Kjetil and I have had a disagreement about whether my answer here should be a comment.

OP's question is essentially one of reading/finding documentation for how xgboost works.*

The full and complete answer to OP's question is that the user's guess is correct: the function works in the way that the user expected. I didn't write more because there's not anything more to say.

Kjetil's contention is that short answers should be comments even if it completely answers the question, citing the Michael Chernick thread. (This is not a straw man argument - it's the content of the comment thread.) I do not believe that contents the Michael Chernick thread apply -- the core contention there is short answers tend to also be incomplete, unhelpful or confusing answers, which are frowned upon.

By contrast, my answer in this thread fully answers the user's question. It took a little digging, but I eventually found a source for the claim (which I knew to be true) that I make in my answer (a fact which may or may not change Kjetil's contention, since it occurred after the comment exchange, but it took a moment to find the relevant doc; regardless, the answer is true even without the doc).

For an example of a short answer which completely answers a question, and was modestly up-voted, see my answer at Closed form for $\mathbb{E}[\ln (1-p)]$, for $p \sim Beta(\alpha, \beta)$ which received no negative feedback, presumably because it completely answers the question.

* Whether that's on-topic is a different matter altogether, but I know from experience that finding xgboost documentation, and even more, documentation for a particular xgboost version, is a bit of a struggle, so I think it's worth providing an answer here.

  • 3
    $\begingroup$ Perhaps the underlying difficulty is that this site is not geared towards addressing yes/no questions. $\endgroup$
    – whuber Mod
    Commented Feb 24, 2017 at 17:50
  • 2
    $\begingroup$ But we still get them, how should we proceeded? I don't see much harm in answering. $\endgroup$ Commented Feb 24, 2017 at 19:51
  • 8
    $\begingroup$ If it answers the question, it should be an answer. Why should it be anything else? (I am not posting this as an answer as I believe a more elaborate explanation is desirable. But if not, this would be in conflict with my own argument.) $\endgroup$ Commented Feb 24, 2017 at 20:18
  • 10
    $\begingroup$ Such a short answer is only possible because the Q is off-topic and will likely be closed as such. I don't see much sense in answering off-topic questions; off-topic questions ideally should not even be here, so why answering them? If the question is closed and not upvoted it will be automatically deleted after some time, which is a good thing. Having an answer with positive score will prevent it from being deleted forever. To me it looks like garbage collecting. At the same time, answering in comments is honourable and helpful because it will help OP without messing with the site's logic. $\endgroup$
    – amoeba
    Commented Feb 24, 2017 at 21:26
  • 5
    $\begingroup$ I fully agree with @amoeba on this. Low-quality questions should be not be encouraged with answers. Matthew's comment was adequate. On occasion I "start answering" the question in the comments (eg. 1, 2 so the OP can formulate/substantiate the real question they have properly and then provide an answer that is constructive for the OP as well as future readers (and on occasion we see there was little to answer anyway (eg. 1). $\endgroup$
    – usεr11852
    Commented Feb 25, 2017 at 12:19
  • 3
    $\begingroup$ As @whuber said, the site is not geared towards yes/no questions but we ask users to research the question and work on it before asking and show us his work. When the question is well researched (as we ask), often the OP has found a possible answer and, if they is right, we only can confirm it. The only way to avoid the issue would be asking users not to research too much their questions, but that wouldn't benefit the site nor the askers. $\endgroup$
    – Pere
    Commented Feb 25, 2017 at 17:16

1 Answer 1


Like Richard Hardy (above), I think that attempts to answer the question should be done as "answers", not comments. This lets people up- (or down-)vote the proposed solution, provides a place for comments on the proposed solution (vs. the question), and makes the "Unanswered" tab work correctly. Putting answers, even short ones, as comments breaks all of the answer-related infrastructure on the site.

If it answers the question in a self-contained way, I don't think an answer's length particularly matters. For very short answers, it would be nice if the answerer provided a bit of "bonus" information (background, support for their answer, etc), but I think this extra effort should be more of a social norm than a hard and fast rule.

Sycorax's answer here seems like a bad test-case for several reasons: borderline off-topic question, the link isn't the greatest (it's essentially another question and answer). However, there are a number of other short answers that seem fairly non-controversial and, judging by the votes, useful and these should be encouraged.

  • $\begingroup$ I agree that the question is marginally topical, but the important part of the answer is who is saying it: a key contributor to the xgboost respo. It's not just another question and answer -- it's more or less definitive to the exact question OP asked. $\endgroup$
    – Sycorax Mod
    Commented Feb 25, 2017 at 23:37
  • 4
    $\begingroup$ See, I totally missed that it was Tianqi Chen answering the question! I think a "perfect" answer might have said that "Yes. predict_proba gives you the estimated probability that an example is in class 1. Here's a conversation with Tianqi Chen (tqchen), one of the authors' of xgboost [link] and here's the documentation he refers to [link]. $\endgroup$ Commented Feb 25, 2017 at 23:51
  • 2
    $\begingroup$ Sometimes I answer in comments because the answer to the question is not what the OP thinks he is asking, and because I have a low tolerance for downvoting, especially when I do not feel the need to explain what I am saying, how I am saying it, why I am saying it and all that in vanilla flavor to all comers. RE answers, frequently less is more, and terminology being ambiguous between disciplines, more is more to misunderstand. $\endgroup$
    – Carl
    Commented Feb 26, 2017 at 2:49

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .