30 replies to this topic

[Hawks]

In 1916 the British mathematician and engineer Frederick Lanchester devised a set of equations which describe the relative strengths between predator/prey pairs.

Simply put, these state that, in the case of ancient melee combat (eg between massed ranks of spear- or pike-armed men) there is a simple, linear relationship between the relative strengths of the opposing forces. Thus a force of ten spearmen will be twice as effective as a force of 5 spearmen (assuming all else is equal).

Great, you're probably thinking, he's stated the obvious. So far, so good.

However, in the case of ranged combat, things get a bit more complicated.

In the case of ranged combat (for example between groups of dreadnought battleships, or tanks, or indeed of battlemechs) individual units can give fire to and receive fire from a multitude of directions. What Lanchester's laws showed was that in such a scenario, the relative strength of such a force was proportional not to the size of the force but to the square of its size. Thus a force of ten riflemen will be, not twice, but four times as effective as a force of five riflemen (ie 100 v 25 - again, assuming all else is equal)

How does this affect us? Simple. In an 8v8 match, if one player disconnects, it's not 7v8. It's 49v64. Two players on the same team disconnect, and it's 36v64.

It obviously doesn't just relate to disconnects, either. If you die stupidly at the start of the match you are utterly shafting your team even worse than you thought you were. So don't do that.

[jtyotJOTJIPAEFVJ]

Interesting maths, bro. Although I'm not sure why this needs a discussion thread. You can't really have a proper discussion on the topic of disconnects being bad.

[El Bandito]

That feeling when your scout dies 45 seconds into the game from multiple PPC blasts.

Pretty bad.

[Shazarad]

I've always had a feeling that the team to lose the first mech most often loses.

[xenoglyph]

So what you're saying is back in the day when we had really imbalanced matches, winning a 5v8 basically made us badasses?

I suspected it all along.

[Hawks]

xenoglyph, on 04 July 2013 - 04:10 AM, said:

So what you're saying is back in the day when we had really imbalanced matches, winning a 5v8 basically made us badasses?

I suspected it all along.

It's possible. Certainly the equation suggests the odds would be 2.5 to 1 against such an outcome.

Edited by Hawks, 04 July 2013 - 05:52 AM.

[William Mountbank]

Doesn't this equation make certain assumptions, for instance that all mechs are able to fire on, and receive fire from, all other mechs at all times? If you take into account confounding variables, then the numbers will move closer together.

Anyway, didn't we already have a thread on concentrating fire?

[Rengakun]

xenoglyph, on 04 July 2013 - 04:10 AM, said:

So what you're saying is back in the day when we had really imbalanced matches, winning a 5v8 basically made us badasses?

Y'know, once, I got into a 4v8. Then someone on my team disconnected. So it turned into a 3v8. We somehow won despite the fact that I had a ****** computer at that time.

[Goose]

Hawks, on 04 July 2013 - 03:35 AM, said:

So don't do that.

Are you advising for or against recon?

[Gaan Cathal]

William Mountbank, on 04 July 2013 - 06:22 AM, said:

Doesn't this equation make certain assumptions, for instance that all mechs are able to fire on, and receive fire from, all other mechs at all times? If you take into account confounding variables, then the numbers will move closer together.

Not exactly. It assumes that the firing- and fired-upon opportunities are approximately equal. But yes, it does equate to the odds of victory in a firefight, not a battle. Ergo if there's a disconnect then the result may be a 4v3 and a 4v4 at different locations.

[Flagrant]

That makes sense in a general sort of view. But doesn't that assume all participants are ready to shoot like a line of rifleman staring at each other?

[MustrumRidcully]

Gaan Cathal, on 04 July 2013 - 08:26 AM, said:

[/size]

Not exactly. It assumes that the firing- and fired-upon opportunities are approximately equal. But yes, it does equate to the odds of victory in a firefight, not a battle. Ergo if there's a disconnect then the result may be a 4v3 and a 4v4 at different locations.

WHich is also why moving in blobs works rather well. You make it less likely to be caught in a (n-x)²vs(n)² situation... And it's requires game modes where this difference is compensated by other rewards / benefits if you split up.

[InRev]

Goose, on 04 July 2013 - 08:18 AM, said:

Are you advising for or against recon?

Most of the time, in this game at least, recon isn't really necessary. Alpine and maybe Tourmaline are the only maps that need a scout to find the main blob and to prevent a caprush. Otherwise, everyone goes to the same ridges basically every time. This is even more so true with the movement changes as certain bottlenecks have become hardened.

So, running off and getting killed in the name of recon is hurting your team more than helping. Role Warfare at its best!

[Alistair Winter]

Did someone actually post a logical argument based on scientific work involving math, and include a source? We don't like your kind around here and we don't take kindly to people using words like 'proportional' and 'linear'.

But anyway, yeah. Thanks for the lesson. It confirms what a lot of us have said for a while - the 8-0 matches aren't necessarily a result of the ELO not working*. This is also part of the reason why the "Blob formation with Focus Fire" strategy has survived any encounter with more complex strategies.

* The point is unfortunately moot, because ELO is clearly not working.

[Deathlike]

Begone! PGI does not take kindly to math. Next thing you know you're talking about how the 6 MG Jager is woefully bad, and how the LPL is worse than a PPC!

[PEEFsmash]

Hawks, on 04 July 2013 - 03:35 AM, said:

In 1916 the British mathematician and engineer Frederick Lanchester devised a set of equations which describe the relative strengths between predator/prey pairs.

Simply put, these state that, in the case of ancient melee combat (eg between massed ranks of spear- or pike-armed men) there is a simple, linear relationship between the relative strengths of the opposing forces. Thus a force of ten spearmen will be twice as effective as a force of 5 spearmen (assuming all else is equal).

Great, you're probably thinking, he's stated the obvious. So far, so good.

However, in the case of ranged combat, things get a bit more complicated.

In the case of ranged combat (for example between groups of dreadnought battleships, or tanks, or indeed of battlemechs) individual units can give fire to and receive fire from a multitude of directions. What Lanchester's laws showed was that in such a scenario, the relative strength of such a force was proportional not to the size of the force but to the square of its size. Thus a force of ten riflemen will be, not twice, but four times as effective as a force of five riflemen (ie 100 v 25 - again, assuming all else is equal)

How does this affect us? Simple. In an 8v8 match, if one player disconnects, it's not 7v8. It's 49v64. Two players on the same team disconnect, and it's 36v64.

It obviously doesn't just relate to disconnects, either. If you die stupidly at the start of the match you are utterly shafting your team even worse than you thought you were. So don't do that.

This is a minor point about disconnects, but a more important point about why so many matches end 8-0 or 8-1. After the first kill, the odds start significantly stacking in your favor. The 8-0 process is a statistically expected one.

[Hawks]

PEEFsmash, on 04 July 2013 - 11:06 AM, said:

This is a minor point about disconnects, but a more important point about why so many matches end 8-0 or 8-1. After the first kill, the odds start significantly stacking in your favor. The 8-0 process is a statistically expected one.

Very much so, yes - especially if the first mech to die is one with a disproportionately high amount of the force's total combat power (eg an Atlas). The n-squared laws assume each individual element is of equal strength. Where this is not the case, the effect may be even more pronounced (and as a corollary, of course, if the first mech to die is a particular weak one then the effect may be much less pronounced).

[Whompity]

That makes sense. I've often wondered the impact of a single player loss on game momentum. Nice to see it in numbers.

[scJazz]

Piipu, on 04 July 2013 - 03:44 AM, said:

Interesting maths, bro. Although I'm not sure why this needs a discussion thread. You can't really have a proper discussion on the topic of disconnects being bad.

You can actually...

I could point out that the same DC Farmer DC'd in 13 of my first 20 matches yesterday. When he was on the other team ours won 6/7 matches and when he did it to my team we lost 5/6. When the same player showed up and started up his routine again later in the day I just quit playing. Even though I had 1 day of premium time I let half of it go wasted.

[Raso]

This would explain why shut-outs are so common. Basically once the first domino falls the rest become that much more at risk. Any advantage the winning team had before the losing team lost a man become increased exponentially.