I am calling for Blizzard to give us the data on how snowball push mode is
I would estimate that whichever team won the first fight their chance of winning spikes to 70% or more.
This game mode is so, so sad to win AND lose in. Especially when you lose the first fight and you are instantly counter picked, so you have to change hero and walk back to the fight for 5 mins (the maps are bad btw) and you know they have the ult economy to win it
if you aren’t changing heroes after a first fight loss then you are doing it wrong.
Whoever wins the most team fights in the game likely is the better team, so it makes sense why the winner of first team fight has much higher chance of winning the game.
They pushed the robot to our spawn in 30 seconds but we won a team fight so we’re still in this
True, and iv seen come backs pretty regularly. Though I must say on a different note the walk back feels pretty painful for heroes like Zen, Sigma and Anna, especially on slow support when there is often a flanker in wait to pick you off and keep you from returning to your team.
30 seconds is literally impossible, fake post.
Winning or losing the first fight does not guarantee a win because both teams usually get ults before one wins
Managing ults key for this mode
When ever I backfill a push map in qp the robot is always at the final stretch and there’s still 4:30+ on the time bank meaning someone thought “screw this” and left
I have the inverse experience I find more often than not the team that loses the first fight tends to win because the team that wins SECOND fight almost ALWAYS wins and that tends to be the team that lost first. I am willing to bet second fight is the best predictor of victory in push.
Does blizzard actually have this data?
This would require an algorithm that can determine what part of the match was the first fight and who won (preferably also the second fight).
Furthermore, they would either need to have run this algorithm over the matches already, or they would need replays to run it over now.
To run the algorithm over replays, they would presumably need to prepare builds of all relevant versions of the engine with this algorithm patched in. (Otherwise, why are replays dropped between patches?)
Anyway, if somehow they could supply this data, I’d like to see it broken down by map as well.
So the desired dimensions are:
And each item should be the count/proportion of games in that bucket.
We could then easily compute per-map and overall percentages from that.
Just checked my last 3 replays of the mode.
43BCSH - My team won the first fight and lost
QR3RR2 - My team won the first fight and lost
HQCK35 - My team lost the first 2, arguably 3 fights and won
Who got first checkpoint first? That tends to be a better indicator of who is going to win. I know early on, during contenders before OWL played push, the early statistic was that the team who got first checkpoint first won about 70% or more of the time.