Speed Vs. Payload Overview
#21
Posted 29 May 2013 - 05:12 PM
#22
Posted 31 May 2013 - 12:02 AM
#23
Posted 31 May 2013 - 01:06 AM
Modo44, on 31 May 2013 - 12:02 AM, said:
Yes they do change a lot for an individual mech, but they don't change much in the comparison between different mechs if you apply these upgrades to every mech.
E.g. using an XL engine, AWS and CTF have the same pay load (- armor) at roughly 76 km/h no matter if they both use endo or both don't.
#24
Posted 31 May 2013 - 01:41 AM
#25
Posted 19 July 2013 - 02:27 AM
#26
Posted 19 July 2013 - 02:55 AM
Edited by Tragos, 19 July 2013 - 02:55 AM.
#27
Posted 19 July 2013 - 01:50 PM
#28
Posted 21 July 2013 - 07:57 PM
#29
Posted 22 July 2013 - 05:47 AM
STD Engine, but with Endo.
I mean, if I think about putting an STD engine in, then Endo Steel automatically becomes mandatory except for the most heavy builds.
#30
Posted 22 July 2013 - 08:52 AM
Windsaw, on 22 July 2013 - 05:47 AM, said:
STD Engine, but with Endo.
I mean, if I think about putting an STD engine in, then Endo Steel automatically becomes mandatory except for the most heavy builds.
I'd disagree that endo becomes mandatory when using a STD engine (most of the time I'm using a STD engine to get 2 more DHS). Of course there are cases this is useful, e.g. AC/20 in side torso.
Nevertheless, I added the graph. Voilà! Put some spoiler tags not to bloat the post too much.
Edit: Creating new graphs is neither a problem nor very time-consuming (any more), since I automated it to a high degree. 80 % armor? 2 % of max tonnage extra weight per 'Mech? Only 'Mechs heavier than 80 tons or faster than 40 km/h? No problem.
Edited by Phaesphoros, 22 July 2013 - 08:58 AM.
#31
Posted 22 July 2013 - 03:36 PM
Personally, I tend to have the same amount of armor on legs as i do the arms, on non-jumping heavies and assualts.
Any who, thanks for working the numbers. Greatly appreciated.
#32
Posted 22 July 2013 - 03:54 PM
DanNashe, on 21 March 2013 - 09:53 AM, said:
Cent: safe to xl at at least 90 kph.
I've Piloted Cents a lot and the XL engine thing with them is pretty debatable. If you're going 98 kph in a cn9-a, people are gonna aim for your legs usually, but getting hit in the side torso can happen too especially if they aim for your CT and your speed throws them off. So I kinda agree 90 kph and higher XLs can be worth it, but it's more so "slightly risky" instead of a flat safe.
#33
Posted 23 July 2013 - 11:14 AM
If you could figure out how to include that, it would be real cool!
EDIT:
example http://mwo.smurfy-ne...a7f0423e78304d5
Edited by Pwnius, 23 July 2013 - 11:28 AM.
#34
Posted 23 July 2013 - 12:02 PM
Pwnius, on 23 July 2013 - 11:14 AM, said:
If you could figure out how to include that, it would be real cool!
EDIT:
example http://mwo.smurfy-ne...a7f0423e78304d5
I wouldn't say they forget to include the extra heat sink slots for >= 300 rated engines. It is an important factor for hard points, but not so much for payload. Yes, it might allow you to use endo.
But these kind of data visualizations always need some restrictions / simplifications, otherwise it couldn't be a useful overview. E.g. including all upgrades (endo, FF) and engine types (STD, XL) in one graph would just result in a mess, even if that is useful for some kind of comparison.
I could make a graph that uses endo-steel structure IFF the engine rating is greater than some value (like XL350 for 4*3 = 12 DHS critical slots) if that'd be useful for you.
#35
Posted 23 July 2013 - 01:27 PM
Phaesphoros, on 23 July 2013 - 12:02 PM, said:
But these kind of data visualizations always need some restrictions / simplifications, otherwise it couldn't be a useful overview. E.g. including all upgrades (endo, FF) and engine types (STD, XL) in one graph would just result in a mess, even if that is useful for some kind of comparison.
I could make a graph that uses endo-steel structure IFF the engine rating is greater than some value (like XL350 for 4*3 = 12 DHS critical slots) if that'd be useful for you.
Well, the weight value of a slot can be calculated. For a dragon with max armor, the armor/skeleton upgrades will save 5.66t (total). This makes each slot worth ~.2 tons.
Each heat sink slot is worth 3 slots of double heat sinks = .6 tons .
Simply adjust the larger engines by -.6 tons per heatsink slot.
However, the value of each slot will need to be calculated for each 'mech.
Edited by Pwnius, 23 July 2013 - 01:29 PM.
#36
Posted 24 July 2013 - 03:53 AM
Pwnius, on 23 July 2013 - 01:27 PM, said:
Here you go:
However, I still think this metric is not useful for everybody. More critical slots doesn't necessarily mean you'd better put endo on it, for example if you need as much DHS as possible (where for most chassis >= 50 t, critical slots are the bigger problem).
#37
Posted 24 July 2013 - 11:32 AM
Edited by Pwnius, 24 July 2013 - 11:35 AM.
#38
Posted 25 July 2013 - 08:48 AM
So it seems a general rule of thumb is that the 300 engine (STD or XL) is the most efficient?
In particular, the QKD and DRG with 300 engines seem to be about the best bang per buck possible.
Particularly heartening to see that for all my fave builds, they seem to sit exactly in the sweet spot for that chassis
Is there any chance this data could be made into a graph that can be examined more closely? Could your automation process be tweaked to create the graphs as a google charts page?
If you are interested in doing this but need assistance with the coding or whatever, I would be willing to help.
#39
Posted 25 July 2013 - 10:55 AM
evilC, on 25 July 2013 - 08:48 AM, said:
So it seems a general rule of thumb is that the 300 engine (STD or XL) is the most efficient?
In particular, the QKD and DRG with 300 engines seem to be about the best bang per buck possible.
Particularly heartening to see that for all my fave builds, they seem to sit exactly in the sweet spot for that chassis
Is there any chance this data could be made into a graph that can be examined more closely? Could your automation process be tweaked to create the graphs as a google charts page?
If you are interested in doing this but need assistance with the coding or whatever, I would be willing to help.
Hmm interesting idea. The program I'm using currently, gnuplot, can already create somewhat interactive graphs via SVG. In fact, I currently do export to SVG, adjust some things manually and finally export to png.
The problems with interactive graphs are:
1) I currently don't have any website to host them
2) can't probably insert them into the forums directly
Is there any way to host the interactive graphs on google? (I've lost track of their services long ago o.O)
From a technical point of view, I don't think it would be a problem to use google charts. I currently extract the data from the game files, process them, and pass the results to gnuplot. Could as well use google charts for the output part. It should even be possible to get the data off smurfy's API and have the whole thing online.
The one problem I don't know atm how to solve - aside from hosting - is that I currently adjust the placement of the labels manually. They're inserted and placed automatically, but my algorithm isn't sophisticated enough so I'll have to manually adjust them. Using a legend only isn't very useful due to the number of lines. I didn't find any feature in google graphs how to place labels (besides the annotation function), do you have any idea?
#40
Posted 25 July 2013 - 11:41 AM
All you would need to do is have a bit of javascript that would generate the graph, and they could easily add it to their page.
Do not forget that html / js works on a local computer - you can build the code required to generate the page, then package it as a zip file and send to the webmaster who adds it to his site.
I would maybe look for a piece of gnuplot -> html software? I have no experience of gnuplot, but if it is a gnu app, chances are someone has written some kind of converter.
The label problem could be solved via a key, and also when you mouse over a line it could show the label as a tooltip.
6 user(s) are reading this topic
0 members, 6 guests, 0 anonymous users