SakuraBot v1.0

Difficulty Beginner

Focus Punish Awareness and Neutral Game Spacing

The intention of this bot is to help train your mind to identify similar looking situations separately, and do the appropriate punish for each situation without just using a default. It also trains the same sort of tendency in the 'neutral' game.

SakuraBot can be considered therefore to be a harder version of KenBot, with only slightly more difficult inputs, for most people. She will not push as much in neutral and will not jump in as often or as predictably, but is still intended to help train reactions, as well as helping players to get used to abilities of new characters they have available before players move on to other bots or Katas. In order to take maximum advantage of the free week of Sakura offered with the release of SFV:AE, this bot's more complex potential has intentionally NOT been explored. It is therefore possible that SakuraBot 2.0 will be available in the near future for those who do not consider this version of the bot to help as much against any new, relatively easy, Sakura technology.

This bot's inputs have quite a lot of leniency but the bot will change relatively heavily as you go from lenient to optimal inputs so be sure to keep this in mind when playing it. We recommend starting with lenient Acceptable inputs and occasionally attempting to re-enter them between sessions or when you feel that the situation is not training you in the aspect of the matchup that you really want. Due to this, the variant data and annotations will be more detailed than usual for an Easy level bot. To save yourself time if you just want to get right to it, don't bother too much with going into the details on those at first, just read the section for Action #3.

Gauge Settings

As with all bots, 1P Health should be set to "Normal". To practice the easy versions, leave 1P CA Gauge to "Auto-Recover", and 1P V-Gauge to also "Auto-Recover".

These settings may change based on your goals, explained in the Difficulty section below.

Settings for 2P are V-Trigger - II, Health - "Auto-Recover", V-Gauge - "Normal", CA Gauge - "Auto Recover".


Completing each action gives 1 point. Try to get to 10 points without dying, and before Sakura gets to 10 points.

Points System


Difficulty Levels

Action #1

Record from very close.


4/5MP (hold this MP to the end. Make no effort to make this FirstFrame, but it is ok if it happens), 4HP (must combo), quick hidden [1HK 1MK], end recording instantly and make sure not holding MK.


4/5MP (hold this MP to the end. Make no effort to make this FirstFrame, but it is ok if it happens), 4HP (must combo), quick hidden [1HK 1MK], end recording.

This input when replayed should mostly repeat one of the moves in it over and over, rather than using any specific sequence, when far away. Optimally, it does a lot of 4HP and nearly no 1HK. If you want this bot to do a lot more of Sakura's V-Skill in neutral, don't hold the MP to the end of this one at all.

Action #2


From far away, 5LK, hidden [623MP (hold this MP, input must be clean or at worst 6323)], end recording instantly


From far away, 5LK, attempt hidden [623MP (hold this MP, input just needs to not cause a Hadoken in neutral when replayed)] end recording

For those who are familiar with buffering, the perfect input is supposed to buffer the 623MP motion for Shouoken within the LK so that it will cancel if the LK hits and not happen if it did not. This is probably one of the largest requirements in terms of getting it right, to improve the bot when trying to deal with its higher levels, but is not so necessary at the lower levels. It does however make the bot a bit 'too easy' at the lower levels so if you find that beating those is simple due to lots of chances to punish her. This input really requires the MP to be held while ending it or she will do even more V-Skill, which is only good for when you are getting used to that move and staying relatively far away from her. Change it as soon as possible if this was not held (use Action #4 along with this one to test).

Action #3

Input this one from close enough to hit. Preferably minimum range.


1LK.1LP+LK, hidden [6MP (hold until end), 6MK], end recording when opponent character hits the ground from the throw


1LP+LK, hidden [6MP (hold until end), 6MK], end recording when opponent character hits the ground from the throw

Since this is a back-throw input, there is only a relatively small window to input the MP and MK in order to have the side switch not happen, but it should generally work without the player needing to be unreasonably fast. The 6MK 'overhead' should mostly be what you see when this is replayed. If your plink is incorrect in the Perfect version you will mostly notice it during her V-Skill. This input generally requires the MP to be held and will be detrimental to the bot with no benefit at all, if it is not, unlike Action #1 and #2 where at least it is useful a bit. Try to get this one right before even starting to use the bot. This particular input should be 'redone' the most often, even if just to test less perfect versions, as it has the most effect on the bot's rhythm and can teach interesting things when 'incorrect' in various ways.

Action #4


FirstFrame 1MP+MK, 9, 66 at peak of V-Skill 'jump' arc, end recording quickly


FirstFrame MP.1MP+MK, 9, 66 at peak of V-Skill 'jump' arc, end recording quickly

This input requires the MP to be firstframe one way or another, but does not actually absolutely require the 1 input to be. The goal is to get a crouching MK when another action with a held MP flows into this one, and therefore give the player the option to control how often Sakura does her V-Skill and which moves go into that V-Skill based on which of the inputs they don't hold the MP in (we suggest Action #1 or #2 for this.

Action #5


236MP+HP (hold the MP until end), hidden [mash 5HP+HK (x2 minimum presses)] 214 end


236MP+LP (hold the MP until end), hidden [mash 5HP+HK] 214 end

In the case of the Perfect Version, the aim is to do it all quite quickly, so that the flows can chain into other things, but the EX Hadoken's state on the screen helps to 'control' what happens. This action should sometimes be able to combine with Action #2 to give a L-Shunpuukyaku, and with Action #4 to give M-Shunpuukyaku.


Still in progress.