Avoid Interception Calculator
 

i put my avoid interception calculator online. enjoy

in the spirit of fof, the interface doesn't look pretty but the underlying code should make your life easier.

sites.google.com/view/tzach/avoid-interception-calculator

this is based on a calibration that i developed using custom draft classes, where i put by hand the rookie bars so i know more or less how they would come out.

I always understood that avoid interception bar correlates to Solecismic Test and Intelligence Scores, nothing to do with the rest?

It's a calculator that uses the bars and overall ratings and determines if the QB is good or bad at avoiding INTs.

When a QB has huge big bars, but is rated 70 then you'll get -11 or whatever. He's bad at Avoiding INT.

When a QB has low bars but is rated 55, he'll get a +10. He's good at avoid INTs.

Do you have any idea if yours is different from Squirrel's calculator?

I see, thanks Garion :)

Oh oh oh, tzach's is different than Squirrel's calculator.

Squirrel's calculator gave a +/- output, which is what my earlier comment relates to. (+ score is good, - score is bad.)

tzach's actually estimates the bar (so, 0-100). That's a neat change and makes it easier to read.

Very cool tzach... A couple Qs if you don't mind my skepticism

So I am guessing you ran the regression to determine how overall is calculated. But what about Avoid Fumbles?

Is that the same bar or not considered in overall. Anyways for estimating purposes it shouldn't change how useful this is.

sorry for the delay garion. they follow the same principle. squirrel's calculator will give you a 'delta', which is the difference between the overall rating implied by the visible bars and the overall rating given by a scout. going from delta to how good a player is depends on overall rating though.


my calculation is a bit different as it calibrates how much avoid AI matters for overall rating, so i can estimate a 'bar' that goes from 0 to 100. i assume that avoid fumbles is not important for qbs.

yup!

sorry for the delay -- i assume that avoid fumbles is not relevant for qbs, so it's not considered at all for overall ratings and calibrating avoid interceptions.

I've added the calculation to player pages in my leagues so that it'll show up automatically.



Example: http://www.fof-ihof.com/playercard.php?playerid=64825



Looking forward to the complaints that either tzach or I have it incorrect when some 45/45 QB with 65 AvInt throws 20 picks in a year. :D

good point ben -- here's my disclaimer from ihof :D



'we should have in mind that avoid AI is only one of the factors that matters for INTs. game plan, passing cohesion and supporting cast/execution matters too, probably as much as avoid AI. As do other QB bars (like read defense, which controls how many times the QB throws into double coverage). so please check if your team fulfill the requirements above before complaining!'

for anyone else that wants to implement in their own way, here's the calculation


AvoidInt = (Cur*1.000 - Screen*0.082 - Short*0.078 - Med*0.091 - Long*0.091 - Deep*0.062 - ThirdDown*0.099 - Acc*0.095 - SenseRush*0.033 - ReadD*0.045 - Twomin*0.029 )/0.296

Props on putting that out there.

Though this means Jim is certain to change things up for the next released. ;)

tzach, why is the timing bar not included in the formula?

I have noticed that the timing bar is also not included in Squirrels calculator as well regarding the weighted bars. I'm curious to why this is and if the timing bar should be less or more of a priority when evaluating a quarterback based on this.

Timing is a tendency bar doesn't affect overall rating

One other silly question. And I understand the calculator only provides an estimate of the avoid INT bar. Are there certain QBs which are outliers where the formula does not work well? We have a QB in the TFL with an estimated 113 AI bar.

The QB is fully developed and the numbers were verified by another GM. So, it would appear that is not scout error in how the bars are viewed.

i'm sure there is some outliers -- my calibration is still an estimate. jim would probably laugh!



i'd take 113 as 100 -- ie he's pretty good.



timing doesn't count for overall ratings so that's why it's not included.



can you post screenshots of the qb ratings ? also, what are the oc/hc scouting bars?

I listed the bar ratings below. As mentioned, in an effort to minimize scout error as a variable, I compared the bar view with another GM. Both of those scout ratings are below as well. The 2nd OC/HC scouting is excellent. And both scouts arrived at the 113 avoid INT bar.

Overall 87
Screen 1
Short 100
Medium 100
Long 84
Deep 94
3D 73
Accuracy 99
SR 72
Read Def 87
2-Minute 47

OC 58
HC 62

OC 79
HC 82

thanks -- i calibrated with OC and HC 95 or so. i don't think it's scouting error though. for now we will have to take guys like these as maxed out in avoid INT. a 10 or 15% error in avoid AI as estimated by the calculation is still pretty good for strategy purposes.

100 doesn't necessarily mean 100 in FOF...

0 also doesn't mean 0.

This thing tzach has done is a better calc than what I had for various reasons. I would hesitate before taking too exactly the 0-100 scale it implies, as tzach has tried to explain. But that shouldn't devalue the calc in any way.

This is super helpful and something that might help encourage more parity in MP FOF leagues. As will the chemistry traits thing posted elsewhere here.

Now we just need to do something about the 'finance' system in the game and the wretched staff draft! Perhaps if we could come up with a change to introduce more randomness to the distribution of good scouting we might be able to bring more parity still

If a tool like conscriptor existed for staff draft, I would just put an upper limit to hc scouting of70 and oc/DC of 50 or so. This would make drafting and roster evaluation interesting.


I do think it would be useful to have an optional tick box to make coaches uniform for all teams, similarly to the economic factors.

Side conversation here, but MOS DEF

Where can I find this program? The links are dead.

This?

https://sites.google.com/view/tzach/...tor?authuser=0

Thanks