I was going to write this one another thread about win rates but I believe the examples I wanted to give would just clutter the discussion so I decided to move it here.
Before I continue, here’s the thing we need to understand about win rates: if you plot the win rate of all players on any given day, you’ll get a perfect bell curve with a mean of 50%. Not that many players will ACTUALLY have a 50% win rate. Some have more some have less, depending on several variables, including but not exclusive to, the archetype they play. Still, it’s a simple descriptive statistics and works every time on games like HS.
What happens if you create the same plot for only a specific archetype? It may be a bell curve or not, but it will be similar to it. However, the mean will not be at 50%. The mean will be something like we see on HSReplay. Some have a higher mean, others a lower one, but all will have (assuming a decent number of users) something similar to a bell curve. Again, the mean will be the true win rate for some players but not the vast majority of them.
Given a large enough sample size, we can say, with some confidence, what affects the individual win rate and why we cannot expect to have a win rate equal to the one we see on HSReplay.
- Overall player experience aka skill. We can say that this affects the win rate because it prevents basic errors.
- Specific knowledge of the archetype (and even the class). We can say that this affects the win rate because it helps to find lines that we wouldn’t otherwise consider, even (or especially) if they go against the overall player experience.
- Knowledge of the meta. We can say that this affects the win rate because it allows the player to anticipate plays and better manage resources on any given match and, more importantly, predict if it’s a good time to play a specific archetype.
On the other hand, there are things that don’t affect the win rate over large sample sizes. RNG is a great example. RNG will never matter to an archetype’s win rate and it will only matter for a player for the specific match where the RNG outcome occurred.
In short: archetype win rates are irrelevant for the individual experience. What matters for your personal experience is your skill and your knowledge. Reducing all this to aggregated data and then having expectations about your outcome is detrimental to the discussion and the growth of the individual as a player.
Nothing wrong about using it to discuss the meta though…
