Tom Griffiths's dashboard

Hi anaverageamerican,

Yes, there have been some changes surrounding voiding to fix a bug where losers on a voided question were not getting their stakes returned.

Voiding should work as follows:

(1) if an open question is voided, all predictions on that question are voided and users have their stakes returned.

(2) if a settled question is voided, users who won as a result of the initial settlement retain their winnings, 'losers' from the settlement get their original stakes back.

The reason you were seeing all zeros was a separate display issue where for a short time predictions that resulted in 'false winnings' under (2) above were being shown in the Settled section of My Predictions, not under voided. This has now been sorted and so under Void in My Predictions will be any predictions made on a voided question (except those that were cashed in).

Hi Ferdi, thanks for this suggestion. Duplicate questions are certainly a problem, and we apologize to any user who has lost predictions due to a question being voided because it's a duplicate. The onus is currently on the question creator to first search for duplicates before creating a question, but we recognize this could be easier to do. We are currently planning to put in a middle step to question creation that will highlight duplicates that we hope will reduce this problem significantly.

Hi guys, we recognize the current system (i.e. relying on us to settle questions) doesn't scale well and is already causing delays and frustration that we want to eliminate asap.

Our solution in the pipeline is again to empower top users that have proven their trustworthiness and ability to settle questions themselves, hence speeding up response times and allowing the site to carry a greater number and variety of questions.

There's clearly potential for much controversy here though and so we're still working how to do this best. As ever, we'll endeavour to keep everyone up to date via Get Satisfaction, and a new notice board feature we're putting live soon.

Thanks for all your comments in the meantime.

Tom

"Are the non-zero entries also possibly cash-ins on questions that were subsequently voided?"

Yes. Cashed in predictions on questions that are subsequently voided currently show in the 'Voided predictions' sections, because it gives all of your predictions on questions that have been voided.

I actually think that should change because it's confusing, so predictions only show in one section (i.e. the cashed in ones you're seeing there will only be in 'Cashed in', and 'Void predictions' will be ones voided with questions).

We'll address this in the next release. Thanks for noticing it.

Hey InfernalMachine,

Voiding works as follows:

- if an open question is voided, all predictions on that question are voided and users have their stakes returned.

- if a settled question is voided, users who won as a result of the initial settlement retain their winnings*, 'losers' from the settlement get their original stakes back.

Hopefully what you're seeing is explained by this - does that clear things up?

Tom

* at least for now, see http://getsatisfaction.com/hubdub/top...

One scenario is going all out on a question, losing, then having funds reclaimed from void of a previous win, sending the balance way negative. There would be no cash for predictions, and no predictions to cash in. That would mean logging in every day for a while to get the H$20 before more predictions were possible, something many people wouldn't bother doing if they were significantly in the red.

Please give us your own thoughts on the original idea first, then a response to this scenario if you have one.

Thanks,
Tom

Hey Rohan, that is correct, you are not meant to be able to cash in on suspended questions. The fact that this was possible before was a bug. This rule is in the spirit of suspended questions - i.e. you should be clear on what position you are taking before the suspend date (e.g. the start of a sports match).

The full list of rules that we've been working on recently with users' help will be uploaded shortly.

Thanks for flagging this up for others in the meantime though.

Cheers,
Tom

Hi tarheelv, I take your points, but what we are stopping is not the betting on these short term questions, only the premature cashing in of the predictions. That means you would still be able to predict on daily stock movements and rapidly developing news, just not cancel your predictions until the question (not your prediction) has been in existence for 12 hours.

I see though that there are situations such as the ones you mention where the new restriction could require a change in tactics.

How do others out there feel? Remember, this was introduced to help stop gamers from taking over the leaderboards unfairly.

@egalloway: This particular question has been voided because the question was unclear. To avoid the same date confusion in future, we have changed the site so that 'End date' now reads 'Suspend date', to make it clear that (if given) this is the point at which the question will be automatically suspended from further predictions.

If the question creator wishes to specify an end date by which the event should occur, they should specify this explicity in the question title or background info.

We are also adding functionality to show users why particular questions were suspended or voided.

@electroaffinity: We're still deciding how to deal with gamers. It's more likely we'll fine them than ban them outright.

@4d: The question has been voided which means you should have had your original stake back.

@Volodya: Your points have been noted, thank you. If one were to assume Fahrenheit and take the correct definition of temperature then I agree the correct answer to the question is No. However, the issue is whether those assumptions were intended by the question creator and universally obvious from the question. In our judgement they were not. On Celsius vs. Fahrenheit, I agree the question creator probably meant Fahrenheit but to an international audience this may not have been clear. In addition, it is in fact questionable whether the question creator was referring to true "temperature", as creating a question about the -50 point makes more sense if wind chill is taken into account (actual temperatures are highly unlikely to reach below the -20s).

Hubdub cares very much about being transparent and responsive. We could have chosen to ignore the lengthy discussion surrounding this question but instead have spent several hours collectively trying to resolve it. Our decision to void the question reflects its indisputable lack of clarity, given there are no established case histories or standards to ensure all users would have made the same assumptions when predicting. As such we are (1) modifying the site to encourage question creators even more to be as specific as possible in their question creation; (2) launching a public 'case history' resource AND a list of standard interpretations in an effort to create the "level playing field and well-defined, evenly applied rules" we and our users want.

Thanks to everyone participating in this discussion. The three issues in play are:

1) Is the question clear enough to be valid at all?
2) If so, what is the correct answer?
3) Was this question used to game the system?

Our conclusions are as follows.

1) Is the question clear enough to be valid at all?
No. Given that two critical factors in determining the outcome are not specified in the question (a: Celsius or Fahrenheit; b: is wind-chill to be taken into account), the question is not valid.

2) If so, what is the correct answer?
Without a clear question, this cannot be answered.

3) Was this question used to game the system?
We are conducting a site-wide audit of questions and users suspected of gaming the system and will deal with that issue separately.

Action:
The question has now been voided.

This is the final action to be taken on this question, and to avoid similar ambiguity in future we urge all users to be as specific as possible when creating questions.

Here at Hubdub we're in the process of learning how best to deal with situations such as this, and all your comments have been very much appreciated. In these early days it's likely we'll err on the side of voiding questions, until we've built up a standard set of conventions for settling otherwise ambiguous questions.

Thanks again for your well-made points and background research.