Blood bowl 2 orc team composition4/21/2024 An approximate way to eyeball it is to see if those 95CI lines overlap at all: if they do then one race cannot be said with confidence to be better or worse than another, wheras if they don't you can say with some confidence that there is a difference.Īnother way to determine "tiers" in the method the OP means is a k-means analysis, which clusters the data into groups. When comparing samples (data from races) one method is to carry out a t-test in order to take those things I previously mentioned into account. The answer depends on a number of factors, not least of which is the environment. Now, what the OP is actually asking is "which races are better than others". Any race where the line crosses the intended tier range of that race cannot be said to be outside the tier with any confidence. In those charts the lines represent the 95CI ranges. That means that in order to say with any confidence that a team is performing outside the given tier it is designed to work within the whole of the 95CI range needs to be outside that tier. So, when we get a win% for wood elves we can look at the sample size, calculate the confidence interval and turn that into a range that represents the range within which we can confidently say wood elves perform. Normally we use 95% and that means the real value of the win% of the race sits within that distribution, given as a range, to 95% confidence. What we can do is measure how confident we are in the win% number given in terms of something called a "confidence interval". Some biases can be controlled for and some can't be so easily, but in most cases taking a really large sample controls for many of these variables. What we actually have with any set of data is a sample, and that sample may or may not be representative of the entire population of games played by that race, ever. That's because there are plenty of errors and confounds in the data including coaching biases, compositional errors and the sheer fact of randomness in Blood Bowl. When you look at, for example, FUMBBL data, you can't simply take the win% there are claim that is the win% for the team. The next question is whether the teams perform within those tiers, and that's where things get a bit more complex. So, those are the tiers and that's what their intended performances are. Tier 3: Halfling, Goblin, Ogre (30% to 35%) Tier 1.5: Slann, Underworld, Chaos Pact (40% to 50%) Tier 1: Anyone not mentioned (45% to 55% win rate in tabletop league play) The list Doomy gave above is close enough, but the actual tiers the BBRC used were: Right, the first thing is "what is a tier?" The tiers are set by the designers as intended performances for the teams and define the only balance parameter for Blood Bowl.
0 Comments
Leave a Reply.AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |