*** VD 15 Commentary Thread ***

[Ken]

Do you have VDXIII or do you mean VDV?

I have XIII

[johnnya24]

I have XIII

Can you email?

... or any other stuff we are missing from the old drafts

First post: http://forum.rotojunkiefix.com/showt...history-Thread

[cavebird]

I take the opposite approach - I try to identify a replacement level player - knowing full well that I will not likely hit exactly the last player picked. But it's close enough. And then I use the objective analysis to compare. I do this across letters, across franchises, across positions, across eras. Some of them are small enough to disregard. And obviously the crossover isn't perfectly accounted for.

But I would say that yes, you can objectively measure this if you take into account some assumptions which are "close".

I think you could in certain drafts (letters only for example, it would be fairly easy to come up with replacement level at each letter), but it is trickier in others--once you get the second layer added on, decades last draft and franchises this draft, it becomes much harder to come up with a replacement level player that is close to accurate. These two-level ones tend to end up leaving a decent amount of really good players just sitting around undrafted because the letter and franchise or letter and decade just don't match. But yeah, theoretically, it could be done as long as you are willing to tolerate a decent margin of error.

[Ken]

I think you could in certain drafts (letters only for example, it would be fairly easy to come up with replacement level at each letter), but it is trickier in others--once you get the second layer added on, decades last draft and franchises this draft, it becomes much harder to come up with a replacement level player that is close to accurate. These two-level ones tend to end up leaving a decent amount of really good players just sitting around undrafted because the letter and franchise or letter and decade just don't match. But yeah, theoretically, it could be done as long as you are willing to tolerate a decent margin of error.

I guess the question is whether an objective approach with an acknowledged known margin of error is better than a subjective review. My experience is that it invariably is.

[Ken]

Can you email?

... or any other stuff we are missing from the old drafts

First post: http://forum.rotojunkiefix.com/showt...history-Thread

your PMs are full

[johnnya24]

your PMs are full

My PM's are no longer full.

[cavebird]

I guess the question is whether an objective approach with an acknowledged known margin of error is better than a subjective review. My experience is that it invariably is.

I think the great question is whether the margin of error is small enough to be useful---and the variance in it. For example, if all letters are off just a little bit, that would seem to be great. But if half of them are exactly right but the other half are off by a good margin, it might be far less useful. But certainly with all of your data sets known (i.e., it isn't a league for this year where players will do better or worse than projected--the stats are already there) there should be an objective way to rank things with pretty decent accuracy.

[Ken]

I think the great question is whether the margin of error is small enough to be useful---and the variance in it. For example, if all letters are off just a little bit, that would seem to be great. But if half of them are exactly right but the other half are off by a good margin, it might be far less useful. But certainly with all of your data sets known (i.e., it isn't a league for this year where players will do better or worse than projected--the stats are already there) there should be an objective way to rank things with pretty decent accuracy.

The error on an individual letter doesn't even matter though - you would need a particular letter to be significantly off relative to other letters. I think it evens out to the point where its close enough to be making the choices we are making.

[cavebird]

The error on an individual letter doesn't even matter though - you would need a particular letter to be significantly off relative to other letters. I think it evens out to the point where its close enough to be making the choices we are making.

Yes, that is the question. But given our data set, sometimes that can happen--for example with K last draft.

[cavebird]

Man that was a tough choice---after Walsh and Joss went, mostly everyone taken was not someone I was keying on. So I got greedy and went for the best chance that a second guy I wanted in round 3 would be around for round 4. I had five under consideration here, so with six picks in between and four left, my chances are....zero, lol.

[cavebird]

Just a heads-up: I am returning to in-person work tomorrow, so starting then, if my spot comes up mid-day I may be delayed til I get home in the evening. I expect to be able to select before going to bed at night, though.

Well, everyone is here through the turn, so hopefully we can get to you, but obviously no guarantees.

[jmaeroff]

Well, everyone is here through the turn, so hopefully we can get to you, but obviously no guarantees.

Right, may not be a factor until Rounds 5 & onward. Just wanted to put it out there now though. Today I'm still WFH.

[cavebird]

Man that was a tough choice---after Walsh and Joss went, mostly everyone taken was not someone I was keying on. So I got greedy and went for the best chance that a second guy I wanted in round 3 would be around for round 4. I had five under consideration here, so with six picks in between and four left, my chances are....zero, lol.

Okay, less than zero. Two for two so far. Nice picks guys. I wonder if Ken's guy made it to him.

[hacko]

I take the opposite approach - I try to identify a replacement level player - knowing full well that I will not likely hit exactly the last player picked. But it's close enough. And then I use the objective analysis to compare. I do this across letters, across franchises, across positions, across eras. Some of them are small enough to disregard. And obviously the crossover isn't perfectly accounted for.

But I would say that yes, you can objectively measure this if you take into account some assumptions which are "close".

I take a third approach- I take in consideration who looks the best in a uni
Who has a hot wife.
Who’s dog would bite Heye.
If anyone would want to be Frae neighbor.
Then I pick.

I think it needs tweak.

[Ken]

Okay, less than zero. Two for two so far. Nice picks guys. I wonder if Ken's guy made it to him.

Yes, Nathan was the guy I was hoping would slide to me. Now on to my second pick.