Boat Momentum Efficiency Challenge

[seanth]

Just now, Ezriilc said:

You ain't kidding.  I used to be able to hit 300 m/s without breaking a sweat, but today I'm lucky if I can get 75 with all my mojo.

So, what about (mass*distance)/time, all measured at a certain destination nice and far away?

I'm working on that now. I had an idea how to deal with it while cooking dinner. Hold tight: I'm looking at some test numbers to see if things are sane

[The_Rocketeer]

Once this challenge works out what it actually is going to be a challenge to do, I shall probably be entering with a modified version of this:

[Ezriilc]

8 minutes ago, The_Rocketeer said:

Once this challenge works out what it actually is going to be a challenge to do, I shall probably be entering with a modified version of this:
-snip-

I like it!

It looks like we have some kind of challenge on!

[The_Rocketeer]

8 hours ago, seanth said:

I can't think of a way to prevent this sort of cheating, so I guess the contest is dead.

There are plenty of challenges that depend on honesty. I think your challenge has a lot of merit.

 

8 hours ago, Ezriilc said:

Divide by zero may not be allowed in the world of computers, but from a practical POV, it certainly works.  If I have 5 apples, and there is zero people to divide them among, then I still have 5 apples.  It's quite the same as dividing by 1, just that no division is happening.

What you've described is not 'quite the same as dividing by 1', it is literally dividing by 1. You are 1, you have 5 apples, 5 apples divided by 1 of you is 5. Now imagine there are just 5 apples, but nobody has them. The number of apples that people have is 0. That's dividing by 0.

In any case, for the purposes of calculating a score there's no problem with using fuel consumption in the formula. If you're going to do a point to point challenge, you'll presumably be using total/average consumption anyway.

[Ezriilc]

2 minutes ago, The_Rocketeer said:

There are plenty of challenges that depend on honesty. I think your challenge has a lot of merit.

 

What you've described is not 'quite the same as dividing by 1', it is literally dividing by 1. You are 1, you have 5 apples, 5 apples divided by 1 of you is 5. Now imagine there are just 5 apples, but nobody has them. The number of apples that people have is 0. That's dividing by 0.

In any case, for the purposes of calculating a score there's no problem with using fuel consumption in the formula. If you're going to do a point to point challenge, you'll presumably be using total/average consumption anyway.

I see what you're saying and I knew there was likely something wrong with my thinking.  Math is not my strong suit.  Meh, we digress.

I think the best way to score fuel consumption, is the act of getting to the target fastest, with the most mass.

[The_Rocketeer]

I agree, tho it would be best to say from one water location to another water location rather than from the runway to a water location. Otherwise the land-scramble to the water could really cock up some people's times. There's still a lot of potential for cheating tho, hence my remarks about honesty - cheaters will always find a way to cheat.

[Ezriilc]

1 minute ago, The_Rocketeer said:

I agree, tho it would be best to say from one water location to another water location rather than from the runway to a water location. Otherwise the land-scramble to the water could really cock up some people's times. There's still a lot of potential for cheating tho, hence my remarks about honesty - cheaters will always find a way to cheat.

I think it's fair to start from the "KSC Beach - Wet" preset in HyperEdit.  If we're going a good long way, then that can be a little flexible as to precise location.

I don't worry much about cheating.  If we always post the .craft file for the ship alleged to have done it, then others can test it.

[seanth]

My kids just went to bed, so I now have some time to think about this and write things down.

See if this makes sense. I'm going to use values from your example rocket-boat, and Watson Your Mind.

In both our screenshots, the resource panel is visible, so I can calculate the total amount of fuel used. WYM used 214L of LiquidFuel, and the unnamed craft used 13,329L of LiquidFuel and Oxidizer combined. Let's assume that we'd also F3ed and saw the total distance traveled. Let's just for the sake of argument say both craft had gone 1km. The images also have velocity and ship mass. Unnamed has 4 parts, and WYM has 80 parts.

Craft	Mass (tonnes)	Velocity (m/s)	Distance (km)	Fuel Used (L)	# Parts	m per L	Momentum (tonnes*m/s)	Momentum/part	Momentum/L	Momentum /L/part
unnamed	114.5	97.4	1	13,329	4	0.075	11,152.3	2,788.075	0.8367	0.209
WYM	34.7	44.0	1	214	80	4.673	1,526.8	19.085	7.1346	0.089
hypothetical	114.5	44.0	1	214	80	4.673	5038.0	62.975	23.54	0.294

Those momentum/L/part scores are not very dignified, so maybe do the momentum cals using kg vs tonnes. So the values would be 209, 89, and 294.

This trade off makes sense to me. Weigh a lot+go fast using lots of fuel with few parts? Great momentum/part score, bad momentum/L score, but good momentum/L/part score. Interestingly, the hypothetical craft ends up getting a score close to the unnamed one. Feels like a good trade-off going on with mass, speed, and fuel usage. This method doesn't need a distance traveled component.

Thoughts?

 

Having said all that, I would absolutely be in for a long distance-no-refueling challenge

Edited July 11, 2016 by seanth

[The_Rocketeer]

Well, I'm a little uncomfortable with raw speed as a measure. If I stick a couple of Hammers on my boat, it'll probably double the speed over 20 seconds or so, but only increase the mass and fuel usage by a fraction. The advantage of using a distance is the law of averages applies to the whole run, which gives a better overall conclusion.

If you're determined to keep it speed-based, I would say entries have to provide 2 screenshots spaced 30 seconds apart of a craft maintaining speed. Speed, time interval, craft mass and the resources panel should be visible in both shots.

Score = (s / f) * m

Where:

s is average of speeds in the screenshots
f is the difference in LF resources between screenshots
m is average of masses in the screenshots

[seanth]

27 minutes ago, Ezriilc said:

I think the best way to score fuel consumption, is the act of getting to the target fastest, with the most mass.

Yeah. I agree, but I'm hesitant about making it a point too far away. I myself am cool with distant points that take an hour or more to sail to, but it might discourage some people from entering.

[Ezriilc]

Seems good to me, but I'm not sure if I fully grok the formula.

 

[seanth]

Just now, Ezriilc said:

Seems good to me, but I'm not sure if I fully grok the formula.

 

(mass*speed)/part number/total amount of fuel used

This gives us three categories for scores:

(mass*speed)/parts = category 1. Efficiency in parts

(mass*speed)/fuel used = category 2. Efficiency in fuel usage

(mass*speed)/parts/fuel used = category 3. Efficiency in both at the same time

 

 

[The_Rocketeer]

I don't really see the relevance of number of parts. If a craft is more fuel efficient surely that's more important than how many parts it's made of?

[seanth]

The downside of doing it this way is that the fuel usage will include the fuel used to get to a good hydroplaning speed, and not show the fuel usage at the time the screenshot was taken....

It too bad it's too easy to cheat on the throttle. Taking the speed and dividing the units of fuel/s in the resource window would give us m/L at the time the image was taken. Assuming the submitter didn't throttle down just before the image was snapped.

[The_Rocketeer]

How is it easy to cheat on the throttle? Throttle setting is visible on the navball. Jet engines will spool down, but if they spool down for more than a second or two speed will noticeably drop. My 30-second interval suggestion directly tackles this problem.

[seanth]

2 minutes ago, The_Rocketeer said:

I don't really see the relevance of number of parts. If a craft is more fuel efficient surely that's more important than how many parts it's made of?

Ezriilc and I did a momentum challenge a while back and we ended up hitting on designs where we just kept adding on more parts in a modular fashion. It was getting to the point that computers were slowing down with part numbers.

The idea is to encourage people to get the job done with fewer parts. Check out the following for the old stuff

 

[The_Rocketeer]

You don't need to use the stated L/s from the resource panel if u have 2 screenshots that tell u 'it was this much 30 seconds ago and it's this much now'. You can calculate average L/s over the 30 seconds from the difference in fuel.

[Ezriilc]

6 minutes ago, The_Rocketeer said:

I don't really see the relevance of number of parts. If a craft is more fuel efficient surely that's more important than how many parts it's made of?

I like the additional constraint of building efficiently as well as building an efficient craft.  But, it's not a deal-breaker.

It seems that elapsed time to a given target, with the most mass (or the least mass lost in transit) is an easier way that doesn't rely on momentary measurements.

[seanth]

1 minute ago, The_Rocketeer said:

You don't need to use the stated L/s from the resource panel if u have 2 screenshots that tell u 'it was this much 30 seconds ago and it's this much now'. You can calculate average L/s over the 30 seconds from the difference in fuel.

I like the dual screenshot idea. That would probably make it harder to have a _very_ low fuel consumption like what happened with the example rocket boat this morning.

I wouldn't say the challenge is just about raw speed. It's momentum. So a ship that is 100kg and traveling an 1m/s would have the same momentum as a ship 1kg traveling at 100m/s.

I think this gives people the ability to adjust speed and mass to match what they are aiming for. The addition of the part #s and fuel usage just add more space for people to play in. 

[Ezriilc]

If you start in clamps and use HyperEdit, then the MET doesn't start until you teleport to the beach.  That makes tracking elapsed time easy.

The dual screenshot idea seems like a pain and difficult to get perfect.  I'd prefer a method that is easier to track and see.

[The_Rocketeer]

vOv to me it's an efficient build if it does the job better. I'd be surprised if modular boats were still competetive. I've got everything in your old thread licked already.

1 minute ago, Ezriilc said:

The dual screenshot idea seems like a pain and difficult to get perfect.  I'd prefer a method that is easier to track and see.

It doesn't need to be perfect. You're using averages, so it doesn't matter if u do 30 seconds or 20 minutes as long as the speed stays pretty constant.

In fact even if it isn't constant, that only diminishes your overall score - you're never going to try to do less than your best speed if u want a high-score.

Edited July 11, 2016 by The_Rocketeer

[Ezriilc]

2 minutes ago, The_Rocketeer said:

vOv to me it's an efficient build if it does the job better. I'd be surprised if modular boats were still competetive. I've got everything in your old thread licked already.

For pure momentum, at least back then, there was no end in sight.

Here's the way I picture the formula, with distance as a constant:  mass/time/part-count.  Or at least something similar.

[seanth]

4 minutes ago, Ezriilc said:

The dual screenshot idea seems like a pain and difficult to get perfect.  I'd prefer a method that is easier to track and see.

Well, one way around this is if we say you _have_ to use MechJeb. That way there can be a window showing max acceleration and current acceleration. If max acceleration is significantly greater than current acceleration when the screenshot is taken, that's a pretty good indicator that there is something fishy going on.

I personally would love if we can use (mass*speed)/(parts*fuel per second)

 

[The_Rocketeer]

I see the issue with momentum - build a behemoth and move it, you win - but fuel efficiency has nothing to do with momentum. This challenge still seems very confused about what it is supposed to be.

[Ezriilc]

1 minute ago, seanth said:

Well, one way around this is if we say you _have_ to use MechJeb. That way there can be a window showing max acceleration and current acceleration. If max acceleration is significantly greater than current acceleration when the screenshot is taken, that's a pretty good indicator that there is something fishy going on.

I personally would love if we can use (mass*speed)/(parts*fuel per second)

 

Honestly, I think it needs to be simpler.  If I need to throttle down during my run, then I should be able to.  The total elapsed time, and total fuel used, is all that should matter.