/40 heals and don't use energy swaps. In those kind of cases, sometimes it is better to have damage.

Cool Name · Feb 16, 2011, 10:51 PM // 22:51

"Recording pauses on win streaks to avoid uneven data"

This is one of the few major flaws in your "statistics". You are stopping when you have win streaks, which most of the time will be when you have a monk. You are not stopping (because it's impossible to know) when the other team has a win streak. By not counting occasions when you have a monk and are on a streak, but still counting the occasions when the other team has a monk and is on a streak you are completely destroying any chance of accuracy.

Also, ignoring this, if you were to take a snapshot of the teams participating in RA at any time then yes, I would think it very likely that the distribution of monks would be unbalanced (in favour of having more monks proportionally than having other professions (except perhaps warrior/assassin?)). This is obvious, intuitively. Teams that have a monk in them do better, of course, so those teams stay "alive". It is analogous to evolution.

So, I don't really know what you were trying to prove with this thread. Even if you had gathered data correctly, you would probably end up with a similar result, thus proving that teams with monks are more likely to survive... isn't everyone agreed on that already?

EDIT: Also, as for your next idea. It might be interesting, but also remember to note down how monks the other team has. If you have 3 monks and the other team has none, then that would obviously be unbalanced because you would think the system would be able to arrange the distribution of professions to be fairly equal. If, however, your team has 2 monks and the other team has 1, then that is only fair. It could have happened that at that time there were 8 players, 3 of whom were monks. Then it couldn't have been fairer.

Furthermore, in the situation where you have more than 1 monk and the other team has none then this could have been caused by the other team being on a streak (even though it's unlikely).

If you were to do this second suggestion, then it would only be interesting when compared with data of the number of monks you got when you were another profession. Even then, however, it would be another concrete. The fact that you are yourself a monk obviously makes it more likely that your team will get 2 monks.

urania · Feb 16, 2011, 10:56 PM // 22:56

his point is he is just an extremely unlucky fellow.

we can relate, mate!

and he is correct about random arenas not really being fully random (something that ra and ta had in common, actually). however, that doesnt hold true just for the cliche tendency that monks flock to your teams when you're monking yourself.

go cubs · Feb 16, 2011, 11:15 PM // 23:15

Quote:

Originally Posted by Del

Cool thread.

Cool Comment....

Also yeah, time of date would be cool too, also itd be cool if you had a section for more than one monk on your team/other team

Cool Name · Feb 16, 2011, 11:23 PM // 23:23

Quote:

Originally Posted by urania

...

and he is correct about random arenas not really being fully random (something that ra and ta had in common, actually). however, that doesnt hold true just for the cliche tendency that monks flock to your teams when you're monking yourself.

What do you mean by not fully random? Either they leave the teams as they are when people click enter (no ordering => still random) or they try order them to make them more balanced. Which of course isn't random, but they are trying to be. Both scenarios I consider to be fairly random, at least random enough for me. Do you think something else is happening?

urania · Feb 16, 2011, 11:34 PM // 23:34

there is one thing about ra that is most definitely not random (and it came from a fairly reliable source), and that is map selection. i believe ra and ta shared the same map "randomisation" code (unless they actually split the coding for them, though i find it unlikely), therefore its no longer fully random ever since the hb map test in ta. after they "reverted" hb map implementation, hb maps were replaced in map selection with the biggest tyria map (the beach map), which explains why one'd play on that map more often than in comparison to the rest (and would very often face shoves on it, amongst other things).
anyhow, there's little to complain about as far as map selection is concerned, since there is no shove in ra.

the other part that isnt fully random is the manner of entering, something that players (ab)use when trying to synch in.

SongOf · Feb 17, 2011, 12:03 AM // 00:03

Quote:

Originally Posted by Cool Name

So, I don't really know what you were trying to prove with this thread.

Exactly, you clearly don't and wasted your time typing all that, despite him saying it in the very first sentence.

He wanted some kind of data to prove he's got lousy luck regarding monks in RA, and while his data collection isn't perfect, it's better and more entertaining than him just saying so.

Cool Name · Feb 17, 2011, 12:18 AM // 00:18

Quote:

Originally Posted by SongOf

Exactly, you clearly don't and wasted your time typing all that, despite him saying it in the very first sentence.

He wanted some kind of data to prove he's got lousy luck regarding monks in RA, and while his data collection isn't perfect, it's better and more entertaining than him just saying so.

I think you also missed my point. My point was it isn't luck, this is how it should be. No matter how many times, how accurately, or by whom, this test would yield the same results. The fact is there are more teams with monks in RA because they last longer. I'll expand upon my evolution analogy:

Groups being randomly created is the RA form of reproduction, with the team being the child. If one of my parents was english and one was scottish, I would be half scottish half english. Similarly, if two of the four members of a team are rangers and the other two are both monks, the team will be half ranger half monk.

Now, in this analogy the equivalent of a generation is one sequence of matches. So if a team (child) can survive that generation, and therefore stay alive then there will be "more" of that child, effectively, since it will be in two generations, whilst the losing team will have only been in one.

Now we can see monks as being a strong gene. A natural equivalent might be... having better eyesight. A team with one monk in it is more likely to survive a match, I think we can all agree on that. If you understand evolution I don't think I should need to explain further. A monk is a gene which makes your team more likely to survive. Since these teams will stay together longer it will seem like there are more of them. So although it may seem like you are in less teams with one monk than everyone else, it feels the same for them. Simply put you are seeing the same teams over and over (even though it might not be the same specific team, it is the same effect).

Summary: It is not bad luck, teams with one monk just survive longer.

Sankt Hallvard · Feb 17, 2011, 12:19 AM // 00:19

Quote:

Originally Posted by Cool Name

This is one of the few major flaws in your "statistics". You are stopping when you have win streaks, which most of the time will be when you have a monk. You are not stopping (because it's impossible to know) when the other team has a win streak. By not counting occasions when you have a monk and are on a streak, but still counting the occasions when the other team has a monk and is on a streak you are completely destroying any chance of accuracy.

QFT.

Assuming there are at any given time more teams on streaks than there are forming teams entering as their first game will obviously change the chances of running into a better team, in the sense containing a monk.

Most, or at least a large fraction, of the players will leave a won(or lost obviously) game when they didn't get paired with a monk, or otherwise a less than decent team. Wammo on your team? Leave. No monk? Leave. This increases the odds of facing a team with a monk and also the chances of facing a good team all in all.

It's random. You're not hexed with bad luck. Not counting the syncing issue obviously, which means you can increase your chances of getting a monk to nearly 100% of the time. Factoring in that a significant amount of other players sync it's even less surprising that you face teams with monks so often.

Infectious · Feb 17, 2011, 12:32 AM // 00:32

Yet nothing about the fact you go XXX class for a few games, get sick of having no monk so you get on a monk, only to have another 1 or 2 monks join your team continously till you go back off monk.
Repeat
Rage quit

Assyl Threng · Feb 17, 2011, 02:01 AM // 02:01

While some of you are calling out the inaccuracy of the data due to me pausing during win streaks, I did mention me doing it for the awareness in the first place. I'd say it's safe to just say, for every point of data that would go to one side during a win streak, there's points going to all the other sides during streaks as well. So I felt no need to continue during streaks, especially when keeping this ENTIRE thing in the mind of, "likelihood of Monks upon hitting Enter Battle." That's actually the main reason I stopped during streaks.

Quote:

Originally Posted by SongOf

Exactly, you clearly don't and wasted your time typing all that, despite him saying it in the very first sentence.

He wanted some kind of data to prove he's got lousy luck regarding monks in RA, and while his data collection isn't perfect, it's better and more entertaining than him just saying so.

Thank you! This was my goal here

Cool Name · Feb 17, 2011, 02:31 AM // 02:31

You are calculating the likelihood of getting a monk on your team, yes. However, the composition of the opposing team is not representing what other people get when they.press enter. The difference is the opposing team may not have been newly formed, it is quite likely they are on a streak, so comparingthe composition of your newly formed team with a team that has been on a streak is skewing your results. Are you denying what I am saying or do you think it is not related?

Assyl Threng · Feb 17, 2011, 03:34 AM // 03:34

Quote:

Originally Posted by Cool Name

You are calculating the likelihood of getting a monk on your team, yes. However, the composition of the opposing team is not representing what other people get when they.press enter. The difference is the opposing team may not have been newly formed, it is quite likely they are on a streak, so comparingthe composition of your newly formed team with a team that has been on a streak is skewing your results. Are you denying what I am saying or do you think it is not related?

It's not that simple.

What you're saying is that this is a comparison of what I get when I Enter Battle compared to what other people get when they Enter Battle.

It's not. It's just me. So, whether or not the enemy Monk team is on a 24 win streak or not is completely irrelevant to me. Because I still did not get a Monk, and I still opposed a team that has one.

Chthon · Feb 17, 2011, 03:44 AM // 03:44

Quote:

Originally Posted by Assyl Threng

It's not that simple.

What you're saying is that this is a comparison of what I get when I Enter Battle compared to what other people get when they Enter Battle.

It's not. It's just me. So, whether or not the enemy Monk team is on a 24 win streak or not is completely irrelevant to me. Because I still did not get a Monk, and I still opposed a team that has one.

No, it is that simple. You just don't understand it.

Let's try one more time. Valid comparisons need to be between apples and apples, but you're comparing apples and oranges. Your team is drawn from the pool of teams entering from the lobby. Your opponent's team is drawn from the pool of teams entering from the lobby, plus the pool of teams that have already won a match. Different pools, different odds. Your team is random. The opposing team is statistically better than random.

Tharg · Feb 17, 2011, 03:44 AM // 03:44

Simple statistics... You mention that you don't play a monk. So you have 3 chances to have a monk on your team. The other team has 4 chances to have a monk on their team - therefore you have a disadvantage.

In other words - if you play a monk you have 100% chance to have at least one monk on your team...