Mythbusters: RA2 Edition

[Clickbeetle]

In this first installment of Mythbusters: RA2 Edition, we're testing the myth that a bigger chassis has more HP than a small chassis.

At first I thought this myth was true because of the stock AI.  They all have a huge chassis, and they're surprisingly durable.  Even the best popup is incapable of one-hit KOing Big Dog.

But then I found out that the stock AI are actually more durable because they all have double strength armor (even DS steel and titanium).  Apparently, in an early version of the game all the armor was twice as strong.  But when the developers cut the armor durability in half, they forgot to adjust the stock AI accordingly.  (Incidentally, this is also why DS aluminum exists in the present game.  The developers also forgot to edit the default chassis armor.)  That evidence makes it look like Myth: Busted.

Now recently Naryar boldly claimed the myth is true, so I decided to find once and for all the truth.

In order to test this myth, I made a modified Omni AI routine that will do exactly 10 damage to itself at the beginning of the match.  Then, it uses the getHealth command to find out the percentage of its HP remaining after the 10 damage, and from that, calculate total HP.

For anyone interested, you can download this AI >here.<

You can just open Bindings.py and replace "Omni" with "Omnihptest" and then play RA2 to find out how much HP the bot has.

Now, on to the myth.  I started testing with the smallest possible chassis with DS Aluminum armor.



So this chassis has 60 HP.  To put things in perspective, extenders have 100 HP and most weapons have 1000.  So 60 is like, nothing.

Now let's test the biggest possible chassis, with the same armor, and see if it has more than 60 HP.



0.0  4259 HP?  That's no random variation.  Looks like the bigger chassis actually has a lot more HP than the small one!

Of course, you're never going to use either of those test chassises on a serious bot, so let's see how much HP is on a more realistic design.

First, Tempus Fugit, a LW with as tight a chassis as you can get.



204 HP.  Pretty crappy, actually.  Good thing TF has that plow in front, because 204 HP will last only slightly longer than Frosty the Snowman on the Sun.

Next I checked Leprosy, a generic HW HS with a slightly larger chassis.



518 HP is better, but it's still not going to hold up long.  That's about as much HP as a razor tip.

For a middling-sized chassis, I tested Hunter-Killer Tank.



1350 HP is actually pretty decent for stock RA2.  It's more than any component, at least.

Finally, I tested the biggest chassis you're likely to use in a serious bot: Osiris and Epic.





Now that's pretty good.  You still don't want to leave your chassis overly exposed, but the difference is great enough that you might want to consider making room for your drive motors inside the chassis instead of stacking them on burst motors.

I think we can safely call this myth: CONFIRMED.



This also brings up another relevant myth, though: Is DS Aluminum REALLY stronger than steel?  Chassis HP could either be calculated by surface area or by weight.  If it's by weight, then steel will automatically be more durable than any type of aluminum.

This myth was easy to test.  I just gave my first two test bots steel armor, and re-checked their HP.





Remember, with DS Aluminum, the small chassis had 60 HP, and the big one had 4259 HP.  So it looks like this myth, too, is CONFIRMED.



Got any more RA2 myths that need testing?  Post them here!

[R0B0SH4RK]

What about burst enhancing wedges? Do the different burst motors/servos work better than the other ones?

[Urjak]

This is great! If I think of any myths I will post them.

[Madiaba]

Nice work, CB.  Now we know...

[Resetti's Replicas]

I've got one:
Can you break a robot in the test area?

[pyscolone]

easily.

[Urjak]

Yes, and the game crashes if you do.

[Somebody]

I destroyed Psyco Sweeper in the test arena once.

[System32]

Which wedges are better? Burst, Hinge or stationary?

[Naryar]

Weren't you the guy that asked for DIVERSITY?

It isn't just the type of wedge that is important. Wedge efficiency takes four things into account, only one being the type of wedge.

I've copied HazCon's wedges several times, but the copies are all inferior to the original.

[System32]

Yeah, but I suck with wedges, and I know all have similer results, but I want to know what makes a good wedge.

[WhamettNuht]

Would a servo be a good motor for a robot like Stinger, or can you use a normal motor?

[Naryar]

Yeah, but I suck with wedges, and I know all have similer results, but I want to know what makes a good wedge.

PROTIP:

You have plenty of wedges in the BBEANS4 AI pack.

[System32]

Another tip: Most of the bots crash my game.

[Reier]

What? Maybe you didn't install it right. Works fine for me...

[Naryar]

You possibly forgot to install Mad's smartzones.

[pyscolone]

IMO wedge effiency is based on many factors. Speed, angle of wedges, and weight distribution.

[Clickbeetle]

Time for another "episode".  This time, I'm investigating what makes a good wedge.  There are a lot of myths about wedges, and I'm pretty sure I didn't test all of them, but this should be useful information in any case.

First up: The myth that Emergency wedges are somehow worse than small wedges.  To test this, I made two identical bots, except one has small wedges and the other has Emergency wedges.  Since Emergency wedges are bigger and heavier than small wedges, though, I put the Emergency wedges on 20 cm extenders while the small wedges are on 60 cm extenders.  This makes the Emergency wedges about the same length and exactly the same weight as the small wedges.

Here are the bots:



To test which has the better wedge, I put them both in the Small Arena (to minimize the distance that one bot could drift off course) and drove them straight forward so they collided head-on, and recorded which bot got under first.

The small wedge got under the Emergency wedge for the first trial.



I'll spare you screenshots of the other trials, since they all look the same.

After 20 trials, the results are:  The small wedge got under 16/20 times or 80%, while the Emergency wedge got under 4/20 times or 20%.

I think it's safe to say that this myth is CONFIRMED.  Small wedges are better than Emergency wedges.



Next up: Weight distribution.  The myth is that if your bot's weight is distributed more to the front, close to the wedges, you'll have better wedges.  I tested this the same way as with the Emergency wedges.   I used the same bots, but one has 80 kgs worth of ballast in the front, while the other has 80 kgs of ballast in the back.



The results: After 40 trials, the front-weighted bot got under 24/40 times or 60%.  The rear-weighted bot got under 16/40 times or 40%.

I did more trials because I noticed that the bot in player slot 1 seemed to have an advantage for some reason (this didn't affect the Emergency wedge test because I alternated which spots the bots started in).  While 60-40 isn't a really significant difference, I'm inclined to call this one CONFIRMED because of the way the bots behaved in different spots.  When the front-weighted bot was in player slot 1, it got under the other bot 80% of the time, and there were almost no draws (where the bots bounce apart and neither gets under the other).  But when the rear-weighted bot was in player slot 1, it only got under 40% of the time, and at least half of the matches I did ended in a draw, probably more.  If I used a greater ballast, I predict the difference would be more pronounced.

So, long story short: Frontal weight distribution makes a better wedge=CONFIRMED.



The next myth I tested was ground clearance.  A lot of people will swear by the rubber wheels, saying they're superior to shiny hubs because they give you lower ground clearance.  Same testing method here: one bot with shiny hubs, the other with rubber wheels.  I added a 10 kg ballast in the shiny hub bot to make up for the extra weight of the rubber wheels.



The rubber wheel bot got under 22/40 times, or 55%.  The shiny hub bot got under 18/40 times, or 45%.

I'm inclined to call this one PLAUSIBLE.  There might be a slight advantage to wedges with lower ground clearance.  I don't think it's enough to justify the extra 5 kgs of the rubber wheels, though.  If you have 10 leftover kgs and nothing to do with it, you might try rubber wheels, but I wouldn't sacrifice anything for a maybe 10% advantage.



Going back to the Emergency wedge myth, let's suppose you have some extra weight.  Are Emergency wedges any good if you just replace small wedges, using the same extenders?

I used the same small wedge test bot as the first myth, and changed the other bot to this:



The results: Emergency wedges are still inferior.  The small wedge got under 14/20 times or 70%, while the Emergency wedge got under 6/20 times or 30%.



Last myth.  Looking at the results for the Emergency wedge, I thought: "Hmm... if the small wedge is narrower than the Emergency wedge and it's better, then is an even narrower wedge better than the small wedge?"  I used spike strip wedges to test this.



The results were overwhelming.  The small wedge got under the spike strips 20/20 times or 100%.  So, this myth is BUSTED.  The reason for the small wedge's effectiveness shall remain a mystery...



To summarize: Factors that make a good wedge include using small wedges rather than some other component, distributing your weight close to the wedges, and possibly having low ground clearance.

This means that using heavier burst motors (DDT's instead of Snapper II's) will also improve your wedges, as more weight will be distributed over the wedges.

[Sage]

do the test with different burst motors.

but really cool.

[Urjak]

Wow.... That is cool.