237 replies to this topic

[Bishop Steiner]

Jman5, on 06 September 2014 - 10:55 AM, said:

So the Hunchback is worse than every single mech in the game except the 35 ton Raven, the 25 ton Commando, and the 20 ton Locust.

....? Rly?

[Mcgral18]

Jman5, on 06 September 2014 - 10:55 AM, said:

So the Hunchback is worse than every single mech in the game except the 35 ton Raven, the 25 ton Commando, and the 20 ton Locust.

Did you submit enough data?

It's not the most popular chassis. Also needed multiple pilots.

[Fire and Salt]

Astrocanis, on 06 September 2014 - 10:54 AM, said:

It is as I said. 10 v 12 is a convenient argument aimed at ensuring superiority of tech - of course with "balancing" to make sure the IS at 12 is not "superior". Which, again, means that 1v1 Clan tech will ensure victory.

1v1 argument is irrelevant if all games are 10 v 12.

Try again.



And BTW, balancing is to ensure that clan tech is not superior at 10v12 just as much as the opposite.

Astrocanis, on 06 September 2014 - 10:54 AM, said:

Care to argue against "I like my OP Clan mech..." again?

My favorite mech (in this game) is the Locust. Mainly because it is extremely fun, and looks awesome. Ask members of my clan. They tease me about it constantly.

Spoiler

Astrocanis, on 06 September 2014 - 10:54 AM, said:

One more thing: you aren't creche-born, you aren't, yourself, clan. You are not a 6'6" indestructible god. You are a guy in front of a computer <more or less> like the rest of us. I'm guessing your favorite superhero is Superman?

The clans weren't indestructible gods.

In fact, the best clan pilots weren't really better than the best IS pilots - its just that the average clan pilots were better than the average IS pilots.

Not everyone likes the clans because of the superior tech. I just think that their ritualized combat is cool. Specifically, I like the trial of entrance and stuff. I also thinks its stupid to extend honorable combat to enemies who don't return the favor. Hence my fondness for CSA rather than some other clan.

TBH, the difference between myself and the average person may be larger than the difference between the average clan pilot and the average IS pilot. IQ wise it is more than 2 standard deviations. This is, of course, completely irrelevant.

But nice attempted strawman fallicy.

And No, I don't have a favorite superhero.

Edited by Fire and Salt, 06 September 2014 - 11:41 AM.

[A banana in the tailpipe]

Mcgral18, on 06 September 2014 - 11:13 AM, said:

Did you submit enough data?

It's not the most popular chassis. Also needed multiple pilots.

Funny story... Last night I had two pug drops in a row with a PGI dev while working on my hunchbacks for stock monday. The second match I decided to mention in chat the hunch could use a little more love in the armor/structure department because overall they are pretty abysmal to play. (usually it's a chore just to break 300 and get 1 kill)

To set a great example for him I got 3 kills and broke 650 damage, served up with a steaming plate of crow on the side for opening my big mouth. Talk about irony! I was like "What did everyone stand still for me? Now the thing will never get buffed. THANKS GUYS I HATE YOU ALL!"

Edited by lockwoodx, 06 September 2014 - 11:29 AM.

[Mirkk Defwode]

Well written and displayed Kiiyor. I'll look to send you some additional games for a larger sample set.

Though I haven't been playing as much lately. I do take a different approach to the use of specific mechs. As I haven't mastered anything but am in the process of getting all my chassis up to Elite status before going back through and getting them to Master. I do this a chassis at a time from Light to Assault as a specific order. I'm currently working on the Summoners - now in a large enough sample set play distribution and behavior based on this shouldn't matter assuming there is enough randomized behavior.

Though if you make the assumption others do this same behavior to a large enough sample size it'd make different distribution statistics entirely.

Just a little food for thought.

Edited by Mirkk Defwode, 06 September 2014 - 11:33 AM.

[Mystere]

Kiiyor, on 06 September 2014 - 05:35 AM, said:

<Great stuff!!!>

That is great work.

But ( ), you failed to directly answer the most important question of all: Are the Clans OP?

Or did you just chicken out? ... I joke, I joke, I joke ... but only somewhat ( ).


Also, what sorceric incantation is this?



( )

Edited by Mystere, 06 September 2014 - 11:41 AM.

[Mazerius]

All the DATA...../mindexplode Seriously though amazing post, and thank you for putting that much effort into the analysis. I'll be taking end of round screenies after this hah.

[stjobe]

I... I... Actually read the whole OP

Now I have a headache and a massive amount of respect for Kiiyor

This thread needs a sticky (although I doubt it'll get one since it shows in great detail (and with SCIENCE) that MWO is far from balanced)

[Jman5]

Mcgral18, on 06 September 2014 - 11:13 AM, said:

Did you submit enough data?

It's not the most popular chassis. Also needed multiple pilots.

There were 413 hunchback data points which comes out to almost exactly middle of the pack in terms of frequency. So he had enough to get a good average.

[Fire and Salt]

Maybe it is because of all this advice I see directed towards new players: "Buy a hunchback"

Clearly these statistics do not include enough Jman5.

[Mirkk Defwode]

stjobe, on 06 September 2014 - 12:29 PM, said:

I... I... Actually read the whole OP

Now I have a headache and a massive amount of respect for Kiiyor

This thread needs a sticky (although I doubt it'll get one since it shows in great detail (and with SCIENCE) that MWO is far from balanced)

Agreed, it does deserve a sticky, and this guy deserves some free stuff in the title for demonstrating what can be done with their metrics - imagine using his tables and math that can be tapped straight into the statistics database for the title. It'd show some interesting outcomes. And limiting the data sets to each build as comparison for the weapons/chassis metrics would prove interesting to see what people favor the most as well as how efficient those combinations are.

[Homeless Bill]

I thought I wrote long posts... Well done, sir. And that formatting...

[Pika]

Huh. But I always use the (C) SHD and it's by FAR my best perfoming 'Mech. In fact it's probably the only Mech I own that has a positive KDR!

Am I a statistical anomaly here or what?

Also this is a lot to read I shall read more and maybe comment again later >.>

Edited by Pika, 06 September 2014 - 01:51 PM.

[Fire and Salt]

Pika, on 06 September 2014 - 01:41 PM, said:

Huh. But I always use the (C) SHD and it's by FAR my best perfoming 'Mech. In fact it's probably the only Mech I own that has a positive KDR!

Am I a statistical anomaly here or what?

Also this is a lot to read I shall read more and maybe comment again later >.>

Well, most people buy the non (C) variants because they don't need to spend real money.

So (C) mechs are usually beginners in trials.

No, you are not the only one. I personally do not own any (C) mechs.

Edited by Fire and Salt, 06 September 2014 - 02:22 PM.

[Dark Jackal]

Pretty impressive amount of work put into the post which must have took a LONG time to put together. While we can only discuss averages, I've hit in some rough patches with 'Mechs and particularly ran some of my IS 'Mechs barely upgraded to be as effective as I know I could be simply because I don't have C-bills to spare and working towards my goal of XP unlocking my 'Mechs. I can simply look over my Commando stats from months back to see the results prior to the Clan addition (pretty dismal stuff). And worse about dismal 'Mechs is that I spend twice as much time grinding away than I do with a nice 'Mech I do well with and we're talking about hours of play time 3 v 6 hours just to unlock all of elite tier.

If we had access to all essential particulars we could home down certain things like whether or not some folks are actually making best use of their 'Mech or not and probably make better assumptions on what really skews the stats one way or the other. I don't think the IS 'Mechs are particularly useful right out of the stock variation in most cases being crippled by default compared to some of the nicer Clan models that can be played very well with little to no modification and come essentially close to maxed out.

[Tkhaw]

Mcgral18, on 06 September 2014 - 09:21 AM, said:

It'll have 26 tons of pod space with near max armour, so more than the Summoner, but less than the Timby. 89 speed, questionable hitboxes (we'll have to see) and no JJs.

It won't be bad, most likely, but shouldn't be anything amazing either. 2 hardwired heatsinks are in the engine, at least, and there are still 9 slots in each arm, and 8 in each side torso. Legs are open to hide ammo, as well.




As always, nice work Kiiyor. I'm tempted to send in some of my pictures as well, although there is certainly some bias in there.
I should probably just start snapping them from this point and and just send those ones in, to avoid bias.

That's what I have started as well

[Kiiyor]

Bryan Ekman, on 06 September 2014 - 08:57 AM, said:

Wow! Love the effort and attention to detail. It would be interesting to compare your findings against the same telemetry we've collected!

I was going to PM you and see how they compare! Even though my sample is quite a bit larger than my first effort, there are still a lot of outliers and vagaries. I was worried that this sort of thing may have an inflammatory effect on some already heated debate on the forums.

If you do check it with your telemetry, i'd be interested in a few general nods towards data that is similar or wildly inaccurate

[Kiiyor]

Jman5, on 06 September 2014 - 10:55 AM, said:

So the Hunchback is worse than every single mech in the game except the 35 ton Raven, the 25 ton Commando, and the 20 ton Locust.

I know man, I know. I feel your pain - the 4SP was my favorite mech for Aeons. I loved it, and that made me keep playing it even as the whole pinpoint huge alpha doom strike started to emerge and allowed my beloved brawler to be effectively one shot. I played it until it's KDR had dropped so low that I had to acknowledge that dropping in it was just giving the enemy free kills.

E N E R G Y, on 06 September 2014 - 11:04 AM, said:

Dude, that's freaking amazing.

PGI needs to hire you lol; how long did this guide take to make?

Too long, lol. After the last batch of data, I approached this one differently. I used lots of copy and paste templates, and I even planned a data model to make it easy to pull out numbers. Then I tried to add the tournament score to almost 40,000 mechs, at once, and the sheet exploded spectacularly, corrupted everything, and I had to pretty much start from scratch.

The worst part was actually getting the post up on the forums. I had broken the whole thing into about 100 images, and... lots of text. It took almost a whole freaking day before I could hit the submit post button!

[Kiiyor]

stjobe, on 06 September 2014 - 12:29 PM, said:

I... I... Actually read the whole OP

Now I have a headache and a massive amount of respect for Kiiyor

This thread needs a sticky (although I doubt it'll get one since it shows in great detail (and with SCIENCE) that MWO is far from balanced)

This is true about balance.

The whole point of me putting this together was to try and... I guess 'baseline' some of the debate. Probably the wrong word. So many people were becoming so vitriolic about the whole thing! If only the anger could be reigned in a little, people from opposing viewpoints might be able to meet in the middle a little more, and start having some actual discussion instead of flaming and put-downs. Where is the love and all that crap.

I wasn't aiming to prove that OMG THE CLANS ARE OP NERF THEM AAAAA, I was just trying to add some context to everything, and show the areas that will affect people's ideas of balance.

For example: the Ember and Misery. Both are great mechs, but why are they so much better than their other variants, and other heroes? Their hardpoints. The ballistics on the Misery obviously has a huge effect on it's performance. It gave me a far better appreciation of what the devs must have to go through when planning mechs - that one ballistic hardpoint added to a mech can wildly improve it's performance. But are they just better mechs? Could it be that the people buying them are already decent players who can recognize the potential of the mech, and less 'bad' players are piloting them because they simply aren't buying the mech? I don't know... but the numbers are at least an indication of performance.

If people are going to argue, they can at least be (moderately) informed!

[Kageru Ikazuchi]

Kiiyor, once again, you rock!