SF6Statsv0.20.15
Preparing fighter pageOpening fighter statsChecking saved stats first. New fighters take a little longer to prepare.

Saved match plans

Previously generated AI plans, with their original supporting stats and data dates. Expand a plan to read both perspectives.

Dhalsim · Latest 100

What to improve

  1. Find the cause of punish events

    Punish Counters landed: profile 0.5 · reference 0.14

    Review five Punish Counters received and five landed. Separate whiffed attacks from unsafe actions on block before choosing what to practice.

    Try this

    Recreate one repeated situation in training. Check the spacing and frame meter, then practice a safer choice or a punish that actually connects.

    Supporting stats
    • Punish Counters landed: profile 0.5 · reference 0.14
    • Got Punish Countered: profile 0.4 · reference 0.1

  2. Make Drive Impact a deliberate choice

    Drive Impact: profile 0.8 · reference 0.23

    Review five Drive Impacts and record the situation and outcome. Keep effective uses; investigate repeated failures rather than aiming for the reference frequency.

    Try this

    Record Drive Impact and an ordinary attack in separate dummy slots. Mix playback and practice recognizing the difference with a response available in your current gauge state.

    Supporting stats
    • Drive Impact: profile 0.8 · reference 0.23
    • Drive Impact gauge: profile 6.8% · reference 1.7%

  3. Give each meter spend a purpose

    SA2: profile 39.5% · reference 15%

    Review five resource decisions with the actual gauge visible. Note what the spend gained and what options remained; usage shares alone cannot measure efficiency.

    Try this

    Practice a familiar sequence with a resource spend and a conservation alternative. Compare the outcome and remaining gauge before deciding which situation calls for each.

    Supporting stats
    • SA2: profile 39.5% · reference 15%
    • SA3: profile 12.9% · reference 24.5%
    • CA: profile 3.2% · reference 6.9%

AI selects and prioritizes reviewed practice plans from the cited stats. Habits cover the whole profile; character records and recent samples have separate scopes. These are ideas to test, not confirmed weaknesses or video analysis. Latest reported trends across the last 100 matches, not lifetime habit averages.

Data saved Sep 6, 2026 · AI: dots-studio/dots-3-note-preview:free

How to beat this player

  1. Use matchup records as scouting leads

    vs Ryu: 15 wins in 34 games

    A lower win rate suggests footage to examine, not a guaranteed counterpick. Check whether the same difficult situation appears in multiple matches.

    Try this

    If you play the cited character, test one familiar situation from that footage. If the player handles it well, abandon the assumption and adapt to what they show.

    Supporting stats
    • vs Ryu: 15 wins in 34 games
    • vs Dee Jay: 20 wins in 27 games
    • vs A.K.I.: 8 wins in 10 games

AI selects and prioritizes reviewed practice plans from the cited stats. Habits cover the whole profile; character records and recent samples have separate scopes. These are ideas to test, not confirmed weaknesses or video analysis. Latest reported trends across the last 100 matches, not lifetime habit averages.

Data saved Sep 6, 2026 · AI: dots-studio/dots-3-note-preview:free