Amped 2/Game Mechanics

Amped_2

Score (bonuses, score decay, ...)

The score of an individual trick is determined by trick type, landing, stance (switch or normal), style, media bonus or not and finally by the score decay.

Score decay

The "spirit of the game" is to promote trick variation. As a general rule, every trick has a default value it gives you the first time. The next time you perform the same trick, the new score is divided by two (this can vary a bit though, depending on the trick) and so on. However, the game mechanic that governs this is full of exceptions and does not seem to be fully consistent. Here are some observations:

• Off-axis spins are immune to score decay.
• Double (or triple) flips are not subject to score decay and don't impact the score for single flips.
• Several kick-flips of one type impact subsequent kick-flips of the same type according to the general "50%" rule.
• There are several cases of inter-dependencies between tricks:

- Spins impact each other in a fairly complicated way. For example performing a standard 720 gives 5.000 points. If you have done a 180 before, you only get 4.800 for the 720. If you have done a 360 before, it goes down to 4.000. The exact details how these inter-dependencies are likely never going to be important for a speedrun and haven't been investigated further.

- Just like the spins, the kick-flips have a complicated inter-dependency. The default value is 2.000. If you then do a different kick-flip trick, you only get 1.600 per kick-flip. The score per kick-flip occasion continues down in a non-linear fashion that will likely not of importance in a speedrun.

- Every off-axis jump impacts future spins and flips (corkscrews only impact forward flips and rodeos only impact backward flips). E.g. the default score for a flip is 2.500. If you do two off-axis jumps and then a flip, you will get 625 for the latter.

- Each grab type along with the three possible tweaked grabs of the same type (for example stalefish, tweaked stale, bs stale and stale method) impact each other (with roughly the 50% decay, even though it's difficult to measure exactly).

• Single spins impact future spins regardless of direction (clockwise or counter-clockwise).
• There is no score decay on any of the sponsor challenges.
• According to http://www.ign.com/wikis/amped-2/Guide_part_3, style is affected by score decay. It's unknown to what degree though due to the difficulties of performing the same trick with the same style twice. It doesn't appear to be a huge decay effect though.
  
  

Bonuses

There are a few ways to boost (or deteriorate) a trick score:

• Media coverage gives +25%. If done during a combo, it's only the part that is caught on camera that will get the media bonus.
• Doing a trick from switch stance gives +20% (one can also find sources on the internet saying +30%, but that is incorrect). This is also true for butters, jibs and rail grinding, but doesn't seem to impact grabs or style points for any trick type.
• There are four types of landing and they impact the trick score (if done as a part of a combo, it's only the jump score leading up to the landing that is impacted):
  

- "Sketchy", -50%

- [Default], no bonus

- "Sick", +15%

- "Stomped", +30%

Bonuses are cumulable.

Butters

Butter scoring is non-linear. The below graph illustrates the score rate.

It hasn't been studied here if the score rate plans out or continues to increase exponentially (?). However, due to the low score rate, butters are generally only used for short distances, which should be covered by this illustration.

Jibs

There are six jib tricks (on snowboard). They can be divided up in three groups based on their score rate. Tricks within the same group have identical score rates. The first group consists of Nose stall, 5-0 Stall and Indy Nose Pick. The second group has Disaster and Blunt. Hand plant is the only member of the third group. Their respective scoring rates are shown below.

Common for all groups is that they start out with an increasing score rate, but then hits a threshold beyond which the score only progresses very slowly in a linear fashion. One can also see that Hand plant is by far the best trick. There is in fact never a reason to use anything else in a speedrun. On top of having the best score rate, the way a Hand plant is performed results in a combo trick, which makes Hand plants even more outstanding compared to the other jib tricks.

As a final remark, the snowskate adds a couple of jib possibilities, but the game never really gives you a reason to use them so they haven't been studied.

Grabs

There are two groups of grabs based on score rate. The first group is stalefish and melon. The second group consists of nose, mute, indy and tail. The score rate of the grabs within each group appears to be identical. The tweaked versions of the respective grabs follow the same grouping. The four different possible score rates are shown in the illustration below.

It's worth noting that the style points are identical for the two tweaked grab groups and the same goes for the two non-tweaked grab groups. One can also see that there is some spread for the base points of the four different groups with the stalefish group giving more points than the nose group and the tweaked stalefish group yielding more points than the tweaked nose group. However, due to style points quickly dominating the base points, the difference is less noticeable when looking at the total score.

As a final remark, the snowskate adds a couple of grab possibilities, but the game never really gives you a reason to use them so they haven't been studied.

Grinding

Of all the trick types, grinding is the most complex to describe and there is likely important details about the mechanics that are not described here. As a first step, one can arrange the different possibilities into four groups. 1. standard 2. edge 3. standard spinning 4. edge spinning. The score rate for these four groups are shown below. For comparison, a switch of groups 2 and 4 have been added, since they show some variations to the expected behavior.

Groups 1 and 3 yield less points and are only shown for comparison, but haven't been studied further.
The base scores all reach a max at roughly the same time, but their respective max scores vary significantly with group 2 well above the other groups. It's actually a bit misleading to talk about a "group 2", because the max base score turns out to be much higher for the Tweaked BS Tailslide (the one used in the graph) than the other members of that group. While the other edge grinds are better than standard grinds, there is no reason to use anything other than the Tweaked BS Tailslide when it comes to non-spinning grinds. The switch points are roughly as expected (so ~20% higher than with regular stance).
If we look at the rate of style points, the spinning grinds seem to never stop increasing, while the non-rotating grinds reach a shoulder, beyond which the style rate progression is much slower. It's also interesting to compare the switch grinds with their respective regular grinds. The style points for switch Tweaked BS Tailslide start out far behind the regular trick, but eventually catches up and goes above. The style points for spinning grinds have a complex pattern, but regular and switch seem to mostly follow each other, with the exception of around 30-40k, where the regular spinning has the advantage for a short a while.
Because of the differences in the base and style scoring for the different groups, switched non-spinning edge grinding (=Tweaked BS Tailslide as described above) is the best choice in the window 80-120k, but then stagnates and is surpassed by the spinning grinds. While it can be hard to see in the graph, the spinning grinds are preferable at low scores (roughly a 25% advantage until 20k, after which the relative difference starts to reduce again).

When it comes to deciding between the type of grind that should be used in a given situation, it can also be mentioned that there are two disadvantages with non-spin grinds. The first is that you often come to unintentionally change the trick while simply trying to keep the balance. When you change trick type, you will also be subject to the new trick's limitations (meaning the score can suddenly plunge if you're above the max base score for the new trick). The other disadvantage is that it can in many situations be difficult to reliably know if you'll end up grinding regular or switch (for example if you start the grind by a few tricks to get the combo multiplier up). In that case, it's safer to go for a spin grind, since the difference between regular and switch is smaller. A possible third (non-verified) disadvantage with non-spin grinds is that they appear to be slightly more difficult to keep the balance on.

Even though not directly related to the scoring rates, it can be worth to leave a quick comment about the balance meter here. As anyone will notice, the balance will be harder and harder to control the longer you grind. However, if you jump to another rail (or any other grindable structure), the "volatility" of the balance meter will reset.

As a final remark, the snowskate adds a couple of grind possibilities, but the game doesn't give you many reasons to use them so they haven't been studied (none of them seem to result in huge score boosts though). The only challenge this could potentially play more than a very minor role would be in Laax legend.

Various observations about the scoring for the other tricks

• All kick-flips give the same amount of points. Same goes for spins, flips and off-axis spins (= regardless of direction of the trick).
  
  

Speed considerations

Movement speed

• The surface impacts the speed. Due to how accurate tests would be quite time-consuming, only rough estimations have been made. The deeper the snow is, the slower the movement will be. Patches of dirt also slow down your boarder. On the other end, newly prepared slopes are faster and ice is the fastest surface. There is maybe a 10-15% range within the movement speed because of the surface. This is enough to warrant going out of the way in certain spots to spend as much time as possible on the fastest surfaces and most certainly to jump over for example patches of dirt, when possible.
• Buttering and grinding are done at 75% of the normal movement speed (regardless of tweaked or not).
• The stance has no impact on the speed. Switching stance has normally no impact either, but in practice often leads to a small time loss.
• Ollie jumps keep the speed intact.
• Even though the animation is different if you press the A-button or push forward on the left stick, the acceleration and max speed appear to be the same (at least within measuring error margin). Combining the two doesn't appear to speed up the character either.
  

Various speed tricks and exploits

• By pressing select, the screen goes black for ~3 seconds and the character is teleported out on the slope. This mechanic was probably put in place to avoid the character being stuck somewhere. Despite the long reset time, the game will sometimes teleport you quite far away and this can be used to get down the slopes faster, as well as combining "own the mountain" targets that would otherwise have been impossible to combine.
• If approaching a "jibable" area at too high speed, trying to jib will result in a "jib jump". You will briefly enter the jib animation, but then get bounced off at high speed. This doesn't save time over normal movement, but is useful for connecting with the end area of gaps that lead to rails (faster than starting to grind and then jump off the rail).
• You can go "out of bounds" in several places. Typically by entering seams in mountain walls or buildings. The game always reacts in a similar way to the select trick above. So the screen goes black for a few seconds and you're teleported out onto the slope again. There doesn't seem to be any documented cases of any extreme cases of this behavior, but at least once on Mt Buller, the character was teleported 10-15 seconds away after having entered a house seam (ktwo, not recorded). So there is at least potential for further findings.
• When traveling at low speeds, you can gain speed by going over small bumps or ramps. The speed increase of the downward part of the bump is generally bigger than the speed loss going up the same.
• As a rule of thumb, you'll lose 1-2 seconds going over a jump instead of around it. The details obviously depend on the dimensions of the jump, but in general it's considerably faster to go around than over.
• Many areas of the game have "imperfect" physics that can be exploited:

- Half-pipes seem to be programmed in a different way to the surrounding slopes. It's generally easy to scale half-pipes by zig-zagging.

- Stretches with a half-pipe like layout (so surrounded by elevated edges) can sometimes be climbed by using zig-zagging and gaining speed on the edges.

- Stairs have sometimes sketchy physics and can be scalable (e.g. Mt Buller a little below the Town Square).

• Some mountains have bent goal lines. This means that the fastest way to end a run isn't necessarily to go straight ahead. Since there is no way to actually see the goal lines, this will have to be tested individually. Below are comments about the individual mountains. The reference point for the directions is for someone looking down the mountain. The same comments apply for both mountain levels, where the same mountain is used for two levels:

- On Millicent, the ending zone reaches down into the forest on the right side.

- The ending zone of Breckenridge makes a sharp turn on the right side, making that side faster to end most runs on.

- On Mt Hood, there is an ending zone on the left side that is much closer than the one on the right side.

- Bear Mtn., Laax and New Zealand haven't been studied in detail, but no particular way other than "straight ahead" seems to be preferable.

- There are three principal paths to the end of Mt. Buller. Starting from the top and with no other constraints, the one in the middle is the fastest. However, most runs will direct in either direction, so the other ending areas will also be used. To be tested in each individual case.

Various unknowns

• Is there a pattern to sitting starts? (or "fast starts"?)
• Is there a way for any given situation to predict if you'll end up riding regular or switch on a rail?