19 replies to this topic

[Artgathan]

With the Medium Vs. The World tournament concluded all we're left with is the leader boards. Congratulations to everyone who placed in the top 10 (especially Eglar, who took home 1st in Shadowhawks and Heavies - go get some sleep now). I decided now was a good time to run some stats on the information available from the tournament and see if it reflected any themes seen elsewhere on the forum.

So I took the Top 10 scores from each category and averaged them. These are the results by weight class (medium score was determined by averaging the average score of each medium):

• Lights: 2463.90
  
• Mediums: 2290.15
  
• Heavies: 2398.50
  
• Assaults: 2548.20

I then compiled average score for the weight classes and used it to generate a standard deviation. Thus, the mean for the scores was 2425.19 and the standard deviation was 94.31.

This generated the following distribution:

• Lights: + 0.41 SD
  
• Mediums: - 1.43 SD
  
• Heavies: - 0.28 SD
  
• Assaults: + 1.30 SD

We can see here that Mediums are scoring far below average, while Assaults are scoring quite high.

If we look at the mediums in more detail - Mean within chassis (Standard Deviation from Medium mean):

• Cicada: 2299.2 (+0.13)
  
• Blackjack: 2310.7 (+0.29)
  
• Centurion: 2275.1 (-0.21)
  
• Hunchback: 2263.5 (-0.38)
  
• Trebuchet: 2218.5 (-1.03)
  
• Kintaro: 2243.6 (-0.67)
  
• Griffin: 2252.7 (-0.54)
  
• Shadowhawk: 2457.9 (+2.42)

The stats for the mediums were a mean of 2290.15 and a standard deviation of 69.18. I ran various t-tests - only the Shadowhawk vs Trebuchet comparison came back significant (p < 0.01). Interestingly, the Cicada and Blackjack managed to outperform every other medium except the Shadowhawk (I did not expect this - I figured the heavier mediums would pull ahead).

What does this say about balance in MW:O? Mediums are lagging behind other classes, while Assaults are pulling ahead. Within Medium mechs the Shadowhawk is indisputably the best in the hands of a capable pilot.

[Levi Porphyrogenitus]

Artgathan, on 03 March 2014 - 06:06 PM, said:

Interestingly, the Cicada and Blackjack managed to outperform every other medium except the Shadowhawk (I did not expect this - I figured the heavier mediums would pull ahead).

The Blackjack has some very strong meta performance for its weight (exemplified by the AC20 / 3 ML JJ build), and the Cicada is a glorified light (which as a class performed strongly).

[Bagheera]

Nice work Artgathan. The numbers tell the story. Not much for me to add, as they sort of reflect what many of us have been saying for a while.

On thing that this highlights: TBT and SHD suffer the same super-size syndrome. What makes the latter so much better than the former? I have my own theories on that (posted a few times about how the TBT got shafted because it was released while SRMs still used broken splash damage), but interested in hearing others.

Edited by Bagheera, 03 March 2014 - 07:15 PM.

[Deathlike]

Bagheera, on 03 March 2014 - 07:15 PM, said:

On thing that this highlights: TBT and SHD suffer the same super-size syndrome. What makes the latter so much better than the former? I have my own theories on that (posted a few times about how the TBT got shafted because it was released while SRMs still used broken splash damage), but interested in hearing others.

I would say that Trebuchet was "feared" internally, because the most notable nerf was with the arm articulation. It's 5 degrees worse than the norm (20 degrees standard, 10 is the "lowest" because of the AC20 arms). SRMs never put the Trebuchet into power (the Cent-A held that honor, due to awesome convergence in just one torso, and the NARC tube on the 7M)... it was probably LRMs that had benefited more at the time. Right now, it does nothing worth mentioning other than being the worst 50 tonner that has JJs. That's a sorry state if you ask me.

Edited by Deathlike, 03 March 2014 - 07:33 PM.

[Artgathan]

Bagheera, on 03 March 2014 - 07:15 PM, said:

Nice work Artgathan. The numbers tell the story. Not much for me to add, as they sort of reflect what many of us have been saying for a while.

On thing that this highlights: TBT and SHD suffer the same super-size syndrome. What makes the latter so much better than the former? I have my own theories on that (posted a few times about how the TBT got shafted because it was released while SRMs still used broken splash damage), but interested in hearing others.

Deathlike, on 03 March 2014 - 07:32 PM, said:

I would say that Trebuchet was "feared" internally, because the most notable nerf was with the arm articulation. It's 5 degrees worse than the norm (20 degrees standard, 10 is the "lowest" because of the AC20 arms). SRMs never put the Trebuchet into power (the Cent-A held that honor, due to awesome convergence in just one torso, and the NARC tube on the 7M)... it was probably LRMs that had benefited more at the time. Right now, it does nothing worth mentioning other than being the worst 50 tonner that has JJs. That's a sorry state if you ask me.

I was aghast when the Trebuchet was released with a paltry 3 missile hardpoints. I strongly suspected that it was due to the dominance of SRMs at the time (which lead to PGI fearing it would supplant every other medium in the game at that point). I had never noticed the Trebuchet had such a terrible arm angle as well until you mentioned it Deathlike (I've never piloted one).

In general I think mechs with articulated arms should have a greater range of motion (which would allow truly skilled pilots to shine as they could protect their CT will firing arm-mounted weapons).

I think the biggest problems with the Trebuchet are:

• Enormous Size
  
• Bad Hitboxes (makes it easy to kill Side Torsos)
  
• Reliance on Missiles and Energy
  
• Medium Mech Mobility

While the Shadowhawk also has two of those problems (Large Size + Medium Mech Mobility), the fact that SHD's can easily carry an XL (freeing up space for a large engine, which then negates some of the mobility problem, or freeing up space for ballistics) and has high-mounted weapons makes up for these faults.


While both mechs are tall (which is a problem) the Shadowhawk is skinny from shoulder to shoulder and wider from front to back where as the Trebuchet is a barn door from shoulder to shoulder and a pencil from front to back. This benefits the Shadowhawk because while in combat you must face your target at some point to fire (which exposes your shoulder to shoulder size). Skilled pilots won't fire at a target that is presenting arms / side torsos (unless they suspect there's an XL in the target), so being pencil-thin is pointless (and thus the Trebuchet's small side profile is useless).

EDIT: To build on a point Deathlike made - consider that Jumpjets are considered to be a force multiplier (IE: if a mech can carry JJs it's usually considered to be vastly superior to non-JJ mechs). The fact that the Trebuchet is the only 50-tonner to carry Jump Jets and is still the worst should set of some serious alarm bells!

Edited by Artgathan, 03 March 2014 - 07:56 PM.

[Davers]

Thank you Artgathan for showing everyone that tonnage limits would not redeem the medium class, as it is outperformed greatly by lights and heavies.

[Deathlike]

Artgathan, on 03 March 2014 - 07:55 PM, said:

I think the biggest problems with the Trebuchet are:

• Enormous Size

Needlessly tall actually.

Quote

• Bad Hitboxes (makes it easy to kill Side Torsos)

Believe it or not, the hitboxes aren't bad and neither are the side torsos. It's actual weakness are the arms considering where most of the firepower comes from them (the 7K being the lone exception). You don't want Centurion sized arms carrying the firepower...

Quote

• Reliance on Missiles and Energy

That isn't really a big deal IMO, unless you're bad at heat management. Higher skill required though.

Quote

• Medium Mech Mobility

Well, medium mech mobility is bad in general, particularly for the slower mechs that can't have the larger engines. The irony is that the Trebuchet has those options (325XL max) and it's still bad.

Quote

While the Shadowhawk also has two of those problems (Large Size + Medium Mech Mobility), the fact that SHD's can easily carry an XL (freeing up space for a large engine, which then negates some of the mobility problem, or freeing up space for ballistics) and has high-mounted weapons makes up for these faults.

Buckets are actually very XL friendly. They simply get cored-CT very easily.

Quote

While both mechs are tall (which is a problem) the Shadowhawk is skinny from shoulder to shoulder and wider from front to back where as the Trebuchet is a barn door from shoulder to shoulder and a pencil from front to back. This benefits the Shadowhawk because while in combat you must face your target at some point to fire (which exposes your shoulder to shoulder size). Skilled pilots won't fire at a target that is presenting arms / side torsos (unless they suspect there's an XL in the target), so being pencil-thin is pointless (and thus the Trebuchet's small side profile is useless).

I don't think that's the issue exactly. One of the things about the Shadowhawk that the Trebuchet DOES NOT have are high weapon firing points. The arms of the Shadowhawk are still low, but the ballistic hardpoint is pretty high up, allowing for an AC20 to the tri-dakka AC2 builds to shine (assuming, you can handle being semi-blinded by that). The Trebuchet has a vastly LOWER firing point (except for the 7K's ballistic hardpoints) which forces you to spend more time in the air to get the shot off. That is the actual problem.

An underwhelming mech (according to the masses) like the Quickdraw has the same problems, but it has high mounted firing weapon points in the torso. Same can be said for the Dragon variants... the torso hardpoints (outside of the missiles) are mounted rather high, which are an understated benefit of the chassis.

Quote

EDIT: To build on a point Deathlike made - consider that Jumpjets are considered to be a force multiplier (IE: if a mech can carry JJs it's usually considered to be vastly superior to non-JJ mechs). The fact that the Trebuchet is the only 50-tonner to carry Jump Jets and is still the worst should set of some serious alarm bells!

Pretty much everything that can be bad about a mech, and not have JJs save it, goes to the Trebuchet (although the Spider-5V is at the very top of the list, despite having hitbox/hitreg issues).

[Bagheera]

I don't recall the state of LRMs when the TBT was released, but that could have been another consideration in gimping the TBT. I was definitely expecting at least one of the variants to have 4M, but that ended up not being the case. Couple that with the single tube launcher on the 7M and the spread launchers on the fast TBT (3C? I forget) and using them for standard SRMs was pretty much out the window. I suppose there could have been some concern that a sufficiently fast TBT could dictate range and use directLOS/Tag with LRMs on a slower opponent. In practice though that's a pretty dubious tactic that really only works well if you catch a lone slow poke. If he's got friends (as a slow mech should) then it's trouble.

In fairness, I used to run a 4SP during the Fastback hayday and did really well with it. iirc it even ran an xl325 just like a Trebbie can. So I may be a little biased.

The high mount weapon is definitely a factor, and I wonder if that is indirectly one of the reasons the SHD is so much more durable than the TBT. IE - the SHD can take more advantage of cover. I know there is a little more armor and a little more IS on the SHD than the TBT, but the durability difference goes well beyond that, imo.

Articulation and twist are definitely big marks against the TBT, but I don't find the SHD to feel that much more nimble either. Of course I'm spoiled by HBK twist so just about everything else feels a bit restricted.

[Artgathan]

Deathlike, on 03 March 2014 - 08:35 PM, said:

Believe it or not, the hitboxes aren't bad and neither are the side torsos. It's actual weakness are the arms considering where most of the firepower comes from them (the 7K being the lone exception). You don't want Centurion sized arms carrying the firepower...

That [reliance on energy and missiles] isn't really a big deal IMO, unless you're bad at heat management. Higher skill required though.

Well, medium mech mobility is bad in general, particularly for the slower mechs that can't have the larger engines. The irony is that the Trebuchet has those options (325XL max) and it's still bad.

Buckets are actually very XL friendly. They simply get cored-CT very easily.

Pretty much everything that can be bad about a mech, and not have JJs save it, goes to the Trebuchet (although the Spider-5V is at the very top of the list, despite having hitbox/hitreg issues).

I find that Trebuchets go down to losing a side torso quite often (when I do encounter them). I agree that the enormous arms don't help.

As for the weapons - the problem is that missiles and energy are not great for dealing the "killing blow". At the moment both energy (except the PPC) and missiles are a little sub-par: all the mechs that rely heavily on them are not amongst the dominant chassis at the moment. Sure, heat management can help but mechs that rely on Energy / Missile combinations need to be able to hit & run (so that they can cool off), otherwise they get stuck with high heat and a significant reduction in their ability to throw out damage.

Looking at their mobility: most Trebuchets have the same amount of torso twist range as the Highlander.

[Asmudius Heng]

Trebs are mainly bad because missiles are mostly bad.

I ran a treb using PPCs and lasers i think and my effectiveness went through the roof.

I am not a brilliant LRM user but even in my best games where everyone ignored me and i could pump out my LRMs til the ammo bins ran dry .... it was incredibly ineffective.

Its role as a mobile LRM boat just does not exist because LRMs are really poor, get eaten alive by AMS, and travle so slow.

I bought my trebs when LRMs were good and it was amazing then, with JJs i could find angles all over the place to put in LRM shots then fade away .... not i can do the same but the damage from the LRMs is inconsequential.

Anyway this was a thread about mediums in general but thats my 2c on the poor Treb.

[Reno Blade]

The results show us that faster/lighter mediums (= closer to light mechs) are stonger than other mediums because they can avoid more shots and also use the meta weapons better (PPC and AC20) while keeping the very high speed and smaller profile.

And that the heavier mediums with best meta hardpoints (= closer to heavy/assault meta mechs) are stronger than other mediums, because they can use more Meta weapons better.

[wanderer]

Why'd the Shadow Hawk do so well?

It's at the "sweet spot" for engines and jump jets for great mobility.

It's the best protected medium (with other 55 tonners) and mobile enough to spread damage.

Best poptart medium with ideal hardpoint type and location for PPC/AC or multi-AC goodness. Alternatively, Streak-o-rama.

Basically, it's the Jenner of mediums.

[Artgathan]

Reno Blade, on 04 March 2014 - 04:06 AM, said:

The results show us that faster/lighter mediums (= closer to light mechs) are stonger than other mediums because they can avoid more shots and also use the meta weapons better (PPC and AC20) while keeping the very high speed and smaller profile.

I'm not entirely sure we can reach that conclusion from the data. The Blackjack is not renowned for it's speed (only one variant can reach speeds over 94 kph). Conversely Trebuchet, Kintaro and Griffin are more mobile than the Cenutrion and Hunchback but scored lower.

I think the more common trend is access to autocannons (this is less true on the Cicada). If we order the mediums based on their average score we get this (from highest to lowest score):

• Shadowhawk
  
• Blackjack
  
• Cicada
  
• Centurion
  
• Hunchback
  
• Griffin
  
• Kintaro
  
• Trebuchet

The three worst performers are the mechs that are restricted to Missile / Energy loadouts (with the exception of the TBT-7K). Given that the purpose of the competition is essentially to put out the most damage (which then garners kills and assists) possible, this would support the notion that energy and missile builds are under performing relative to builds that utilize ballistic weapons.

Likewise mobility doesn't seem to come much into play, as the two least mobile mediums (the Centurion and Hunchback) scored firmly in the middle. That said, discussing the role of mobility in combat during a competition about dealing damage may not be the most productive.

[Bagheera]

Re: the ordered list.

If the deviation was narrower that would probably be fine for the way the game eschews roles. 3.45 is way too much of a difference within a single weight class in the current game play.

[Charons Little Helper]

Artgathan, on 03 March 2014 - 06:06 PM, said:

So I took the Top 10 scores from each category and averaged them. These are the results by weight class (medium score was determined by averaging the average score of each medium):

• Lights: 2463.90
  
• Mediums: 2290.15
  
• Heavies: 2398.50
  
• Assaults: 2548.20

So basically - you're taking the average of the top 10 scores for lights'/heavies'/assaults' vs the average of the top 80 mediums' scores and think that proves mediums suck?



You should look into a little something I like to call 'statistics'.

(frankly - while your reasoning is BAD - there simply isn't enough knowledge of the number of players / games involved to come to any real conclusions from the top score info unless the differences were FAR more dramatic than they are)

Edited by Charons Little Helper, 05 March 2014 - 01:50 PM.

[Svidro]

Yeah, I see three major problems here. 1. The lowest scoring chassis also had the lowest number of players. I had one game on BJ and one game in a light. As a light I was at 2000something-th place, and as a blackjack I was around 600. With a much smaller population, the competition will generally be lower.
2. Being the meta, more players, and more importantly more *good players had the chassis/weight classes that are most popular. I imagine Eglar or Proton would have blown away the trebuchet bracket if they had bothered to try, partially due to knowing where and when to airstrike!
3. The other weight classes look stronger because you aren't splitting them. How many of the top 10 Assault pilots do you think were in an Awesome? Top 10 light pilots in a Locust? You are comparing the worst medium chassis against the best light/heavy/assault chassis.

Edited by Svidro, 05 March 2014 - 02:05 PM.

[Charons Little Helper]

Oh - and if anyone's interested - the average of the top 10 medium scores was actually 2466.8.

That must mean that mediums are better than heavies.

[Grits N Gravy]

Notice the top two other medium mechs are typically the fastest. The medium mechs are the most hampered by the engine restrictions imposed in closed beta. Give them the ability to run at speeds of 115 kph and they all become much more survivable and hence viable. The need for the engine restrictions for mediums has largely been mitigated by improved hit detection and various balancing factors imposed since closed beta, ghost heat, ect. Without change, even in a 3,3,3,3 environment the medium mech will still be a black-sheep because it offers no advantage to the player. Self interest being what it is, we are all going to be waiting for someone else to roll mediums before we can launch.

[Corvus Antaka]

The treb is fine, great mech really.

problem is only 1 treb has a ballistics slot, and that treb doesnt have jumpjets.

you figure out the rest

[Firewuff]

And you make 2 classic statistical mistakes. A) the sample set is WAY too small to have any significances so any alanysis is fundamentally flawed. Secondly you are looking at the tail of the population by looking at only the winiers. These are effectively outliers in the first place. So not representative and highly volatile data set and not representative of the population

So ... Um... Not useful and proves nothing