Notes

LCS json doesnt have info on summoner spells used over time
API only reads ranked games, so we good
next time change if lane = bot, then append role
track 3 time marks for soloq - 6 min (before first back), 20 and final
SoloQ analysis to see what player is safe/greedy/tilting, etc.
op.gg and match history have all the data I could display. So instead, just analyse what i need and just give a summary report on performance. My job is to find smurfs and merge data, and analyse it
so, there are item optimisers with unknown to me weights. I could work with those, but could also make a simplistic one myself, based on gold available and opponents' stats in pro games
if we think that opponents might also use a similar code to scout for smurfs, then share smurf accounts with teammates to make it impossible
enemy might ban because of their scouting, general meta bans or off of scrim experience they had
early boot upgrade seems to suck on adc
armor and mr dont scale as good as dmg, so bruisers are better.
Trundle is good splitpush vs tanks, but is worse in tp plays. so team either has to make sure that there are no flank wards, or have a good disengage.
whever there is no aoe threat, teams prefer to stick together in tf. So i assume that is a natural behavior and aoe threat forces them to cautiosly change their playstyle (ori good)
try to back earlier if you have a chear buy than your opponennt. or to force game go on longer without a sightstone on either side