Two possibilities:
First, the idea that, yes, you can draw conclusions from clusters in data. That’s why the actual post says “[Cluster analysis] is commonly used in market research”, for instance.
Second, from the lower half of that post: Why the use of categories in this manner is important. It’s the same as from earlier, that mean averages are frequently bad because they get poisoned by size oddities in the sample data easily.
I was given information and a chance to absorb it, and after tanking on the information a few days came to a realization about what it actually said and what its underlying tenant was despite having been originally stated poorly.
I was additionally spurred to that realization by factors not relating to the evidence, but it is important to note that I haven’t actually changed my position at any point. It has been and will always be that the three races are approximately balanced*, and that any oddities in a player’s - or even a statistically aberrantly large group of players - is owing to some number of factors that are coincidentally common in that group.
* Balanced is not the same as well designed or fun to play. I’m always happy to rag on how the Disruptor and Widow Mine are garbage unfun units that I’d rather the game not have, but they’re not meaningfully imbalanced.
Well, that’s a different tack entirely since you’re talking about a tiny group of elites instead of the general population, but I understand how it’s related.
I do, actually, mildly agree with Miro’s statement on this: That the current pro Protoss players are equal to the average rest of the pro player population, but that both Terran and Zerg have extremely strong current representatives. Maxpax is the Protoss player, head-and-shoulders above, and he just refuses to participate in the relevant tournaments – and that’s a huge factor, because it means that one of the top five Protoss is basically eliminated in advance.
As an example, take a 16-player tournament: Top 5 representatives of each race, plus an extra. For Protoss, you then immediately lose one of your best, and that means there’s either two ‘extras’, or your new Protoss player is likely below the rest.
So it only seems logical that Protoss’d be under-represented in winners. To me, the thing that then cinches it is that the Disruptor is a horribly designed unit that Protoss is extremely dependent on in a lot of situations - victory depends on enemy mis-micro, and that increasingly does not happen as skill rises.
Or, without it, the Carrier, which is extremely heavily countered by Corruptors, Vipers, Thors, and Vikings. My pro-viewing experience has been that the Tempest simply doesn’t do the job despite its increased safety.
No, the statement’s likely, or find believable, and not once certain.
Because we didn’t do a proper survey of the player base, we can’t be certain.
And that’s just not true. Reasons were listed, and completely ignored or dismissed with no affair.
Please provide evidence for your claim. You have pulled it out of nothing and given no context for why you believe that.
Which, you have claimed, is not what you do:
Ergo, please give me the evidence that drew you to this conclusion. Hold yourself to the standard.
