Jump to content

The MVP race


Steelersfan43

Recommended Posts

56 minutes ago, Duluther said:

Honest question: Who the HFIL cares who the most valuable player is? That's such a weird concept. I'd care a lot more about a most outstanding player.

Finding the most valuable player is like finding the most valuable employee at a selection of companies. It'd always be the CEO because they steer the ship just like the QB. But wtf would care about an MVP for CEOs?

Just do a Most Outstanding Player award.

Just a Player of the Year award?
Otherwise it’s always going to be the QB with the most eye popping stats if you go by ‘valuable’.

Link to comment
Share on other sites

On 11/18/2023 at 11:24 PM, Soggust said:

Hey guys - plz don't ban me if this is against forum rules (I don't think it is bc I'm obviously not making money or trying to drive traffic away. In fact, I'm on a free plan so I probably can't handle too much traffic lol). I literally just made this for the forum bc I was bored.

But anyway, I threw together a web app last night to try to help rank the QBs / decide MVP and I thought others might enjoy checking it out as it's kind of coming together. You might want to zoom your browser out a bit as I haven't really done much responsiveness or really much of anything lol. Don't judge me.

2023 QB Ranker (qbrankings-6719e.firebaseapp.com)

The basic idea is that I feel like we always apply criteria to one player, but we never seem to weight it the same across everyone. So, I tried to make an unbiased-ish ranker based upon a custom algorithm. 

Basically, I just took some super basic stats and normalized them to where the value 1 would be the leader in that stat and everyone else would be a fraction of that according to performance. Then, I added them all and subtracted the negative values (turnovers, biases) to create a composite score.

I feel like maybe it's kind of close to what I'm thinking but I'm also a total moron so idk. It's truthfully hard to tell because it's one of those things where I'm probably too stupid to know how stupid I am. But check it out! 

p.s. - oh and before "omgz you think my [coach,weapons,o-line] are [good/bad/neutral]? Do you even watch [my team]?" - no i don't. I'm an MMA and Fantasy Football guy. These scores are what I pulled out of my rear at 3am. But you can change them by clicking on them. You might want to disable the stat first so it doesn't fly around on you.

Really cool thing and thanks for creating it. Out of interest how do you score recency bias and support bias. That seems hard criteria to quanitfy

  • Like 1
Link to comment
Share on other sites

4 hours ago, Brit Pack said:

Really cool thing and thanks for creating it. Out of interest how do you score recency bias and support bias. That seems hard criteria to quanitfy

ty friend! 

so the current calculation is 

Spoiler
      recordWeighted +
      +(
        // Increasing by 100% to increase severity of slider
        (compPercentWeighted * 2)
      ) +
      totaYardsWeighted +
      ypaWeighted +
      ratingWeighted * 1.25 + // Increasing by 25% to increase severity of slider
      totalTdWeighted * 1.25 - // Increasing by 25% to increase severity of slider
      turnoverWeighted -
      (weaponsWeighted +
        coachingWeighted +
        olineWeighted +
        defenseWeighted -
        responsibilityWeighted * 1.25) *
        +(
          // Increasing by 25% to increase severity of slider
          (this.filters.supportValue / 150)
        ) + // Dividing By 150 instead of 100  to reduce the severity of slider
      recencyWeighted;

The reason for some additional weighting is that the difference between #1 vs #30, for example, in record is inherently more weighted than completion percentage, for example. In that specific case, maybe that's a feature not a bug but it happens to a lesser degree across all stats.

I think the right answer is to get the lowest value of everyone and do something like (statValue + lowestPlayerValue) / (statValue + maxPlayerValue), but I have my daughter this week so I haven't gotten around to playing with it. Therefore, the quick answer was to just increase the slider's effect a bit.

---

So how are we weighing recency? 

Basically, most recent game = 100% weighted, next 90%, 80%, 70% then 60% (W = 1, L = 0). Then, that sum is totaled and normalized to 1. 

----

What about support? 

Well support on its own is just a slider that controls the 5 values (weapons,coaching,oline,defense,responsibility) concurrently. Those values individually are given an initial value of 0 (F) through 12 (A+) and then normalized to 1 (responsibility is a positive for the QB, everything else is a negative). These are all changeable by the user, so I am not too concerned with being correct on initial analysis.

----

It's not close to perfect at all, but in a week or two I think it might be pretty decent. I still need to correct the min value issue in a better way and clean up the code to make it presentable without completely embarrassing myself, but then I might just open up the repo to reddit or something and let smarter people fix it lol.

 

----------

Edit - also just to keep this post on topic - the algorithm currently has:

1) Dak
2) Jalen
3) Brock
4) CJ
5) Lamar 

And I don't know that I hate that list right now lol
 

Edited by Soggust
  • Like 1
Link to comment
Share on other sites

4 hours ago, Steelersfan43 said:

 

Brock Purdy can't be MVP because look at his weapons, coaching and defense.

So, instead our leading candidates are -

1) Jalen Hurts (Top 5 WR, Arguable best #2 in NFL, Best Oline, Great Run Game, Sirianni 33-13 as a HC)
2) Patrick Mahomes (Arguable GOAT TE, Arguable GOAT Offensive Mind at HC, Solid Oline, Great Defense)
3) Dak Prescott (Top 5 WR, Solid Run Game, Solid Oline, Great Defense) 
4) Lamar Jackson (Top 3 TE, HoF HC, Good Oline, Great Defense)
5) Tua Tagovailoa (Top 1 WR, Arguable best #2 in NFL, Top 5 offensive mind at HC, Solid Oline) 

Link to comment
Share on other sites

9 minutes ago, Soggust said:

Brock Purdy can't be MVP because look at his weapons, coaching and defense.

So, instead our leading candidates are -

1) Jalen Hurts (Top 5 WR, Arguable best #2 in NFL, Best Oline, Great Run Game, Sirianni 33-13 as a HC)
2) Patrick Mahomes (Arguable GOAT TE, Arguable GOAT Offensive Mind at HC, Solid Oline, Great Defense)
3) Dak Prescott (Top 5 WR, Solid Run Game, Solid Oline, Great Defense) 
4) Lamar Jackson (Top 3 TE, HoF HC, Good Oline, Great Defense)
5) Tua Tagovailoa (Top 1 WR, Arguable best #2 in NFL, Top 5 offensive mind at HC, Solid Oline) 

What? So Brady shouldn’t have won 07 MVP since he had prime Moss 

Link to comment
Share on other sites

Imagine thinking Trey Lance was going to take Dak’s job this season. The predictions of which week it would happen this season was comical. Now Dak should be the current clear front runner in the MVP conversation. It’ll be interesting to see how he can fare in December with a challenging schedule, but the man has been shredding out there.

  • Like 1
Link to comment
Share on other sites

1 hour ago, NoFlyZone said:

Imagine thinking Trey Lance was going to take Dak’s job this season. The predictions of which week it would happen this season was comical. Now Dak should be the current clear front runner in the MVP conversation. It’ll be interesting to see how he can fare in December with a challenging schedule, but the man has been shredding out there.

We don’t see eye to eye often but agreed. Dak is #1. Hurts close behind. 

Edited by Tank4Drake
  • Like 1
Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...