|
On February 16 2013 01:01 sylverfyre wrote:Show nested quote +On February 15 2013 08:05 Ryuu314 wrote: So how do you check how many points you have? As far as I can tell, if you're not in TL's top 20 or so, you can't? That makes no sense to me. Posting over there to see if we can't resolve this - the admins have at least been pretty responsive. If they keep the same system in the future, I'd be willing to run our own fantasy league using ol'fashioned spreadsheets using our own Fantasy Proleague style rules (set a roster, plus trades) This would be awesome.
|
We congratulate Redox on winning week 2 of LCS Fantasy! 97 points is pretty good. Runner-up is EpicPleasure, scoring 95.
Seems unlikely they'll change the system for the site, though, according to the few posts in their forums. I'd gladly participate in a League run by sylverfyre instead.
|
my lu for this week:
DoubleLift nRated sOAz Chauster xPeke
|
Jeez, so hard to pick with like 80% of all players being +20 bucks.
My lineup: Jimbz Aphromoo Chauster ecco Spontexx
ecco pls carry Clakeyfeed ;;
|
Lineup: HosaN Babeta HotshotGG xPeke Morden
Feels like all the good choices for AD carry and top became just abit too expensive.
|
On February 16 2013 01:01 sylverfyre wrote: If they keep the same system in the future, I'd be willing to run our own fantasy league using ol'fashioned spreadsheets using our own Fantasy Proleague style rules (set a roster, plus trades)
i'm currently working with some people on a fantasy site that *makes sense*, with the above concept
>draft a set roster >trades >a way better UI
just give me a little time and i'll pull through, TL!
|
On February 18 2013 21:37 DragoonTT wrote: We congratulate Redox on winning week 2 of LCS Fantasy! 97 points is pretty good. Runner-up is EpicPleasure, scoring 95.
Seems unlikely they'll change the system for the site, though, according to the few posts in their forums. I'd gladly participate in a League run by sylverfyre instead. Don't hold your breath on one from me, I'm still not sure exactly how we should be valuing players, and I'd need help collecting stats. I think it's more likely going to be for the second half of the season that I put together one. I'll maybe run a test-run on a week or two of this one jsut to try and get
Things I wish I had access to for valuation: Gold per minute (but casters dont show gold screen often) I like their KDA/minute system (though I think CS is greatly undervalued in the current system, and rewards bloody messes of games greatly over clean play.
I think win/loss should be factored in - ultimately, this is where true skill factors in, right?
I think dragon/baron should be factored in somehow (best way would be if I could find a way to access gold/minute from players)
|
On February 19 2013 04:01 sylverfyre wrote:Show nested quote +On February 18 2013 21:37 DragoonTT wrote: We congratulate Redox on winning week 2 of LCS Fantasy! 97 points is pretty good. Runner-up is EpicPleasure, scoring 95.
Seems unlikely they'll change the system for the site, though, according to the few posts in their forums. I'd gladly participate in a League run by sylverfyre instead. Don't hold your breath on one from me, I'm still not sure exactly how we should be valuing players, and I'd need help collecting stats. I think it's more likely going to be for the second half of the season that I put together one. I'll maybe run a test-run on a week or two of this one jsut to try and get Things I wish I had access to for valuation: Gold per minute (but casters dont show gold screen often) I like their KDA/minute system (though I think CS is greatly undervalued in the current system, and rewards bloody messes of games greatly over clean play. I think win/loss should be factored in - ultimately, this is where true skill factors in, right? I think dragon/baron should be factored in somehow (best way would be if I could find a way to access gold/minute from players)
I'd say win/loss should be valued the most, and KDA/personal performance second. If you constantly get high KDA and farm but cannot win a game, are you even among the best players?
|
United States47024 Posts
Scoring mechanism as from DotA-Academy's Fantasy League:
+ Show Spoiler + How do we get points?
Now, this is the tricky part, but we'll make it so that you don't have to do any math at all. First of all, for each match, we track 5 things about each player:
Which hero he picked. How many Kills he had. How many Deaths he had. How many Assists he had. How much GPM he had.
Gee, Bruno, that looks a lot like the stats you track in the site! And it's for a reason! The points you'll get will depend on how well that player did, and compare those totals with the averages for that hero. Depending on how much worse/better they were from the averages, you'll get less/more points.
There's a different system for each stat. Click here to know how it goes.
For Kills and Assists:
If you get more than 2x the Average for that hero, you get 5 points. If you get between 1.5x and 2x the Average for that hero, you get 3 points. If you get between 1x and 1.5x the Average for that hero, you get 1 points. If you get between 0.5x and 1x the Average for that hero, you get 0 points. If you get between less than 0.5x the Average for that hero, but more than 0, you get -1 points. If you get 0, you get -3 points!
For Deaths it's exactly the opposite
If you get 0 deaths, you get 5 points! If you get between less than 0.5x the Average for that hero, but more than 0, you get 3 points. If you get between 0.5x and 1x the Average for that hero, you get 1 points. If you get between 1x and 1.5x the Average for that hero, you get 0 points. If you get between 1.5x and 2x the Average for that hero, you get -1 points. If you get more than 2x the Average for that hero, you get -3 points.
For GPM, the score is trickier. While Kills Deaths and Assists don't have any hard high constraints (A player can end up 99-76-153, although really unlikely) GPM is generally more rigid so the formula used before doesn't really work too well. So the calculation we use, involves the concept of Standard Deviation. Simply put, Standard Deviation is a value calculated, similarly to how you'd calculate an average, and it helps you calculate for a specific number, what percentage of the values from which you're calculating the average. So, for example, if a certain group of values has an average of 10 and a standard deviation of 2, it means that: ~67% of the values in that group are between (10 - 2) and (10 + 2). It also means that ~50% of the values are between (10 - 0.674 * 2) and (10 + 0.674 * 2). The 0.674 comes from a table ( 95% is 1.959, 90% is 1.644) and the good thing about it is that it allows us to calculate the value that is, for example 90% better than all the previous existing values. And that sounds awesome to us, because we can assign points based on how well that player ranked compared to all the previous games with that same hero.
So, the system goes like this, suppose your score ranks somewhere between 0% and 100%, 0% meaning you were worse than EVERYONE ELSE that played with that hero and 100% meaning you were BETTER than everyone else, while 50% means that you were exactly in the middle (The average):
If you're rank between 0% and 10%, you get -3 points. If you rank between 10% and 25%, you get -1 points. If you rank between 25% and 50%, you get 0 points. If you rank between 50% and 75%, you get 1 point. If you rank between 75% and 90%, you get 3 points. If you rank between 90% and 100%, you get 5 points.
Finally, you get additional points depending on match length and whether your team won or lost. While this might be controversial, due to some teams being structured on winning late, this mainly helps offset low scores in stomps and it won't affect most matches in the 20-50min range. And pretty much no team is structured to win in more than 50 minutes.
The system goes like this:
If you won in less than 20 minutes, you get 5 points (As a bonus for stomping + the fact that your stats will probably not be that great, because you couldn't rack enough kills/assists/gpm) If you won in between 20 and 50 minutes, you get 2 points (A standard win, more points than a loss, nothing special) If you lost in less than 50 minutes, you get 1 points (Consider it a consolation prize) If you lost in more than 50 minutes, you get 5 points (A bonus for holding out for so long)
You can optionally also have a CAPTROLL player. The CAPTROLL in your team will get more points the WORSE he performs. The amount of points you get for your CAPTROLL is 15 minus his actual score.
I like their scoring system, but I wouldn't copy their buy/sell system--the original buy/sell system for DotA-Academy's league was made for TI2, with fixed player cost (since in a 1 week event, player value isn't likely to change that much) and 80% sell-back. A trade-based system makes more sense for an extended league like LCS, because I expect player value to fluctuate much more over a long period of time.
|
Win/Loss should have value, but picking up good players on "bad" teams is part of fantasy. It's like picking a star quarterback/running back on a bad team, because you know he'll give good numbers, even if his team is bad. League is only 5 players though, so if you are that good you should have more impact.
I like the idea of drafting a "coach" for win/loss. Since there aren't actual coaches in league it would be like drafting a team. Say I could draft EG as my "team", because I think they'll win a lot, but my team is members of Curse, TSM, and Gambit, SK, etc. because I think individually players on those teams are stronger. That way points for wins/loss would be on what team I drafted, but points for kills/etc would be on what players I drafted.
Dragon/Baron should be factors, and so should turrets gained/lost. Sometimes its worth giving up a dragon for a turret, etc. But where they should be assigned is difficult.
The biggest problem I think is there isn't a good stats tracking page for any of this. The LoLesports page basically has 3 stats. KDA, GP/Minute and Total gold. You have to go through the VoDs to get any other stats, and if value is going to be given to Dragons/Barons/Turrets/Double/Triples etc. then it is going to be very time consuming.
|
United States47024 Posts
I don't see why Baron/Dragon/Turrets need to have separate value. KDA have value because they are indicators of individual performance, and can signify a strong player on a weak team (he did well when his team lost). Baron/Dragon aren't indicators of anything like that. They are simply team objectives toward victory. I don't see why a team that got more Dragons before they lost should get more points than one that didn't get as many if it's not reflected by better performance via other stats (longer game length, better kill stats due to being able to do better in fights with more gold), etc.
|
Lineup-
Doublelift nRated sOaz Xpeke Snoopeh
|
On February 19 2013 04:42 TheYango wrote: I don't see why Baron/Dragon/Turrets need to have separate value. KDA have value because they are indicators of individual performance, and can signify a strong player on a weak team (he did well when his team lost). Baron/Dragon aren't indicators of anything like that. They are simply team objectives toward victory. I don't see why a team that got more Dragons before they lost should get more points than one that didn't get as many if it's not reflected by better performance via other stats (longer game length, better kill stats due to being able to do better in fights with more gold), etc.
Perhaps *solo* Baron/Dragon should count for points- rewarding Baron steals, dragon steals, solo dragoning, and the rare solo Baron, while not rewarding people who just showed up as the team killed it.
Thoughts?
|
I don't see how you can re-draft the lineup as player's salaries continue to go up, but your salary limit stays the same...
|
The problem is the costs are swinging wildly based on having a good game/bad game. Players costing between $1 and ~$35 is a huge difference. I think they expected a $15-25 variance but the scoring system has just caused this drastic differential. I believe the average player cost is still $20, but there's just this huge swing because its changing the costs so drastically.
|
United States47024 Posts
Costs should be normalized per role, IMO. A particular role doing well or poorly is going to cause some weird shifts in the cost distribution for players.
|
LOL I fucked myself in this league by not realizing how LCS works and only some teams play in a given week meaning I had like 3-4 players in week 2 that didnt even play hahaha (fnatic i think)
|
Yay 97 points last week, most in the TL league. :D
But I guess points gathered so far do not count towards the total right?
|
On February 19 2013 10:33 Redox wrote: Yay 97 points last week, most in the TL league. :D
But I guess points gathered so far do not count towards the total right? Nice job :D what was your team last week?
|
On February 19 2013 07:23 Slayer91 wrote: LOL I fucked myself in this league by not realizing how LCS works and only some teams play in a given week meaning I had like 3-4 players in week 2 that didnt even play hahaha (fnatic i think) Points didn't start accumulating til week 3 anyway.
|
|
|
|