How I hit 100th Percentile
in DotA TI fantasy.
The Secret Sauce
Want the Complete Dataset?
I’ve analyzed 650+ matches over the last 6 months from all 16 teams that are competing in The International 2025. Now with only 8 teams remaining, here’s the biggest mistake I see: picking based on averages.
Fantasy rewards ceiling games, not consistency. A player who averages 9,000 points but hits 15,000+ occasionally beats someone steady at 11,000. My Composite Scoring System weighs both baseline performance and standout games.
Infinite Scaling Wins
Target stats with no ceiling: creep score for cores, wards for supports, and stuns for teamfighters. Avoid capped stats like kills – even stomps rarely exceed 15 kills, but a Nature’s Prophet can hit 1000+ last hits.
Stats Over Percentage
When rerolling fantasy banners, prioritize getting the right stat first. A 300% bonus on “Smokes Used” for a midlaner (who uses 1-2 smokes) is worth less than 150% on GPM. The base matters much more than the multiplier.
Why My Scoring Works
This captures both average and ceiling scores. Since only your best series counts, you need to manage the risk of choosing explosive players with consistent ones. The composite score finds the perfect risk-managed middleground.
My Personal Picks
My Main Stage fantasy picks are strictly data-driven based on the composite scoring system’s data, targeting specific metrics that excel in long games.
For cores, I’m prioritizing players with the potential for 1000+ last hits, a feat often achieved by dedicated Nature’s Prophet spammers (*ahem* skiter *ahem*). My supports are chosen for consistent warding, efficient camp stacking, and full smoke usage. For the mid lane, I’m maximizing stun duration while balancing a high GPM ceiling with kill potential—a necessary compromise since my card rolls for Creep Score or Madstones weren’t ideal.
Analysis of statistics from the last 15 games reveals that Team Falcons is a clear outlier, consistently playing longer matches (60.5 minutes average) where their players post incredible scaling stats. This data has led me to a high-risk, high-reward strategy: I’m all-in on a full Falcons stack (skiter, Malr1ne, Cr1t-). It’s a gamble, but the numbers suggest they have the highest fantasy point ceiling. If I were playing it safe, a more conservative draft would involve swapping in Shiro (Team Tidebound) for my Core and xNova (Xtreme Gaming) at support to diversify my team choices.
Want to Do Your Own Analysis?
I’ve made my complete dataset public. Sort by any metric, filter by team, or find hidden gems based on your specific fantasy roster and banner bonuses.
Note: The Lotuses Grabbed Stat isn’t accurately tracked by Valve, hence it is excluded from my dataset.
KeenPlay 