Warrior General Games

Originally posted by tautology
Given these numbers, I would say something like Clutch is better *if* you think your base chance is extremely low. Clutch could give you a substantial bonus, albeit not terribly often.
Yeah, it's entirely possible that Clutch could be better than 3DS, but impossible for any of us to say for sure. Then Long Reach is boosted by the fact that 60-70% of pass attempts happen on 1st or 2nd down.

Originally posted by tautology
Also, your chance of getting the PD if you fail your vision check or if you are out of position is as follows:

Base
Chance____3DS________LR________Both
0% ________0%________0%______0%

Not entirely true, as deflections and interceptions happen along the pass plane, so you don't have to be beside the receive to get an attempt, just along that plane at a point where the player is able to reach the ball. That's why bullet passes on streak routes are so awful this season. But sure, Ball Hawk would seem to have a lot of value, while it's much harder to guess how much an extra 6 points in vision affects your vision check. Presumably it's significant, but we'd have to know more in order to make any mathematical comparison.

Originally posted by jdbolick

60-70% of pass attempts happen on 1st or 2nd down.

MPHD begs to differ

So we're using made up numbers to try and prove something that can't be proven unless we know what the real numbers are?

That being said: BH, TDS, FG, HJ, Clutch, PTP.

Originally posted by Worker 3
Originally posted by Hate Sighed


5. Prime Time Player http://goallineblitz.com/game/replay.pl?game_id=1275453&pbp_id=15044847

PTP wouldnt have activated on that playo_O


Yes. It works when the player is in the endzone.

Originally posted by Longhornfan1024
So we're using made up numbers to try and prove something that can't be proven unless we know what the real numbers are?
No, but you're precisely the type of superstitious, irrational influence in the CB sub-forum that I was talking about earlier. If the VA descriptions are accurate then we know what those equations are, we just don't know what the base chance for deflections may be. We can figure out a general average by looking at PDs vs. pass attempts, but it's probably a variable figure anyway, which is why scales like what tautology posted are useful.

Originally posted by jdbolick
Originally posted by Longhornfan1024

So we're using made up numbers to try and prove something that can't be proven unless we know what the real numbers are?
No, but you're precisely the type of superstitious, irrational influence in the CB sub-forum that I was talking about earlier. If the VA descriptions are accurate then we know what those equations are, we just don't know what the base chance for deflections may be. We can figure out a general average by looking at PDs vs. pass attempts, but it's probably a variable figure anyway, which is why scales like what tautology posted are useful.


A critical question:

Does 3DS actually give a % increase to your bare % ability to deflect a given ball (which is how it reads), or does it give a % bonus to your pass Deflection Quotient (which presumably is opposed by Pass Quality and the receiver's Anti-Swat abilities).

Because I doubt that the resultant comparison of these things is on a linear scale, and this could make quite a difference.


Also: Does a better PD score result in a less catchable deflection? If so, this gives a dramatic edge to 3DS, as so many deflections are caught anyways....

Originally posted by tautology
Originally posted by jdbolick

Originally posted by Longhornfan1024


So we're using made up numbers to try and prove something that can't be proven unless we know what the real numbers are?
No, but you're precisely the type of superstitious, irrational influence in the CB sub-forum that I was talking about earlier. If the VA descriptions are accurate then we know what those equations are, we just don't know what the base chance for deflections may be. We can figure out a general average by looking at PDs vs. pass attempts, but it's probably a variable figure anyway, which is why scales like what tautology posted are useful.


A critical question:

Does 3DS actually give a % increase to your bare % ability to deflect a given ball (which is how it reads), or does it give a % bonus to your pass Deflection Quotient (which presumably is opposed by Pass Quality and the receiver's Anti-Swat abilities).

Because I doubt that the resultant comparison of these things is on a linear scale, and this could make quite a difference.


Also: Does a better PD score result in a less catchable deflection? If so, this gives a dramatic edge to 3DS, as so many deflections are caught anyways....


I'm pretty sure it's to the Deflection quotient, or score, which is then used in a competitive roll.

The math is interesting to look at if you assume a competitive roll like hte following:
Deflection score X
Pass Quality score Y

Roll for deflect where X+Y is the probability range (meaning normalize X+Y to 1 and roll from 0 to 1.)

Then look at things like +45% X and what that does to the results. The short answer is - always less than 45% more pass deflections. For very low X/Y (meaning the Deflection score is small compared to the pass quality), it asymptotically approaches 45% more. For very large X, the 45% boost yields 0 more deflections.

And I don't even know that that's the competitive roll model he uses, though something like it probably makes sense.

Originally posted by kurieg
Originally posted by tautology

Originally posted by jdbolick


Originally posted by Longhornfan1024



So we're using made up numbers to try and prove something that can't be proven unless we know what the real numbers are?
No, but you're precisely the type of superstitious, irrational influence in the CB sub-forum that I was talking about earlier. If the VA descriptions are accurate then we know what those equations are, we just don't know what the base chance for deflections may be. We can figure out a general average by looking at PDs vs. pass attempts, but it's probably a variable figure anyway, which is why scales like what tautology posted are useful.


A critical question:

Does 3DS actually give a % increase to your bare % ability to deflect a given ball (which is how it reads), or does it give a % bonus to your pass Deflection Quotient (which presumably is opposed by Pass Quality and the receiver's Anti-Swat abilities).

Because I doubt that the resultant comparison of these things is on a linear scale, and this could make quite a difference.


Also: Does a better PD score result in a less catchable deflection? If so, this gives a dramatic edge to 3DS, as so many deflections are caught anyways....


I'm pretty sure it's to the Deflection quotient, or score, which is then used in a competitive roll.

The math is interesting to look at if you assume a competitive roll like hte following:
Deflection score X
Pass Quality score Y

Roll for deflect where X+Y is the probability range (meaning normalize X+Y to 1 and roll from 0 to 1.)

Then look at things like +45% X and what that does to the results. The short answer is - always less than 45% more pass deflections. For very low X/Y (meaning the Deflection score is small compared to the pass quality), it asymptotically approaches 45% more. For very large X, the 45% boost yields 0 more deflections.

And I don't even know that that's the competitive roll model he uses, though something like it probably makes sense.

Yep...something like this is how I have always imagined that it works. Similar for tackle/break tackle too I would guess. A model like this makes sense to me from what I have seen in the sim over all these seasons.

Do the KL rolls work the same?

Originally posted by jdbolick

No, but you're precisely the type of superstitious, irrational influence in the CB sub-forum that I was talking about earlier. If the VA descriptions are accurate then we know what those equations are, we just don't know what the base chance for deflections may be. We can figure out a general average by looking at PDs vs. pass attempts, but it's probably a variable figure anyway, which is why scales like what tautology posted are useful.


I'm not sure how you could label me as superstitious or irrational. The issue is simply that LR is a VA that requires the CB to fail a roll, meaning that it will rarely ever trigger except on plays where the CB has a poor PD score to begin with. On top of that, it's only a 45% chance of re-rolling; the re-roll isn't guaranteed and the PD isn't guaranteed to succeed even if you hit your re-roll. Since you've already failed the roll the first time around, odds are that you're going to miss it the second time around.

On the other side, you have TDS which is limited to 3rd and 4th down, but gives 45% bonus to your PD score on every passing play on those downs. If your team is organized, you can run TDS on your starters and force them on third down. So you have a guaranteed bonus to your PD score on the plays where you will most likely see a pass versus having a less than 50% chance to re-roll a roll that you are likely to lose the second time anyways.

Originally posted by Longhornfan1024
Originally posted by jdbolick


No, but you're precisely the type of superstitious, irrational influence in the CB sub-forum that I was talking about earlier. If the VA descriptions are accurate then we know what those equations are, we just don't know what the base chance for deflections may be. We can figure out a general average by looking at PDs vs. pass attempts, but it's probably a variable figure anyway, which is why scales like what tautology posted are useful.


I'm not sure how you could label me as superstitious or irrational. The issue is simply that LR is a VA that requires the CB to fail a roll, meaning that it will rarely ever trigger except on plays where the CB has a poor PD score to begin with. On top of that, it's only a 45% chance of re-rolling; the re-roll isn't guaranteed and the PD isn't guaranteed to succeed even if you hit your re-roll. Since you've already failed the roll the first time around, odds are that you're going to miss it the second time around.

On the other side, you have TDS which is limited to 3rd and 4th down, but gives 45% bonus to your PD score on every passing play on those downs. If your team is organized, you can run TDS on your starters and force them on third down. So you have a guaranteed bonus to your PD score on the plays where you will most likely see a pass versus having a less than 50% chance to re-roll a roll that you are likely to lose the second time anyways.

LH1024:

There are several arguments to favor 3DS over LR for a variety of reasons, but you aren't making any of them.

Your anecdotal analysis of "Since you've already failed the roll the first time around, odds are that you're going to miss it the second time around" are absolutely accounted for (and refuted) in the table.

Chance____3DS________LR________Both
5% ________7%________7.1%______10.3%
10%_______15%_______14.1%______20.1%
15%_______22%_______20.7%______29.4%
20%_______29%_______27.2%______38.3%
25%_______36%_______33.4%______46.6%
30%_______44%_______39.5%______54.6%


If we expand our thinking to other ways in which the PD roll works (ways which are in somewhat of a contradiction to the text of the VA) and speculate on a more "robust" deflection with a higher PD% score, then the debate is furthered. But denying the simple math in this table with incorrect statements gets us nowhere.

Now please carry on building excellent CBs, as you always do

Originally posted by tautology
LH1024:

There are several arguments to favor 3DS over LR for a variety of reasons, but you aren't making any of them.

Your anecdotal analysis of "Since you've already failed the roll the first time around, odds are that you're going to miss it the second time around" are absolutely accounted for (and refuted) in the table.

Chance____3DS________LR________Both
5% ________7%________7.1%______10.3%
10%_______15%_______14.1%______20.1%
15%_______22%_______20.7%______29.4%
20%_______29%_______27.2%______38.3%
25%_______36%_______33.4%______46.6%
30%_______44%_______39.5%______54.6%


If we expand our thinking to other ways in which the PD roll works (ways which are in somewhat of a contradiction to the text of the VA) and speculate on a more "robust" deflection with a higher PD% score, then the debate is furthered. But denying the simple math in this table with incorrect statements gets us nowhere.

Now please carry on building excellent CBs, as you always do


How are you assuming the PD score works when you input your base % chances in that table? From previous Bort posts we know that Jumping + Vision + SDC (flat bonus) + SB (flat bonus) = the CB's PD score. From there you add the % bonus given from AEQ and TDS, if using it. That score is then rolled versus the WR's catch ball score, both likely being modified by the pass quality, RNG, and positioning (this one is questionable). There is no flat % chance at the PD that a CB has on every play that would be modified by the two VAs to get a higher % chance. Instead TDS gives a % bonus that increases in value the higher the CB's PD score gets. LR allows the CB the retry the roll, but that essentially means having another chance at the RNG. Assuming that pass quality and positioning have an impact on the success of the roll, re-rolling wouldn't make any changes to those two inputs.

I'd also like to know how you are accounting for LR only being a 45% chance to trigger and how you account for my "Since you've already failed the roll the first time around, odds are that you're going to miss it the second time around" argument in the table.

Now please carry on doing crazy complicated math as you always do

I guess it still comes down to building a better dot through base attributes and EQ distribution. VA's are something that can be argued, sure. Unless the "base" dot is worth a crap, its all moot.

Originally posted by tautology
A critical question: Does 3DS actually give a % increase to your bare % ability to deflect a given ball (which is how it reads), or does it give a % bonus to your pass Deflection Quotient (which presumably is opposed by Pass Quality and the receiver's Anti-Swat abilities).
I'm not sure how much that would matter. It's true that if so you would have 3DS modifying only X whereas LR would interact with the RNG, but if I'm doing the algebra correctly then it doesn't look like that should produce a huge difference in the results, obviously depending upon the values of X & Y.

Originally posted by
Also: Does a better PD score result in a less catchable deflection? If so, this gives a dramatic edge to 3DS, as so many deflections are caught anyways....
Given that a lot of those caught deflections happen on 3rd down against guys who already have 3DS, I'm doubting that it makes much of an impact, but it's possible.

----------------------------------------

Originally posted by Longhornfan1024
I'm not sure how you could label me as superstitious or irrational.
Because you and the other denizens of the CB sub-forum argue from the basis of anecdote and feeling, not sound logic or mathematics. You dismiss LR because you don't feel like it has a pronounced effect, but while your feeling may or may not be correct, it's not useful to a legitimate discussion of the merits. That's not to say that experience is irrelevant, as it often clues us in to various truths, but it can also be extremely misleading. GLB history is littered with ridiculous notions that we look back and ridicule, but which a lot if not most of the player base felt at the time. By sticking to mathematics, we can be objective instead of subjective. Of course, that goes back to depending on the VA descriptions correctly representing these mathematical equations, however, which is no sure thing.

Originally posted by
The issue is simply that LR is a VA that requires the CB to fail a roll, meaning that it will rarely ever trigger except on plays where the CB has a poor PD score to begin with.
That's not the issue at all, but you're showing precisely what I mean about your thinking being irrational. The fact that you failed the roll is pretty much irrelevant. In theory you wouldn't need 3DS either unless your unmodified result was too low to successfully deflect the pass, so that's the same scenario. The question is how much the two VAs modify that score, and the mathematics indicate that there isn't an enormous amount of difference between the two. Given that LR would apply much more often, that would mean that it has the greater overall impact.

Originally posted by Longhornfan1024
How are you assuming the PD score works when you input your base % chances in that table? From previous Bort posts we know that Jumping + Vision + SDC (flat bonus) + SB (flat bonus) = the CB's PD score. From there you add the % bonus given from AEQ and TDS, if using it. That score is then rolled versus the WR's catch ball score, both likely being modified by the pass quality, RNG, and positioning (this one is questionable).
None of that except the RNG matters to the equations because they're all constant. The CB's jumping, vision, SDC, Swat Ball, and deflect chance AEQ all stay the same from one roll to the next, so all of that can be simplified to some hypothetical base chance for a deflection, or "deflection score" if you prefer kurieg's terminology. Neither Long Reach or 3rd Down Stopper modify only one part of that collection of factors, they modify the entire collection. Therefore the individual parts of the collection don't need to be addressed or noted separately. They can be grouped into the variable X, from which we can then apply these modifiers.

Originally posted by
I'd also like to know how you are accounting for LR only being a 45% chance to trigger and how you account for my "Since you've already failed the roll the first time around, odds are that you're going to miss it the second time around" argument in the table.
If you're familiar with algebra, then you'll realize that what you're describing is what the bold portion of the equation means: Chance of deflection with 15 in LR = (1 - (1 - X%) * (1 - (0.45 * X%))) I just don't think you're grasping the mathematics here sufficiently to understand that you're asking questions that don't even need to be answered because they're self-evident to anyone who understands the equation.

I officially feel dumb.