15
\$\begingroup\$

Some questions have a number of bonuses stated as percentages. For code golf (where the majority of these bonuses are seen) these will be percentage reductions of the score.

If an answer achieves several of these bonuses, how should the resulting score be calculated?

There are two main ways I can imagine:

  1. Apply each percentage in turn

    • 10% reduction followed by 20% reduction results in 90% of 80% = 72% of the original score.
  2. Sum the percentages before applying them.

    • 10% reduction and a 20% reduction results in a 30% reduction overall - that is 70% of the original score.

To me, (1) makes more sense, but it also happens to result in a higher score, so an answerer has an incentive to use (2).

Does it make sense to standardise here and settle on a default way, or should it be up to the question writer to specify each time?

| |
\$\endgroup\$
  • 2
    \$\begingroup\$ Obligatory \$\endgroup\$ – mınxomaτ Jan 10 '16 at 20:11
  • 1
    \$\begingroup\$ I've changed the title of this because we absolutely CAN standardize it (we can make whatever policies we want ;)). The question is SHOULD we standardize it (and how). \$\endgroup\$ – a spaghetto Jan 10 '16 at 20:11
  • \$\begingroup\$ @quartata your pedantry is most welcome here - thank you :) \$\endgroup\$ – trichoplax Jan 10 '16 at 20:18
  • 12
    \$\begingroup\$ Remove bonuses, problem solved \$\endgroup\$ – user45941 Jan 10 '16 at 20:31
  • \$\begingroup\$ Method 2 gives better score but it's mathematically nonsense. If you get a 50% discount and on top of that another 50%, are you going to pay 0? \$\endgroup\$ – edc65 Jan 11 '16 at 20:19
10
\$\begingroup\$

This should be left up to the question writer

Each question has its own specific situation, so there is no formula to fit all cases.

The other issue is that many posters may not be aware of this meta post. I'm not against having defaults, but it can also be odd for a new user to suddenly hear that rule X from meta post Y applies to their challenge.

I know occasionally a poster makes a silly mistake with bonuses that leads to a negative score. Often someone points this out to them, but if it goes unnoticed and someone gets a negative score, it doesn't really matter (unless everyone ends up with the same score.)

| |
\$\endgroup\$
  • 3
    \$\begingroup\$ I agree with this. We already have so many implicit rules buried somewhere in meta, that I think this one is a bit too specific to add to the list. That said, I think having a discussion about the different ways bonuses can be combined and if some of them have advantages over others is probably a good thing, so we can direct people here who can't decide how to make their bonuses work. \$\endgroup\$ – Martin Ender Jan 10 '16 at 21:27
  • 1
    \$\begingroup\$ @MartinBüttner this "implicit rule" has very little side effects: If somebody posts a question without knowledge of this post AND doesn't post guidelines on how to score, then we have a backup "good enough" scoring. If a potential answerer isn't sure how the bonuses work on a question, then the worst I see happening is they leave a comment, or they miscalculate the bonuses on their answer (which would be super rare in my opinion). \$\endgroup\$ – Nathan Merrill Jan 11 '16 at 14:37
  • 1
    \$\begingroup\$ @NathanMerrill I don't think it's necessarily a good thing to have a backup "good enough" scoring. Especially for a new user, it might be surprising for people to make an assumption that was not their intention (and once a few answers have been posted, it's hard to correct their oversight). That's why I think it's better to ask the OP for clarification. \$\endgroup\$ – Martin Ender Jan 11 '16 at 14:40
9
\$\begingroup\$

Multiplicative bonuses should be the default for percentage bonuses

This is in line with how multiple percentage bonuses are typically applied, both here and in other competitions. Getting two 50% bonuses shouldn't leave you with an answer score of 0; you should get a total of a 75% bonus.

Flat bonuses are additive

If you get 10 bytes off for one bonus and 20 bytes off for another, you get 30 bytes off total. Simple as that.

Percentage bonuses first, then flat bonuses

Order of operations: multiplication before addition.

That said, challenge authors are welcome to use whatever method of applying bonuses they desire. These are just defaults.

| |
\$\endgroup\$
  • 1
    \$\begingroup\$ That's what I would go for as a default (if you think there should be a default, which I'm not convinced of.) If there are both multiplicative bonuses and flat bonuses, which gets applied first? It's a minefield :-D \$\endgroup\$ – Level River St Jan 10 '16 at 20:51
  • 2
    \$\begingroup\$ I agree with multiplicative percentages as a default, but I can't upvote due to the inclusion of the percentage then flat rule, which I don't have an opinion on and don't wish to implicitly vote on. \$\endgroup\$ – trichoplax Jan 10 '16 at 22:56
-1
\$\begingroup\$

The default for percentages should be multiplicative

  1. Apply each percentage in turn

    • 10% reduction followed by 20% reduction results in 90% of 80% = 72% of the original score.

I have no opinion on the application of flat bonuses

This answer has been added because I cannot upvote the answer that also expresses an opinion on flat bonuses without implicitly voting for that too.

| |
\$\endgroup\$

You must log in to answer this question.

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