Rhino v. Psyback

[Inquisitor Fox]

That's what I'm telling you guys...

 

To me, my brain just does that. I'm more concerned about the unit of comparison (kills per point) than the amount of zeros or decimal places, because it's easy to just mentally erase some zeros.

 

A: 0.0160 deaths per point spent

B: 0.0163 deaths per point spent

C: 0.0111 deaths per point spent

D: 0.0166 deaths per point spent

E: 0.0171 deaths per point spent

F: 0.0157 deaths per point spent

G: 0.0167 deaths per point spent

 

is the same as

 

A: 1.60% deaths per point spent

B: 1.63% deaths per point spent

C: 1.11% deaths per point spent

D: 1.66% deaths per point spent

E: 1.71% deaths per point spent

F: 1.57% deaths per point spent

G: 1.67% deaths per point spent

 

Wipe two zeros, move the decimal point. It's the unit of "deaths per point spent" that make it readable. It's EASY to adjust the scale of a chart.. the labeling of the axis though, that's what's important. Just because you can generate a number doesn't make it usable data. It's important to be able to interpret the results and the numbers you're working with. As one of my college instructors essentially told us: "Well what did you find? Great... so what does it MEAN?"

 

It's about meaningful data and the ability to actually have an interpretable result. Comparing apples to apples (variations of the same unit as we're invoking Ceteris Paribus and simply testing a few variations on size, transport, and presence of psybolts) so that the numbers really mean something. It's too easy to get lost in a sea of numbers and lose touch with what the data is actually telling you.

 

Ignoring the obvious outlayer that none of us use, these numbers tell me the real difference between an MSU unit of GKSS and a full squad in a Rhino is less than 0.14% per point. Less than One Percent per point. It's measured in hundreths of a single percent. In the grand scheme of things, statistically I'd say there's essentially no difference. This means the whole thing is really inconsequential until we've exploded the scale so much that the exercise becomes relatively meaningless. My brain thinks in "per point spent" so that's what I use. Generating kills isn't important, because the number of kills available depends on factors I cannot control, but the number of points I can spend on something are. Cost/Benefit maximization and all B)

 

That's why what really makes me wonder about jhrovii's data is the lack of a consistant unit measured throughout in different situations. It's a snapshot of a GKSS unit versus a snapshot of a Purifier unit then trying to claim that one shows that MSU is worse than the other, which isn't the case. It shows an MSU Purifier's unit is less efficient per point spent than a full sized GKSS unit depending upon moving or shooting (which really isn't telling me per point spent or anything useful, he's just proving moving state of the unit, with two totally different things). Totally different conclusion. It's like taking a measurement of a brown haired basketball player, taking a measurement of a redheaded horse jocky, then telling me redheads are more likely to be short.

[barenone]

Now I do not have my codex handy at the moment and I am not sure exactly how it works but maybe you guys would know if this was applicable. Now you are allowed to take a 10 man unit and then break them into 2 squads right? That being the case are you allowed to choose which models got to which squad? if so you could out 2 psycannons into one vehicle but only have to have a 5 man squad in there leaving another 5 man squad to do whatever with. Assuming you are able to do it that way would it make the actually psycannon choice even cheaper following the assumption of having a 5 man squad not pinned up in the rhino? or does this not apply at all?

[jhrovii]

Obviously it's the same pair of numbers just a different ratio.

 

And Inquisitor, you're again missing the point of what I'm measuring. But you're a "math geek," so maybe I can't hope to grasp your argument.

[Inquisitor Fox]

And Inquisitor, you're again missing the point of what I'm measuring. But you're a "math geek," so maybe I can't hope to grasp your argument.

Alright... I am saying that you are making a table comparing apples to oranges. While you correctly are comparing a full sized squad to a "combat squad", you incorrectly are comparing a GKSS to a Purifier squad. This is incorrect because you're changing the unit type, making it no longer a like squad.

 

Let's look again at what you did:

 

http://img824.imageshack.us/img824/3206/ex1f.png

 

Here we see that comparing that the difference between moving and not for the bigger squad has less of an decrease of efficiency (only 35.69%) compare to the MSUs loss of efficiency (59.73%). Not only that, but the decrease in unsaved wounds allocated is by 35.69% compared to 59.73% respectively. Here you see that no matter the point cost, the fact is an EQUAL AMOUNT of Psycannon shots means more in smaller squads.

Do you see how you are saying that a Purifier squad is MSU, and a GKSS is full sized? You aren't measuring the difference in squad size of the same unit, you are measuring the efficiency of Purifiers versus GKSS with psycannons. If you were measuring squad size for MSU comparison, you would have two tables. The first would have a 5 or 6 man GKSS with Razorback, and a full 10 man GKSS without Razorback (or with Rhino). Then on a separate table you would make a 5 or 6 man Purifier Squad with Razorback, and compare that to a full 10 man Purifier Squad with Rhino. That keeps a like unit compared to a like unit, and since squad size is the variable we're changing (using the appropriate transport), we can see the real effect of just a squad size change. You are changing two variables at once, squad size as well as unit type.

 

If I was using that method, I could argue (without any measuring) that if I had a 5 man Purgation squad with psycannons, that their efficiency when moving drops more because they loose a greater percentage of their shots, so they suck compared to a GKSS full sized. But since I'm actually changing the unit type (which allows them more psycannons per man) it skews the data.

 

If you were to redo these tables comparing a 6 man GKSS with Razorback to a 10 man GKSS with Rhino, and a second set comparing a 6 man Purifier Squad with Razorback to a 10 man Purifier Squad with Rhino, you would have more meaningful data. That way you're measuring the real effect of MSU vs Full Sized to efficiency to two different (but common!) unit types.