https://trio-explorer-epm3pwevznw9nf7rw3krvj.streamlit.app
Back with a quick offering! I like the concept draftgap has, taking the normalized synergy deltas from lolalytics and predicting what champs are best based on synergies and match-ups. I play a lot of comp and I wanted a way to effectively visualize my options a couple picks in advance (namely picks 1-3). I made a tool that shows the synergy connections for up to 3 champ pools all connected. You can threshold which trios show up using a slider. If a trio is above the thresh, the 3 champions will be connected. Note that the individual connection between 2 champs may be bad but compensated for by the other connections in the trio.
Like here, Kai’sa-Thresh-Elise is a trio with a mean synergy above the threshold. However, this is mostly because Thresh-Elise is a good pairing, good enough to compensate for Elise-Kai’sa being a bad pairing.
You can selected up to 15 champions for each position’s pool and see positions connected at once. Note that you can lock champions in as they are selected and as you do they will condense graph accordingly. For example, below I locked Syndra and Aatrox at top and mid. You can see how well they synergize with Sup, Adc, Jun by the colored M and T next to the champ names.
Finally, at the bottom of the page you can see how all the champions you have locked in synergize.
I will note, I’ve done a few things differently than draft gap here.
- This is all diamond+ data from 2025. 11 million games in fact.
- This is all normalized synergy delta (delta 2 on lolalytics). Basically, it corrects for how strong or weak a champion is and instead focuses on how good their synergy is. Champion strengths change from patch to patch. My data supports interaction terms being nearly invariant. You can factor champ strength in new patches in on your own.
- This is 2025 data. I haven’t pulled enough 2026 data to include yet, but I don’t think it matters. My 2024 deltas match my 2025 deltas. Champ strength changes, champ interactions generally don’t.
- I did some fancy under-the-hood stats to wash out the fact that some match-ups are played less than others which makes their interaction terms further from their true mean and therefor more volatile.
- I also limited how much really high interaction terms can affect a trio using some log transforms.
These are my caveats! It’s probably not the most useful thing I’ve ever made, and streamlit might crash if you all use it at once, but here it is! Have fun making comps!
(By the way, I didn’t post this to reddit to keep the user numbers down since streamlit deploy is freeware. If it gets overloaded, I apologize. Also if you like it, please comment here and let me know. I may improve or host it if you do.)
Leave a Reply