The six month window was a first ditch attempt on my part to think of some way to soften the blow to the newer people who dropped so severely in ranks.
It was also an attempt to get the conversation started, and it's a great conversation!
Of all the ideas presented so far, I am liking the: shorter window (coupled with Maddes greatest folders list) and the exponential points (wow, nicely thought out, infjamc).
Another (perhaps hare-brained) idea to throw out here may be to keep the <15 & <150 puzzles separate from the normal puzzles for global points, period. Have a <15 global points scoreboard, a <150 one, and our usual normal one. The benefit would be that a new player can judge themselves against their peers, and be able to get their feet wet in normal (with a more realistic rank) - and do both at the same time.
That's really what the current system was designed to do, and it functioned well with fewer players - with the current larger numbers, it now fails in encouraging players who start out slowly in terms of "getting it". And yeah, I was one of those players when I started - I would have been one who dropped the 100 ranks. It took me 2 months to have one finish in the top 30… 4 months for a top 20… 8 months for a top ten (then another four before I was able to do that again)… almost a year & a half for a top 5. My progression was very slow, but it was steady, and that encouraged me. I was able to eventually gain global rank #7 at one point, so I guess that I'm not so bad at the game after all - perhaps it would have been a shame to have chased me away with a scoring system that I didn't understand when I started.
I think you forgot that in your second example (5 instead of 7) almost all players would gain more global points, so the 400 points were good for #208 in the existing system but not in your example.
That's a good idea but it would be impossible for players to catch a flawed leader-board with such a complicated system - and we all know we can't really trust the leader-board.
Another idea of dealing <15 <150 and normal puzzle: if player go for all 3 versions of one puzzle only best score (global points earned) counts. This way we allow to play on all 3 and this also prevent sudden rank drop if player do all of them high ranked and puzzle left scoring window.
That's a very good point… also, it has been brought to my attention that there are actually 53 puzzles in the last four months (I only counted off three months in my previous post).
Obviously, it would be difficult to estimate the new rank under the case where the exponent is 5 (since it would require recalculating everyone's scores). but I would expect that the general trend–that the gap between a slightly above average performance and an outstanding performance would be narrower–would still apply. So, the new rank will mostly likely be lower than 208, but still higher than 445.
Alternatively, the "regular" version of the puzzle could be hidden from those with <150 global points to prevent the case where the best Foldit score is achieved on the "regular" version of the puzzle. Also, as brow42 has already mentioned, the <15 and <150 puzzles could be weighted less– say, 25% and 50% of that of the regular puzzle, respectively.
I'm going to throw another thought into this mix while the subject is being discussed and say that while puzzle scores are a good thing for hooking players in and getting them trying to improve, they are probably ultimately a bad thing when it comes to actually producing the desired results from the game.
Score chasing tends to get us constantly digging down into local energy minima and all narrowing down to the same structure rather than experimenting and producing a diverse range of options in the solutions to a puzzle.
Problem is to find a way to reward people for puzzle results other than on the basis of the best score - which is not always an absolute indicator of the best solution anyway.
I haven't got any answers, just a feeling that it would be a good thing if we could do something of that nature in order to get better solutions for CASP and beyond.
I like the idea, but the tricky part is that relative contribution to science is difficult to quantify:
-
For CASP puzzles, one possibility is to calculate global points based on deviation from the native in terms of GDT_TS or some other metric. The problem with this approach is that it requires manual recalculation on the part of the Foldit staff (since only the top Foldit models are submitted to CASP); plus, it obviously doesn't work for puzzles where there is no native to compare to.
-
Yet another possibility is to create a new achievement type for individuals whose results represent significant contributions to a study that ultimately results in a published journal article. (Obviously, "significant" is subjective, so this will have to be awarded manually by the researchers.) For extra visibility, perhaps the player could be allowed to be adorned with a star next to their username (just like the ribbon for admins). The idea is that there would be a tangible reward for actually "doing something."
No, they can not be scored less, they are identical. <15 is not easier than <150
the achievement idea is great infjamc, partcularly if there can also be one for coders.
My understanding is that in CASP the known structure has been found, just not released, so whilst it may take a while the deviation from the native may be calculated?
Also in the olden days (18 months ago) beginner <15 and <150 puzzles were only up for a weekish, so it was harder to do well in the <15 and then transfer that to the <150 and the unrestricted puzzles, thereby scoring well in several puzzles.
I am not suggesting that beginner puzzle times should be shorter, I don't think that is a good idea, but it may be a factor in beginners getting higher scores at first, then being disappointed when they drop.
