Research is weak in the early game — what should I bet on instead?

True, and combat calibration confirms it. In the early game, research percentages are small — in one of Adel's real fights, Spearmen Damage was only worth +5.2%, comparable to the ±7% matchup and far behind other levers.

What dominates early:

1. Dragon↔troop affinity (+20% to stats) — pick a dragon with positive affinity for your type.
2. Auras — Champions (−8% damage taken), Rhysarion (−27.5% damage) weigh far more than research.
3. Your dragons' level and stars (raw stats).
4. The triangle matchup (±7%) — field the type that beats the enemy.

➡️ In the late game (research in the hundreds of percent), it's the reverse: research on the type you field crushes everything. So weight accordingly: early = dragons/affinity/auras; late = research.