49 replies to this topic

[Hammertrial]

http://mwomercs.com/...ost__p__1680369 for raw data and results.


The most numerous chassis is the Cataphract, followed closely by Catapults/Stalkers and then the Atlas.

The most numerous class are heavies, followed by Assaults, and lights and mediums are basically equal and far behind the other two.

The least numerous chassis is the awesome, with a sad and pitiful 2

Trial Mechs were ~15% of all mechs (not precise, some of the mechs may have been custom of the trial variant)

Hero Mechs were 4% of all mechs

ECM mechs were 22/135, or 16.3% of all mechs.
12 of 51 mechs were not seen.

Discuss here or in the other thread.

[soarra]

nice work, i would like to see this done over a large period of time.
some days i see a ton of lights others none, so after about a week or 2 im curious how it would look.. ill try to keep track

[hammerreborn]

I plan on running more trials in the future (probably another 20 matches in each class to see if that causes any effect). But for now I'm all worn out doing all of this in notepad.

I really need excel...that would have helped so much in doing this.

[Gaeb]

Neat stuff. I echo Soarra's comments. Definitely agree notepad is not sustainable for any sizable data collection work

What I would give to have a crack at data mining whatever PGI is logging ... admittedly, mostly out of curiosity since I already do such things for a living...

[soarra]

ill start to keep a tally.
grabbed a tally app on my ipad to keep track

[hammerreborn]

Gaeb, on 30 December 2012 - 06:05 PM, said:

Neat stuff. I echo Soarra's comments. Definitely agree notepad is not sustainable for any sizable data collection work

What I would give to have a crack at data mining whatever PGI is logging ... admittedly, mostly out of curiosity since I already do such things for a living...

I've done it twice now

To do the counts I just posted the results and used CTRL+F on my browser to figure out the numbers =P

[Flyby215]

Cool stuff! I want to know more about this so that I can play the underdog mechs. I sort of assumed the less played mechs were Centurion and Awesome; though since I do not yet have a medium mech I bought a Centurion.

Hmmm... Playing devil's advocate here; wouldn't the Stalker stat be slightly skewed because it is the new mech and everyone wants to try it?

[stjobe]

Hammertrial, on 30 December 2012 - 05:49 PM, said:

The most numerous class are heavies, followed by Assaults, and lights and mediums are basically equal and far behind the other two.

Interesting. It's supposed to be Medium, Light, Heavy, Assault - if you want to go by BT lore. Clearly MWO has a bit to go yet in getting the weight classes right.

Thanks for the hard work, it's always interesting to see some statistics.

[hammerreborn]

Perhaps, anything is possible.

A good test would be to repeat when the spider comes out in the middle of next month.

The stalker is just a really solid mech, so I don't think the newness of it is really driving its numbers.

The awesomes stats are just abysmal.

My biggest surprises was the complete lack of Wangs (I didn't see any on the opposing team either, not just on mine), I figured there would be more commandos then there were, and the medium count (though that was boosted heavily by the trial centurion).

Also, this clearly showed its not lightwarrior online, and that class distribution is farly even, but leaning heavier than lighter.

Edited by hammerreborn, 30 December 2012 - 07:01 PM.

[Gaeb]

Flyby215, on 30 December 2012 - 06:52 PM, said:

Hmmm... Playing devil's advocate here; wouldn't the Stalker stat be slightly skewed because it is the new mech and everyone wants to try it?

Actually there's a couple of macro trends that would need to be kept in mind for any analysis;

1) Holiday season. Should be a shift in the makeup/volume of playerbase, and I'd hypothesize you see more new/casual players in recent data samplings with <250 games played, likely driving a higher % of trial mechs and a relatively lower % of FOTM mechs

2) Recent macro game changes driving FOTM trends. Not just the Stalker release, but the addition of ECM (rise of Raven/Cicada/Commando variants as folks work to elite previously 'poor' chassis), the tail of the Cataphract release, the PPC projectile tweaks, TAG changes, etc. (edit: As pointed out below, also the $8.9M cbills given out last patch)

3) Player base growth. This went to open beta in mid-Oct (I didn't learn about it till just after the founders program closed!). I'd hypothesize your population is ramping up gradually, with a higher relative % of newbs than the last 10 weeks.

4) Timezone. Is the european meta different than the east coast american meta and the west coast american meta? Is there a korean playerbase? Etc.

5) Playerbase goals. While the data you can get is interesting, I'd caution about reading into it too deeply in absence of any competitive ranking system - some people play the same mech for 1000 matches, some people build troll builds, and right now a win or loss is immaterial in terms of your ranking so there's less incentive to optimize.

Anyhoo. Its a point in time set of observations, neat to read regardless. We could figure out a sampling that you could actually derive some legit statistical data from, but you'd need to keep all of the above in mind before drawing any conclusions.

Edited by Gaeb, 30 December 2012 - 07:15 PM.

[hammerreborn]

I think the bigger trend is premade vs non premade time. As seen in the first experiment I ran, prior to 6 on a weekday premades were seen in roughly half of matches. After 6 it went up to about 3 out of every 4 matches.

If you look at the makeup of some of the teams seen today you could clearly see some clear evidence on premades on my team. The match with 3 D-DCs, the one with 4 ravens, and a few others seem very off from the normal makeups seen in the typical matches.

I'll be doing a total of probably 40-80 trials in this diversity experiment as I get the urge to PUG large numbers of matches. I may do another 20 in each mech chassis to see if theres any effect, or keep in the Jenner to be consistent.

[Kaspirikay]

I have a fract atm. I can see why they're popular.

[Flyby215]

hammerreborn, on 30 December 2012 - 06:59 PM, said:

Perhaps, anything is possible.

A good test would be to repeat when the spider comes out in the middle of next month.

The stalker is just a really solid mech, so I don't think the newness of it is really driving its numbers.

The awesomes stats are just abysmal.

My biggest surprises was the complete lack of Wangs (I didn't see any on the opposing team either, not just on mine), I figured there would be more commandos then there were, and the medium count (though that was boosted heavily by the trial centurion).

Also, this clearly showed its not lightwarrior online, and that class distribution is farly even, but leaning heavier than lighter.

Heavies can deal a beating and take a beating without feeling like a turret... good all-rounder. I am disappointed on the lack of mediums... hmm... I purchased a Centurion recently and I intend to prove it a capable and deadly mech.

Devil's advocate again (sorry); in our most recent patch update players who had over 26 matches were given a free 8 million. Players who once lacked the financial discipline to save for an assault mech suddenly and immediately have the resources to purchase an assault mech: almost exactly priced at 8mil, the brand new Stalker.

The Stalker also benefits from having some badass looking art probably driving up its popularity. Though I would expect to see a good number of Spiders out there upon their release as well, I don't think it will even come close to Stalker numbers (heck I'd never buy a Spider) since it's not a badass looking assault mech.

Don't get me wrong, I think the stats are fantastic! I'm merely hoping to rationalize the Stalker's popularity since in recent matches, personally, I am seeing less and less of them every day.

[soarra]

well i just started my tally. so ill try to keep track best i can. given a week ill post them and hopefully with all our data we can see some good stuff.
thanks for the work hammer ill see if i can add to yours.

did you just do your team or enemy too?

[hammerreborn]

Just team, and feel free to add yours to mine in the other thread once you've compiled.

[soarra]

hammerreborn, on 30 December 2012 - 07:21 PM, said:

Just team, and feel free to add yours to mine in the other thread once you've compiled.

will do..

[BCOVertigo]

In leiu of Excel, you may look into getting Open Office to assist with this. It's FREEEEEEE!
http://www.openoffice.org/

[soarra]

https://itunes.apple...d370856329?mt=8 FTW

[hammerreborn]

The thing I don't like about a pure tally is that I like to show the raw data as well as the tallied results.

[soarra]

i have one with all the variants.. took a bit to type out