Nerd sheets here! (DT stats)

[Nipple Nibbler]

Who wants stats?
You want stats?
I got stats for you here!

Season:
https://docs.google.com/spreadsheets...it?usp=sharing

Playoffs here:
https://docs.google.com/spreadsheets...it?usp=sharing


If anyone sees any errors reach out, I will fix them ASAP. Wirah's DT website has his stats and bitsaver's API has his as well. Here is your chance Claushouse to get me back



<TIMESTAMPS>
I use the timestamps on the far right side of score boxes for consistency since this is where !subs !switches !changes are timestamped. This might be off by a second here or there due to bot limitations.

[WillBy]

I don't think the time stats on DTB are correct. In particular, how can demonic have 30 minutes for, but only be 1-1? By a quick check, they should have 29:58 time for, as they lost to Haze by 2 seconds. Haze's time against is similarly messed up- 30 minutes, when it should be 29:58. Not any big differences, but this could cascade for future weeks.

[Nipple Nibbler]

WillBy There is a few second delay in the timestamps in boxscores. The delay never seems to be the same so for consistency I just use what is in the scoreboxes. This is an old bot bug. In the TWL nerd sheets post I talked about it, I should have put it here too .

[Tiny]

hey nerd stats...add me to KO leader. ty for this

[Nipple Nibbler]

hey nerd stats...add me to KO leader. ty for this

Last time I did this manually, this is automated now and I fubar the formula

<FIXED>

[Nipple Nibbler]

~updated for week 2~

[Vehicle]

Isn't it kind of weird that the stats are biased to games that are closer? Unless I'm mistaken, minutes played are weighted positively towards your rating, but in basing I would think the least amount of minutes played should be rewarded? Wouldn't you be considered better if you blow out your team instead of have it be 14:30-15:00.

I also think it's more than likely the case that whoever has the most *k/dpg is also going to end up with the most *k/dpm, that's basically what it looks like now after two weeks. I think, at least for basing, the formula is wrong and should reward fewer minutes played rather than closer games and that change would also ideally be reflected in those other four categories.

Maybe I'm just too stoned, but that's my two cents.

[Nipple Nibbler]

Vehicle Long reply sorry.

I appreciate any input for the rating I am working on and I am putting the formulas used right in the visible sheets to be transparent. I want to keep it simple, so anyone can look at it and go, "hey that makes sense", even if they don't agree with it. But not so simple it only uses k/d. It certainly needs more thought put into it and nowhere near close for the basing ships. The hardest part with basing is keeping it so it is similar across all 3 ships to attempt to compare.

I divide the rating by time to get an idea of how well the players actually play each minute they play instead of just having someone get 130 kills simply because it was a 30 minute match. I give players with small amounts of time played a 0 rating to avoid someone subbing in and going 15-1 in like 2 minutes having a crazy high rating. As the season progresses I increase the minimum time played amount. I put pretty good weight into Teks to avoid campers sitting in the back stealing the show (like Henry). I put good weight into wins to avoid selfish players playing just for rec (again like Henry).

Obviously no matter how I weigh anything in the ratings it will opinion-based unlike the rest of the stuff in the sheets and was honestly just a way to sort the players. I did reach out to some Warbirds and Javelins for opinions on their ratings, and I am happy with it for them (at least for now).

With time in the equation it doesn't reward short or long games, just how well people consistently play. Per game leaders aren't always going to be (but probably will be most the time) the same as per minute unless you make it per 30 minutes. You might be right about rating players with shorter games higher up, but that wont always be true, sometimes you just play against a really crap team and get an easy win . I will make sure to set some time aside to work on this some more and hopefully by playoffs I will have something I am happy with for all 3 basing ships, any changes I make to the rating will be changed in any other sheets I have done as well, like TWL, so it will be consistent.

Any basers with honest non-trolly opinions on how heavy to weigh the stats used for basing feel free to reach out. Initially when I started nobody seemed to care lol.

[Efhat]

Isn't it kind of weird that the stats are biased to games that are closer? Unless I'm mistaken, minutes played are weighted positively towards your rating, but in basing I would think the least amount of minutes played should be rewarded? Wouldn't you be considered better if you blow out your team instead of have it be 14:30-15:00.

I also think it's more than likely the case that whoever has the most *k/dpg is also going to end up with the most *k/dpm, that's basically what it looks like now after two weeks. I think, at least for basing, the formula is wrong and should reward fewer minutes played rather than closer games and that change would also ideally be reflected in those other four categories.

Maybe I'm just too stoned, but that's my two cents.

how is the formula wrong?
i have more kills and less minutes played (10 or more minutes) than the other 9 spids in the top 10... the kpg/kpm seems on point.
for example, autopilot has 10 more minutes played than me but hes #2 in the kill race at 445 and I’m at 511..
im sure if I played the same amount of time he has in basing the kill gap would be dramatically more spread if we stayed at that same pace.
which I would believe is why I lead the kpg? Unless RR can correct me if I’m wrong?

[Vehicle]

Nipple Nibbler Efhat I may be wrong this is just how I was reading it. I am by no means smart when it comes to this stuff. And I appreciate all the time and effort you've put into it RR.

It just seems to me that basically all the stats except for the per minute ones are reward longer games. But I guess you're right that you do work some stuff in to counteract that.

I wasn't making an attack on your stat sheet I was just asking an honest question / opening up the discussion. After reading your response, I think this is your best solution to the problem I mentioned and it may even be the only solution.

Like I said, I'm dumb when it comes to statistics, so my post wasn't meant as some veiled insult.

[Kassius]

Any chance you can unhide the raw data tab(s)?

[Vehicle]

how is the formula wrong?

Because at its core a dominant basing performance is one in which fewer minutes are played. I don't know how you fix statistics to reflect that but rewarding "minutes played" is rewarding close games.

[Nipple Nibbler]

Efhat
KPG is just kills divided by games played, KPM is kills divided by minutes played. So longer games means you have a chance at having higher KPG. Your personal KPM won't really change much if you're consistent. Every stat is completely raw stats unbiased except for the rating stat, which is a formula I am trying to come up with. A better example of comparing KPM and KPG would be Cripple and Ease both have almost the same kills per game (Cripple leading by a pubic hair), but Cripple nets almost a full extra kill per minute. If Cripple played longer games his KPG would be a lot higher as he is only averaging just under 24 minutes per game, while ease had almost a 29 minute game so he had 5 extra minutes to get higher KPG. Out of the two KPM is a better stat to go with, but some people like KPG so I put it there.

There is no one raw stat to see who is the best, so I try to have as many stats as I can in there to paint a bigger picture the best I can so people can form their own opinions. The closest we can come to finding the best player is to come up with a rating formula that is well balanced, but that again will be opinion based, intentionally or not.

Vehicle I didn't take offence to your post, I was only trying to explain my thought process and the stats in question. I literally break stuff for a living, which is one of the reasons I am trying to make it a simple formula for rating so I can follow it too . I COULD add a small flat bonus to ratings for having victories under x amount of minutes played, it can be done. The question becomes if I should. It be nice to have some more opinions on it before I do. The KPM stat rewards players with high amount of kills in shorter games, which is what Cripple has. That stat alone might be what you are looking for.

Kassius Do you mean where I put in all the boxscores data? Yes I can unhide that, in TWL it was 3 sheets, this time I managed to cram all 3 divisions into the same sheet, it is a little messy, because it wasn't really meant for people to view. It is still better than just boxscores crammed together I guess. You could hit ctrl+f to search for games with someones name in it. The sheet name is simply DATA. Be sure to let me know if you find any typos.




If anyone has any other questions, don't hesitate to ask.

[Kassius]

Thanks nipple, I wanted to see time played for individual games, which tour sheets show but the main site scoreboxes omit.

[Efhat]

Efhat
KPG is just kills divided by games played, KPM is kills divided by minutes played. So longer games means you have a chance at having higher KPG. Your personal KPM won't really change much if you're consistent. Every stat is completely raw stats unbiased except for the rating stat, which is a formula I am trying to come up with. A better example of comparing KPM and KPG would be Cripple and Ease both have almost the same kills per game (Cripple leading by a pubic hair), but Cripple nets almost a full extra kill per minute. If Cripple played longer games his KPG would be a lot higher as he is only averaging just under 24 minutes per game, while ease had almost a 29 minute game so he had 5 extra minutes to get higher KPG. Out of the two KPM is a better stat to go with, but some people like KPG so I put it there.

There is no one raw stat to see who is the best, so I try to have as many stats as I can in there to paint a bigger picture the best I can so people can form their own opinions. The closest we can come to finding the best player is to come up with a rating formula that is well balanced, but that again will be opinion based, intentionally or not.

Vehicle I didn't take offence to your post, I was only trying to explain my thought process and the stats in question. I literally break stuff for a living, which is one of the reasons I am trying to make it a simple formula for rating so I can follow it too . I COULD add a small flat bonus to ratings for having victories under x amount of minutes played, it can be done. The question becomes if I should. It be nice to have some more opinions on it before I do. The KPM stat rewards players with high amount of kills in shorter games, which is what Cripple has. That stat alone might be what you are looking for.

Kassius Do you mean where I put in all the boxscores data? Yes I can unhide that, in TWL it was 3 sheets, this time I managed to cram all 3 divisions into the same sheet, it is a little messy, because it wasn't really meant for people to view. It is still better than just boxscores crammed together I guess. You could hit ctrl+f to search for games with someones name in it. The sheet name is simply DATA. Be sure to let me know if you find any typos.




If anyone has any other questions, don't hesitate to ask.

But Ease has one game played as spider
I have four.
Isnt that a contribution factor too?