World Beyblade Organization by Fighting Spirits Inc. Beyblade Discussion Beyblade Customizations

Testing Vectors

Jul. 18, 2012  5:11 AM
 
I've been talking to a few users about the idea of having "testing vectors" for use when creating customs, and I wanted to post here to see what everyone thinks of the idea and of what would be good candidates for vectors.

Testing vectors are standard customs that everyone has an idea of how well they work. They are used when testing new or random custom ideas to get an idea of how good a new custom really is. By having a standard list of testing vectors, answering the question "What should I be testing my customs against?" would be very easy; if your custom can beat the appropriate testing vector, then it is worth looking into.

A good vector custom should be good, but not great, at whatever it's designed to do. Let's take a stamina vector as an example: wheels like Duo would be bad because they're oppressively effective, but wheels like Death would be great since they have good Stamina without being broken.

The idea is that the testing vector should be good enough to beat customs that aren't really that great, but not so good that they're unbeatable. If your custom actually beats a given testing vector, then you know your custom has potential, and then you should test it against the top-tier.


I personally have a few vectors I test with when I'm coming up with random custom ideas. These are by no means ideal testing vectors, they're just what I happen to use:

Stamina:
Death ____ AD145WD

Defense:
MF-H Death ____ BD145CS (This is technically high tier, but it's more beatable than most and I am having trouble thinking of something better)

Attack:
MF-H Beat Lynx GB145RF

To give some examples, I always test spin equalizers and new stamina customs against Death ____ (I use Pegasus II, personally, since I have an exceptionally heavy one) AD145WD; if they can beat Death, then they have potential and I test them against higher-tier stuff. If I'm testing an attacker, I first test against Death _____ AD145WD to make sure it's competent, and then I test against MF-H Death _____ BD145CS to see if it can KO that. If I'm testing a defense combo, I see if I can KO it with Beat; if I can, then it isn't up to par.


If you think this idea does or does not have merit, post your opinion! If you think this is dumb, that's totally fine (but please do explain why)! If you think it's a good idea, give your feedback as to what you think would be good vectors and we'll see if we can make a standardized list.


Important points from later posts:

(Jul. 18, 2012  5:41 AM)Ingulit Wrote: The main criticism I'm expecting for this idea is that it might end up becoming something of a mid-tier list, which is not at all what I'm proposing; rather, I'm hoping that we come up with one, maybe two customs for each type of Beyblade that are "standard" to test against.

I want to doubly emphasize that there would only be one testing vector for each type, so rather than someone having to look at the top tier list and think "Hmm, should I test against Duo Cancer AD145SD or 230D or yada yada," they would have a very definite "This is a standard thing to test against" to work with.
Jul. 18, 2012  5:31 AM
 
Hm, why would you not just test against whatever is currently top-tier though ? I know you elaborated on it in your post, but if it does not do much against the current top-tier combinations no matter at what point you are in the timeline, then it is probably just not worth it at all. For instance, Duo is winning a lot more tournaments than Death. I understand that certain parts are "broken", as you wrote, but even if they have a higher win percentage than a given custom, you just always need to compare with what is currently considered to be the best Beyblade against it, just like with what we did with Gryph.

I think nobody here is going to discourage anyone from testing anything and everything if they can and want to, but if time and resources are limited, then in my opinion, they should maximise their effort by only playing against top-tier customizations. However, in cases like for Beywiki articles, of course Beyblade parts should be tested as much as possible, and while it might not be worth a thread to post a combination that does well against only "testing vectors", I do have an idea that will probably at least help slightly.


Anyway, did I miss something : why Pegasis II ?
Jul. 18, 2012  5:41 AM
 
(Jul. 18, 2012  5:31 AM)Kai-V Wrote: Hm, why would you not just test against whatever is currently top-tier though ? I know you elaborated on it in your post, but if it does not do much against the current top-tier combinations no matter at what point you are in the timeline, then it is probably just not worth it at all. For instance, Duo is winning a lot more tournaments than Death. I understand that certain parts are "broken", as you wrote, but even if they have a higher win percentage than a given custom, you just always need to compare with what is currently considered to be the best Beyblade against it, just like with what we did with Gryph.

I think nobody here is going to discourage anyone from testing anything and everything if they can and want to, but if time and resources are limited, then in my opinion, they should maximise their effort by only playing against top-tier customizations. However, in cases like for Beywiki articles, of course Beyblade parts should be tested as much as possible, and while it might not be worth a thread to post a combination that does well against only "testing vectors", I do have an idea that will probably at least help slightly.


Anyway, did I miss something : why Pegasis II ?

Your first paragraph was the main response I expected about this idea, and why I didn't post this until now. The reasoning behind not testing against high tiers first is to see if an idea you have has potential or not by testing against a custom that A) everyone knows about, so it's easy to ask for recommendations, and B) is beatable, so you can see if an idea has merit. The advantage to this comes when you're trying out a really far-out idea based on a single part and you're trying to come up with the best parts to support the main piece. By testing against testing vectors, you can get a faster idea of what works and what doesn't without everything just straight-up failing because the thing you tested against was broken.

What you said next is very true; one should absolutely not post a thread if the custom only beats testing vectors. Like I said earlier, once a combo works against a testing vector you should then move on to high tiers, and if it doesn't work then then you go back to the drawing board (or realize you've struck gold if it wins).

On Pegasus II, sorry, I keep forgetting to mention this: I have a very heavy Pegasus II (3.4 grams, my second heaviest clear wheel) that gives me good results in both defense and stamina customs. Maybe putting ____ instead of Pegasus II would have been better... I'll change that now.


EDIT: The main criticism I'm expecting for this idea is that it might end up becoming something of a mid-tier list, which is not at all what I'm proposing; rather, I'm hoping that we come up with one, maybe two customs for each type of Beyblade that are "standard" to test against.

EDIT 2: Again, I won't be butthurt or anything if you guys don't think this is a great idea, it's just something I do that someone recommended I put out there to see if it can be useful.

EDIT 3: I want to doubly emphasize that there would only be one testing vector for each type, so rather than someone having to look at the top tier list and think "Hmm, should I test against Duo Cancer AD145SD or 230D or yada yada," they would have a very definite "This is a standard thing to test against" to work with.
Jul. 18, 2012  6:03 AM
 
I just want to post my support for this idea, it's excellent for initial testing of concepts and so on, especially in plastics, before you decide to commit the time to thorough testing.

I'm out for lunch at the moment, so I'll add more later, but yeah, wanted to express some support, haha.
Jul. 18, 2012  6:07 AM
 
(Jul. 18, 2012  6:03 AM)th!nk Wrote: I just want to post my support for this idea, it's excellent for initial testing of concepts and so on, especially in plastics, before you decide to commit the time to thorough testing.

I'm out for lunch at the moment, so I'll add more later, but yeah, wanted to express some support, haha.

I remember you telling me this was good for plastics and it's why I kept up with the idea. I was talking to someone else about how I test stuff and they said I should post it, so here we are!

The out for lunch thing caught me off guard since it's 12:06 AM here! I need to remember how international this community is, haha!
Jul. 18, 2012  7:26 AM
 
It's also winter here, though the sky is clear and it's about 15°C. Good excuse to wear jackets.

Anyway, now I'm back at the keyboard, I shall ramble appropriately. It's hard to get these thoughts together cohesively on the spot, and as I'm sure you all know, most of my testing time is focussed on plastics, and while there is a lot of crossover, doing certain things in MFB is harder and different approaches work better.

Now, I do take a slightly different approach to testing vectors, though most of my testing is done with plastics, so it's also easier to do there. I tend to use a single combination, generally a balance type, to sort the good from anything below that level, when testing combination ideas. The typed vectors Ingulit discusses here are more appropriate for testing more solidified concepts (and concepts is a key word - once you get to specific combinations and the formal testing, you must use competitive combinations for obvious reasons. This is basically for seeing if your ideas could work), whereas the balanced vector thing I tend to do depends less on having that direction (i.e. you're not sure what types you really need to test it against/you just want to see how it acts in battle).
Ingulit's is generally more appropriate for MFB where typing is much more defined and unified, whilst the approach I take is arguably more useful for the broader metagame and eccentricities of plastics. Not that they don't work for others, either, in fact I often use exactly the sort of thing ingulit has discussed here after I've done the initial checks - and a more balance-styled approach would also work for MFB too.

I guess I should better explain what I mean with this balance-styled stuff. I wish I could come up with a more appropriate name, but you'll just have to live with this for the time being.
Basically, the main criteria I look for are
-Balanced Performance in all Attributes - definitely easier to achieve in Plastics, though even there, I tend to focus less on attack for vectors, simply because it isn't worth the other tradeoffs.

-Consistent Performance - the vector should not be prone to wild behaviour variations (think metal change tips and some CS's), and provide the same behaviour round-after-round.

-Stable, Mostly Passive Movement - The requirement for consistency also means no fast moving tips, as how they connect with opponents can alter results dramatically. That said, it is still good if it can go on the aggressive, but to do so should require significant alteration of launch. Ideally, though, it should still be able to KO opponents that have extremely poor defense (B: D for example) that wouldn't be reliable at a tournament as a result.

-Consistent/Understandable Reactions: the vector should allow you to see the effect the beyblade you are using against it is having. It should respond to hits in a consistent manner, for example.

-Familiarity - while testing vectors generally wouldn't be used for formal testing, for their indicative use and when choosing what qualifies as a good vector, the combination should be something most bladers will understand the behaviour of. This is much harder for MFB, but there's the possibility of something being endorsed as a good vector making people become aware of it. Basically, while helpful, it's not as important as the others.

In a word, these combinations would be "plain". Nothing fancy, nothing exciting, but easily understood and always the same.

Now, I don't really have much to offer in the way of a good example for MFB, I was messing around with something along those lines when my launcher broke and I haven't found something I'm 100% satisfied with yet, but for those that understand, here's what I use as my primary testing vector for plastics:

AR: Tiger Defenser
WD: 10 Heavy or 10 Balance
SG: Neo Right SG (Heavy Metal Core)
BB: SG Semi Flat

It's a very generic compact. It's not too powerful, though, SG Semi Flat isn't a particularly great base, but it's perfect as a vector - extremely consistent in behaviour, and very equal in all aspects without excelling in any. Metal Ball Base with four metal balls could be used too, though I tend not to due to its more angular shape. Tiger Defenser is good all around, has a fairly regular shape and no gimmicks. The use of 10 Heavy and and an HMC is generally just because it needs that defense to be balanced. Unlike more defensive compacts, it moves when hit and does so in a way that it's easy to judge how hard the hit was. It would do decently in a tournament against the less-thought-out opponents you see there, but doesn't do so well against more competitive opponents.

Now, it is perhaps slightly defensively biased (not that much, SG Semi-Flat's defense isn't great), but getting a perfect balance is hard, and in this case the consistent performance, and the level at which is performs, is well worth it anyway (plus there really isn't anything that works that much better).

It's a lot easier to do in plastics where consistent balance types are a dime a dozen, but something similar could definitely be done with MFB.