94 replies to this topic

[TruePoindexter]

Thanks for the diversity breakdown!

The Awesome does need some love. I am surprised with the Ciacada 3C popularity...

[Fastidious]

Nice stats, interesting to see hard data for once. Thanks all.

[hammerreborn]

TruePoindexter, on 02 January 2013 - 12:42 PM, said:

Thanks for the diversity breakdown!

The Awesome does need some love. I am surprised with the Ciacada 3C popularity...

The cicada non 3M varients are likely people trying to unlock the elite/master tiers of the class before heading back to the 3M. Similiar to the Raven 4X. The Raven 2X has the effect of people both trying to unlock the class, as well as using the current trial mech, which is why theres likely just a huge gap between the two.

Most other mechs are probably people playing a preferred/trial variant. The stalker main 3 (the other two basically being worthless) are the only ones seen while the 2 inferior ones aren't.

And to the guy who commented about ECM. The thing about ECM is that it's a two-fold issue. In most matches (look at my raw data) where there is 0, 1, or 2 ECM mechs, the result wasn't really determined. When you start getting into the higher numbers does things start looking a little fishier. This is due to premades more than likely. A 4 assault match means that we were paired up against probably another 4 D-DCs or similiar, which is almost without a doubt a premade.

As in my first experiment, the team with the premade wins roughly 60-70% of the time, so people have basically attributed ECM with winning, rather than attributing large amounts of ECM with well coordinated premades which win because in my experiments

DC > coordination > ECM > pug, in terms of effectively determining the winner.

Edited by hammerreborn, 02 January 2013 - 02:20 PM.

[TruePoindexter]

hammerreborn, on 02 January 2013 - 02:18 PM, said:

The cicada non 3M varients are likely people trying to unlock the elite/master tiers of the class before heading back to the 3M. Similiar to the Raven 4X. The Raven 2X has the effect of people both trying to unlock the class, as well as using the current trial mech, which is why theres likely just a huge gap between the two.

Most other mechs are probably people playing a preferred/trial variant. The stalker main 3 (the other two basically being worthless) are the only ones seen while the 2 inferior ones aren't.

And to the guy who commented about ECM. The thing about ECM is that it's a two-fold issue. In most matches (look at my raw data) where there is 0, 1, or 2 ECM mechs, the result wasn't really determined. When you start getting into the higher numbers does things start looking a little fishier. This is due to premades more than likely. A 4 assault match means that we were paired up against probably another 4 D-DCs or similiar, which is almost without a doubt a premade.

As in my first experiment, the team with the premade wins roughly 60-70% of the time, so people have basically attributed ECM with winning, rather than attributing large amounts of ECM with well coordinated premades which win because in my experiments

DC > coordination > ECM > pug, in terms of effectively determining the winner.

Those observations match my experiences in particular the issue with disconnects. The game desperately needs a reconnect/resume option to help mitigate it.

[hammerreborn]

TruePoindexter, on 02 January 2013 - 02:24 PM, said:

Those observations match my experiences in particular the issue with disconnects. The game desperately needs a reconnect/resume option to help mitigate it.

Not sure how much that would help to be honest. Most disconnects are due to client crashes, and the time it would take to reconnect with the current match lengths would probably be 7-0 by the time you log in to be inevitably killed.

On the other hand, theres a solution to hte 4fps bug.

[TruePoindexter]

hammerreborn, on 02 January 2013 - 05:42 PM, said:

Not sure how much that would help to be honest. Most disconnects are due to client crashes, and the time it would take to reconnect with the current match lengths would probably be 7-0 by the time you log in to be inevitably killed.

On the other hand, theres a solution to hte 4fps bug.

I manage to restart my client in less than a minute but it still is a time loss. Now that the dev's are back in the office I can't wait to see what will be coming in the next patch.

[hammerreborn]

Well we know the following is coming down the pipe:

-elo matchmaking
-spider (15th)
-JJ adjustments
-Weapon overhauls (mainly, Ppc Emp effect, and heat adjustments for large pulse, ppcs, and er ppcs. This doesn't have current dates though).

[Taizan]

I'm a sucker for statistics like that and not surprised a bit over the Awesome unpopularity.

One thing to keep in mind that these statistics will always be skewed a bit by the trial mechs offered. Currently seeing way more CN9-Ds and DRG-5Ns due to the trial mechs.

Anyway good effort, maybe doing it once every round when new trial mechs are made available will give you more useful data.

[Thirdstar]

Checking your thread after NY break. Great all round info, glad to see some stats out there.

If I would disagree about one issue is the impact of ECM. My personal experience is that the complete loss of radar cripples PUGs far more than it does premades. Specially so in situations where friendly mechs aren't in LOS.

I have won just about every match as long as I knew where I was and where my team was. When you take that away I will honestly admit to being overwhelmed. I can't type out a strategy worth a damn if I don't know where everyone is. The ability to glance at the map and know where the rest of the team is and what they're fighting is essential to a PUG's ability to fight given the lack of in-built comms.

[hammerreborn]

Taizan, on 03 January 2013 - 07:32 AM, said:

I'm a sucker for statistics like that and not surprised a bit over the Awesome unpopularity.

One thing to keep in mind that these statistics will always be skewed a bit by the trial mechs offered. Currently seeing way more CN9-Ds and DRG-5Ns due to the trial mechs.

Anyway good effort, maybe doing it once every round when new trial mechs are made available will give you more useful data.

Indeed, I would expect the 2X numbers to fall to about the 4X were it not for trials (basically, people playing the other variants to master the 3L), I would expect the 9D to stay roughly the same percentage of the chassis, but be lower overall. The 9D is just an overall good mech, because of it's speed. The dragon is most certainly getting a boost from the current trial, and the Stalker is probably just as many people in trial mechs as there are people in legit versions of it. The Stalker is popular, it's new, it's shiny, and can blow your *** to kingdom come with a single alpha.


I'll probably run a bit more after the spider is released, which will also probably come out with a new set of trials (it will but I hope it won't include the spider in the trial rotation so I can actually get solid numbers of it), to see the impact of new vs old. Though with the current JJ implementation and the spiders stock loadouts (if PGI doesn't change the hardpoints) as well as the current speed cap extremely hampering it, I'm not sure if it'll get much looks.

Quote

Checking your thread after NY break. Great all round info, glad to see some stats out there.

If I would disagree about one issue is the impact of ECM. My personal experience is that the complete loss of radar cripples PUGs far more than it does premades. Specially so in situations where friendly mechs aren't in LOS.

I have won just about every match as long as I knew where I was and where my team was. When you take that away I will honestly admit to being overwhelmed. I can't type out a strategy worth a damn if I don't know where everyone is. The ability to glance at the map and know where the rest of the team is and what they're fighting is essential to a PUG's ability to fight given the lack of in-built comms.

Well....ya. I wouldn't go so far as to say ECM cripples PUGs, and in a pure PUG vs PUG match barely makes a difference, but a team using ECM effectively (as in your example) is most likely a premade or very coordinated, and as I detailed

DC > coordination > ECM > PUG in terms of guessing the winner.

ECM only works when coordination is involved, that lone raven ramboing thinking he's immortal cause Ravens are OP right now isn't helping his team and will get squished by any decent pilot that can aim, and the same goes to any D-DC that thinks he's god because he's in an atlas and can't be LRMed (cause most PUGs don't use TAG) and just waltz into the other team.

I put ECM > PUG because ECM gives a slight edge due to light v light combat to the team with the ECM, and an Atlas D-DC is a harder target to bring down (as using LRMs and streaks become difficult to bear on him) than an Atlas K. That may be enough to swing a match as an unimpeded Raven can wreak havok in the back lines.

Edited by hammerreborn, 03 January 2013 - 08:33 AM.

[TruePoindexter]

hammerreborn, on 03 January 2013 - 08:32 AM, said:

I'll probably run a bit more after the spider is released, which will also probably come out with a new set of trials (it will but I hope it won't include the spider in the trial rotation so I can actually get solid numbers of it), to see the impact of new vs old. Though with the current JJ implementation and the spiders stock loadouts (if PGI doesn't change the hardpoints) as well as the current speed cap extremely hampering it, I'm not sure if it'll get much looks.

Looking forward to it. I'm not optimistic about the Spider though - I suspect it will be underwhelming.

[Thirdstar]

TruePoindexter, on 03 January 2013 - 11:07 AM, said:

Looking forward to it. I'm not optimistic about the Spider though - I suspect it will be underwhelming.

I'm not entirely sure what niche the Spider fills. I suspect that it actually doesn't.

[TruePoindexter]

Thirdstar, on 03 January 2013 - 11:14 AM, said:

I'm not entirely sure what niche the Spider fills. I suspect that it actually doesn't.

That's really the issue. Aside from going insanely fast I don't know that it will have much of a place on the field. Maybe in conquest just running around capping nodes.

[Thirdstar]

TruePoindexter, on 03 January 2013 - 11:17 AM, said:

That's really the issue. Aside from going insanely fast I don't know that it will have much of a place on the field. Maybe in conquest just running around capping nodes.

It's a raider mech. Which only makes sense if the battlefield had something worth raiding like rear echelon command areas, ammo dumps, radar installations etc etc.

[hammerreborn]

Quick and dirty because the site going down or something messed up my writeup.

ECM/Premade effect on Win Percentage and Prevelance (40 trials):

All 40 trials were done in my Jenner D. All premades were self-reported (thanks guys to all those who did!). As they were self-reported, numbers may be off.

Shocking details observed outside of data:
-Lots of non-ECM lights abound, a bit of the Spider 5V, but a lot of 2Xs and 1B commandos running around. Also, Jenners are back in style <3

-Awesomes, way more than I'm used to seeing. Inflated by current trial, but still lots of 9Ms and other running around probably in preparation for the Hero mech

-Fatlas, Fatlas everywhere. Assaults easily dominated the numbers game (probably at least around 4 a game each side).

And now for the data. If the images suck I'll reupload either in the morning before work or after.

Enjoy!