Matt Haberland

> Another option might be to convert these objects to bunches, and add support for aliases I think you can convert them to bunches and just add the attribute manually...

@JozsefKutas I added two more to the list above. I guess we missed the things in `scipy.stats.contingency` on the first pass. Anything else to add?

Issue scope is up to you! Personally, I'm ok with leaving the rest out of it. There is hope of a consistent interface among the statistical test functions, but the...

I think we should add `statistic` as an alias to all of them. There's no harm in it, since it's unlikely that a user would accidentally use that attribute name...

@dschmitz89 I'd be happy to merge this, but the global benchmarks don't follow the standard ASV pattern, so it's tough for me to tell whether they work as intended. >...

Ok, then someone will need to try running these. @dschmitz89 you don't need to run the whole suite; it would be ok to modify the code so that only some...

Here are the results (1 trial for each problem) ``` · Running 1 total benchmarks (1 commits * 1 environments * 1 benchmarks) [ 0.00%] ·· Benchmarking existing-py_opt_anaconda3_envs_scipy-dev_bin_python [100.00%] ···...

Seems to be working. Thanks @dschmitz89. > These benchmarks are not run as they need SCIPY_XSLOW=1. See BenchGlobal I'd suggest a follow-up PR that runs a subset of the benchmarks...

> Did you test that too? You can see the results in "Details" above, right? It looks like the stochastic solvers are running fine. | | avg nfev | success...

I ran all problems once, and it is the average of the "success %` for all problems with the given solver. You may need to look at the code to...