Tweaking is a bit hard to define precisely...

When you do a grab, you earn x points during each frame of animation. And x increases each second. So after 1 second, a Nosegrab is earning 50 points per frame (ppf), but after 2 seconds it earns 60 ppf. After 3 seconds it's 70 ppf, after 4 it's 80 ppf, etc.

I think that tweaking just changes the ppf of the grab, but I'm not sure exactly how. I have 2 theories:
1. Each grab has a normal ppf and a tweaked ppf. Tweaking just looks up a number from a different table.
2. Tweak modifies the normal ppf. It uses the same base ppf but then multiplies it by a "tweak constant" (which is the same for all tricks).

I think 2 is more likely, but I can't say with certainty yet. I just haven't collected enough data yet to figure it out.

I do know though that tweaking is not just adding a flat point bonus to your score. If you hold tweak for the whole grab, you'll get a huge bonus. If you just tap tweak early in the grab, but then let go, you'll get very little bonus. If you wait until late in the grab to tweak, you'll get very little bonus. This is why I'm pretty sure that tweaking changes the ppf you earn only while you're holding tweak & the rest of the time the grab is scored like normal. (Along with the fact that grabs can't be tweaked until after they're held long enough, this makes it hard to collect consistent data. Every score I record is a mix of a tweaked & an un-tweaked grab; I'll need a ton of data to de-couple the two grabs.)


TLDR:
Long story short, though, tweaking often appears to roughly triple a score. Any tweaked super uber is better than a sig trick.


The sig trick does have one advantage, though. The animation continues & you keep earning points (until it reaches a natural stop) even after you let go of the grab button. But even if the animation has not reached a natural stop, as long as you let go of the button early enough, you still land safely. So it's often possible to effectively 'hold' the grab until the actual very last frame before landing. I doubt this comes close to making up the score difference, though.

So Signature tricks do better tricks points now?

_________________
Well now, let's see if we can't do something about my boredom.

Skype-XEShadowX

It says in big, green text that it doesn't.

_________________
"There's absolutely nothing they can do to change me from being right in the middle" - MurkaDurkah

Sig tricks have the highest base score of any trick, but they can't be tweaked. The bonus points for tweaking make all other super ubers better than a sig trick.

That's a very major oversight. Who in his right mind doesn't hold down the tweak button 99.99% of the time?

_________________
Give yourself a high five!
[hipster] I liked Fall From Grace and Burnout BEFORE they lost the O2 requirement! [/hipster]
https://ssx.ozmonet.com/user/190015544

Thanks mahkra!

Mahkra, can I just say that this is an amazing piece of work?

I haven't had time to go through this whole thread yet, but wanted to let you know that I haven't ignored/forgotten your request for me to verify some of your deductions.

I will. As soon as I have enough time to give it the full attention it deserves for an obviously huge and thorough job researching everything.

Funny story, a few days after the demo hit and this thread started I forwarded it to a few of our gameplay engineers. We were all thoroughly impressed. Seriously blows me away.

So again.

_________________

Thanks Todd!

I didn't really expect you to be able to confirm the details of the scoring algorithms, but I'm glad to know you did see my post in the other thread.

Even more than the specific numbers, I'd actually love to hear about how you reached the final scoring model, if you can speak to that at all. I love games of all types -- board games, card games, video games, sports, etc -- and I'm a huge math geek. I think it's fascinating how a basic set of rules ultimately defines complex gameplay strategies, and how tweaking the basic rules will reward / encourage different types of play.

I tried asking "Tots" in the Thank you Merqury City! thread a few weeks ago, but this is probably better directed to you, now that I've got your ear:



... ... ... [/blush]


ps - You wouldn't happen to have any leads on game design jobs in the Chicago area, would you? (I know this is a total shot in the dark, but no harm in asking, right?) I'm moving there with my wife & kids in a couple months. Trying to get a job teaching HS math right now, but what I'd really love to do professionally is work on games.

Seeing as no one noticed when I posted this. It was before he had even put the green lettering saying it doesn't give anymore points that a tweaked grab.

_________________
Well now, let's see if we can't do something about my boredom.

Skype-XEShadowX

I think that bit was bold originally, then I went back and changed the color instead because it didn't really stand out. Hard to recall such a tiny detail though as that post was a full week ago...

Wow, how time flies. Including the demo, we've been playing this game for a month already! Certainly doesn't feel like it's been a month.

We dissect every game's mechanics, Todd, every time. Mahkra is the latest and one of the best. This is why I was confident we would blow away any scores that came out pre-release.

_________________

I managed to do some testing of the regular ubers (and even had time to set a new personal best on the Beyond the Fall GE -- 31,189,392 ) while my daughter was napping today. In short, each level of grab (basic / better / best) is exactly halfway between the normal version of that grab and the super uber version of that grab. 'better' ubers are about equal in score to 'basic' super ubers. and 'best' ubers are almost as good as 'better' super ubers.

If anyone's interested in exact numbers, I've added those to the first post in this thread.

Mahkra, one thing I noticed you assumed incorrectly: grinds are scored on time rather than distance. It probably would have made more sense to be scored on distance, I think it got set for time b/c air tricks are scored that way and we just never got around to changing it since it doesn't really make a huge difference either way.

Anyways, this means if you're trying to max your score you should never boost while on rails. The faster you go, the shorter time you will be on the rail, the less points you will score.

One other thing, someone noted that to score max you should avoid rails completely. That may be true for the super eilte player, but a neat trick for building your multiplier up quickly is to get into Tricky/Super Tricky and then just cycle between a few grind tricks for @1s each. Doing this on the long first rail at the top of Serenity for example, can get you from about 8-9x (what I usually enter at) to 20x by the end pretty much risk-free.

Almost everything else looks spot-on. Nice!

_________________

well il be damned... i've been doing Sig tricks every other trick or so on trick runs... just assumed it scored more lol.

thank god for this thread... might finally diamond Serenity... one day

_________________

Hmmm quite interesting that you scored rails by time. No wonder the badges are for grinding for X seconds rather than Y metres.

An even neater trick is jumping off the rail to do a couple of flips and spins with a grab and landing back on. While it nets the same +1 to multiplier, it probably scores a lot more points but there's an odd chance that you won't magnetise back onto the rail. Food for thought when building multiplier.

Hehe... so was I right or was I wrong?

Thanks for taking the time to look this over!

To further Batty's point that rails can be good, I think that also depends on the track and whether or not it is really hard to do tricks. Such as Africa - Corkscrew. Everyone knows that massive spiraling rail right? If you do a trick down the drop you get ~300K with 20X flow. If you grind the rail as slow as possible, you get ~ 1-1.5million because of how long the rail is.

By the way, Award for most involved and greatest video game director goes to Todd Batty. Someone get the man a medal! Gondee?

_________________

I've... Been... Doing... It... All... Wrong...

_________________

I haven't read the entire thread, only skimmed through and I'm sure these have been asked already but here goes,

Is there any reason why Blackbird and All you can eat were given a higher scoring rate than the rest of the tweaked double tap ubers?
Why aren't all double tap uber tricks worth the same amount of points? Same goes for the other types of tricks.

In my opinion, every trick of it's type should be worth the same amount of points. For example, all single button grabs should be worth the same as each other, then all double tap grabs should be worth the same as each other, etc etc. I also think the sig trick should be worth the same amount as a tweaked double tap uber.

It's a bit tedious doing the same 2 tricks over and over, just because you know they give you the most points. I'd like to be able to mix it up but what's the point when you need the highest score possible? It doesn't make any sense to me as it is.

Like mentioned above I'd love double taps to be worth the same just for mixing it up. While we're at it, I'd like spins and flips to be the same as well for variation. (Nothing to do with me using new controls and not being able to get spins out of my tricks completely. Really. )

I don't have any insider info from the devs, but the difference seems to be that Blackbird / AYCE are "harder" -- you have to tap 2 buttons that are opposite each other instead of tapping 2 buttons that are adjacent to each other.

If it were up to me, I'd add two more "best" tricks (we have left -> right and right -> left; let's add top -> bottom and bottom -> top) to fill out the system much more elegantly.

Then we'd have, at each level of Tricky:
1. 4 easy tricks
2. 4 medium tricks
3. 4 hard tricks
4. 1 signature uber (only in Super Tricky)

This could work in Standard or in Classic controls:

Standard:
1. single-button inputs (4 possible)
2. adjacent 2-button inputs (4 possible)
3. opposite 2-button inputs (4 possible)
4. L2/LT + any button (4 possible, so there's some redundancy)

Classic:
1. one shoulder button (4 possible)
2. two shoulder buttons (6 possible, so there's some redundancy)
3. three shoulder buttons (4 possible)
4. four shoulder buttons (1 possible)

This would involve scrapping the "first button selects the hand, second button selects the side of the board" idea, but does that even matter?

mahkra. Does this verify your calculations on how many points you get for spins and flips?

This is the name of the trick before I bank it in. Sorry, didn't have enough patience to do 1 million degrees

This shows the points I got.

_________________

Emu, we just discussed this in chat, but I'll note it here for anyone else who's interested.

Just looking at the flip / spin (because you said you only held the grab for a second or two), I'd expect your trick to earn a bit over 510 million (base trick score). Your combo shows 521 million after the trick, which seems reasonable. (I don't know how many points were in the combo before this trick, but around 10 million is totally possible.)

The 2.1 billion is NOT 510 million x20, though. It looks like you're maxing out a variable used for an intermediate step. A 4-byte integer will only have enough space allocated to store values from -2,147,483,648 through 2,147,483,647. So if the "new points to add to the bank" variable is a 4-byte (signed, because "no flow" penalties end up negative, so it can't be unsigned) int, then the most points you can ever bank at one time will be ~2.1 billion.

Bump! Wanted to thank you guys, especially Mahkra for this thread. Alot of interesting theory here that I' definitely gonna put to use. Thanks alot again!