Small Game Numbers

[Doc]

I just got my latest edition of New York Outdoor News and was reading about how NY's small game hunting is getting smaller. That was where I found out how the DEC wants us to believe that they know how many rabbits, squirrels, grouse, and other small game critters that exist. Check out the accuacy of some of these numbers that they publish:

Talking about rabbits, they have these amazing numbers: "Last seasons kill of 96,742 was well below the 5 year average of 123,738 although up slightly 2010-11's talley of 95,809."

Notice the precision of those numbers. And this, a species that has no requirement for harvest reporting.

Check these numbers on squirrels: "New Yorkers bagged an estimated 187,350 squirrels last season slightly higher than the previous season (187,350) but below the five year average of 226,061 and well down from the 2008-09 high of 315,947"

They go on to give other examples of precise counts on coons, red foxes, gray foxes, coyotes, phesants, varying hare, and grouse. They also gave some numbers on participating hunters and how many hours they spent hunting these species.

While all this stuff may make interesting reading, is there anyone who places any credibilty in these numbers at all? I mean really .... Coming up with a squirrel count? They don't even have the luxury of mandatory harvest reporting (thankfully), and yet they publish numbers like this? Oh I understand the magic that supposedly is in statistics, but really, just how far are we supposed to buy in to this hocus-pocus? This borders on a sales pitch from snake-oil salesmen. But then who is going to refute those numbers .... lol. And yet, limits and seasons will be determined based on those studies, and who knows what other decisions get based on this stuff.

And then let's get back to a subject that represents the largest management activity in the hunting world, which is the estimation of deer season results. Whatever black art they are claiming for their small game numbers is the same thing they use to come up with deer populations and densities. I have always had a real problem assigning any level of credibilty to their procedures. But after reading this article on small game populations, I'm afraid that whatever potential credibility that I might have had is gone now.

So what's the deal? Do any of you swallow this stuff? I know we are expected to, and there is not a one of us that has the resources to disprove any of it. But at some point you have to start wondering if they aren't simply treating us all like fools.

[growalot]

direct hit....

[landtracdeerhunter]

Doc, do they even send surveys out to get some feel on numbers? Never have heard of any. Think this stuff is invented to give employees work. Maybe they have to many employees?

[Doc]

Doc, do they even send surveys out to get some feel on numbers? Never have heard of any. Think this stuff is invented to give employees work. Maybe they have to many employees?

I'm sure they send some surveys out to somebody, although I have not really heard of too many people who have ever received one. But even at that, given the tiny samplings, and wide variations of habitat and other regional situations, it is ridiculous to base anything on numbers gathered in that way, especially when it comes to things like squirrels and rabbits. It's ridiculous to even report those wild guesses. I understand the importance of statistics, and I apologize to all whose careers are in the field of statistics, but I do believe in the saying that "figures don't lie, but liars figure". I think we all are so mystified by the field of statistics that we are primed to believe anything that comes our way that is derived through statistics. But this nonsense has gone over the line of believability. This article on small game population numbers looks like an example of the abuse of statistics and has gone way too far to receive any sort of credibility. Unfortunately since the same organization has used these same techniques in other areas of wildlife management that we have grown to accept, it casts doubt on all of that as well.

[phade]

While I can't say anything for the DEC, I have found a significant respect for experts in anlaytics. I recently took a promotion to work on a team within my company that performs analytics. Although I do the qualitative portion of the work, I see what they can do with little information, and be scary accurate. Granted, the group is one of the best analytic teams in the country for its industry and 50% of the team has Phds from Ivy league schools...but the fact remains that they can give you information you never deemed possible. And, really, it's gotten to the point, they can make non-numeric recommendations - text mining is huge now.

One particular model that my team built involved predicting sales reps who would quit within the company within the next six months. Keep in mind the company has 13k employees and about 5k salespeople. 98% accuracy...unreal.

Again, not dinging or crediting DEC, but a very small sample can be used to calculate totals based on very complicated metrics and calculations. I think that is why you see the precise numbers. I think the real value here is the trend data...as long as their collectionmethods remain constat, valid, and reliable, then the can base that off of the trends.

Edited August 22, 2012 by phade

[growalot]

!f one would look at the variables at work in predicting sales reps quitting in a company and numbers of animals in the wild...Well it's ridiculous...not saying you are but the numbers are...There are so few variables in one compared to the other...calculating one over the other is worlds apart..IMO

disease...weather... births....predation...hunting ...accidents...ect

They are trying to go by bow hunter surveys on what is seen in the fall...how much do you think bow hunters really move around...and how many hunt private over public lands

[fasteddie]

Like I said 97.3 percent of statistics are made up .

[k burke]

I don't see it as condescending at all. Yes, they are providing numbers down to single digits, but they say they are estimates and I'm sure they are based on a limited survey to hunters. Are you suggesting that wildlife sciences should only be rounded to the 100's, 1,000's - or just "a lot", "a bunch", etc.

I'm being facetious, but what do you see are the other options? A census of the animals and if you're not certain every animal is counted there's no benefit to trying to quantify them?

I'm with Phade, I think the main benefit is the trending of the estimates. And I'm sure there are just as many unlimited variables in a population of individual sales people as there are with wild animals. From my understanding he was referring to 98% accuracy of predicting the individuals who quit vs. just the number of sales people who quit.

I guess that's all just to say it's better than nothing and better than some individual just giving a general opinion of "a lot", "a little", etc.

[phade]

!f one would look at the variables at work in predicting sales reps quitting in a company and numbers of animals in the wild...Well it's ridiculous...not saying you are but the numbers are...There are so few variables in one compared to the other...calculating one over the other is worlds apart..IMO

disease...weather... births....predation...hunting ...accidents...ect

They are trying to go by bow hunter surveys on what is seen in the fall...how much do you think bow hunters really move around...and how many hunt private over public lands

I'm in a $2.5B a year company - there are unlimited contributors to why reps quit - and indicators lay hidden in the data fields. The DEC has models in place to account for what variables there are as best as possible. As long as it is done consistently, there is significant value from the calculations.

Value for Sales Rep Model: Fixing the contributors that can be modified, such as commission or bonus structures, car allowances, benefits, territories, training, coaching carries with it a cost savings because rep turnover reduction directly correlates to improved performance. Certain KPI can show what contributor is leading a rep to the liklihood of leaving. Intervention can then take place to "save" a rep.

For the little bunnies, squirrels, etc. 20 years of trend data mashed into a model may reveal that certain snowfalls, average temperatures, hunter densities, etc. have X impact. X impact then translates into what expectations are for the following season, or even 3-5 years out. Maybe it signals a season structure change, maybe it signals need for reinvestment in educating sportsmen the fun and value of plugging a few bushy tails...maybe it prevents a population threat. Bottom line is...these numbers play a significant role in how not only the species are managed, but also the humans, and finances.

Edited August 22, 2012 by phade

[Doc]

I don't see it as condescending at all. Yes, they are providing numbers down to single digits, but they say they are estimates and I'm sure they are based on a limited survey to hunters. Are you suggesting that wildlife sciences should only be rounded to the 100's, 1,000's - or just "a lot", "a bunch", etc.

I'm being facetious, but what do you see are the other options? A census of the animals and if you're not certain every animal is counted there's no benefit to trying to quantify them?

I'm with Phade, I think the main benefit is the trending of the estimates. And I'm sure there are just as many unlimited variables in a population of individual sales people as there are with wild animals. From my understanding he was referring to 98% accuracy of predicting the individuals who quit vs. just the number of sales people who quit.

I guess that's all just to say it's better than nothing and better than some individual just giving a general opinion of "a lot", "a little", etc.

First of all, with todays combination licenses, the DEC really doesn't even know what numbers of small game hunters that are actually targeting what species of small game. That is one huge set of combinations and variables. There is also huge variations in habitat from one corner of the state to the other ..... another huge set of variables. When it comes to expanding small samplings to encompass the entire state, such variables have to be accomodated. Weather (climate ) is another monstrous set of variables that can effect harvest accumulations. Then there is a wide variation in the intensity of each hunter's efforts ..... another huge set of variable. How about survey respondants that don't recollect all that accurately. I'm sure I am not thinking of many other massive sets of variables. But you get the idea that the level of interacting combinations of all these variables are absolutely staggering. The sampling size would have to be massive to have any credibility as data entries into any statistical analysis.

Not to belittle the effort, but all that makes a simple statistical study of predicting the percent of sales reps that will quit, kind of like child's play. Also, there is the factor of verification. Predicting exiting sales people can be verified as being correct or not, after the fact by simple counting. Not so with small game harvests. Also there are definite records from past time periods that the statistics can be based on with the sales rep predictions. That's a huge benefit to accurate statistics. There are no such actual counts when it comes to rabbits and squirrels and foxes and such. So the comparison is badly flawed.

So, it comes down to the question of whether bad data is better than no data. My guess is that it is a 50-50 crap-shoot as to whether decisions and actions based on flawed data will be positive or negative. Yes, it is possible that bad estimates could trigger reactions that could be harmful. I think there are quite a few cases in deer management where bad estimates have caused inappropriate permit allocations which have had serious negative reprocussions in herd stability both in terms of under-estimation and over-estimation.

So what would I do different? ...... I have no idea. It isn't really my job to to come up with an accurate way to count little critters. Nobody pays me for that. But then I'm not the one that's publishing numbers and perhaps pretending a level of knowledge and expertise to feed the public. Is it wrong to question something that seems too incredible to be true? I don't know. I just read these kinds of reports and simply have to ask, "How in hell could they possibly know that?" Perhaps like many, I should be satisfied with the answer that I don't really have a need to know ..... lol.

[growalot]

I know this would be a never ending debate...so I'll just make one more observation on your reply...

Value for Sales Rep Model: Fixing the contributors that can be modified, such as commission or bonus structures, car allowances, benefits, territories, training, coaching carries with it a cost savings because rep turnover reduction directly correlates to improved performance. Certain KPI can show what contributor is leading a rep to the liklihood of leaving

What you have stated here are what I would refer to as "hard" numbers...actual fixed numbers that can be accounted for...ie...number of product sales to cash in the register

For the little bunnies, squirrels, etc. 20 years of trend data mashed into a model may reveal that certain snowfalls, average temperatures, hunter densities, etc. have X impact. X impact then translates into what expectations are for the following season, or even 3-5 years out.

The numbers you are referring to are so widely variable and as set up now unknown..... getting a true 3-5 year plan isn't feasible....there is no accounting for the hunter density...I buy a LT...or you buy a Sportsman...does that mean you and I small game hunt?...of coarse not...weather trends and averages...Have you been living in this state the last few years?...even the meteorologist can't get it right half the time...I check the weather several times a day and it changes a few times a day. You speak of trends...but if they are making such comments on just the last few years trends...being weather...this weather has been unprecedented in it's extremes...I would love to know how they could possible enter predation into the figures...I haven't seen any numbers nor know how they would even come up with the numbers on ferrel cat kills ...let alone all the other predators...

[phade]

I know this would be a never ending debate...so I'll just make one more observation on your reply...What you have stated here are what I would refer to as "hard" numbers...actual fixed numbers that can be accounted for...ie...number of product sales to cash in the registerThe numbers you are referring to are so widely variable and as set up now unknown..... getting a true 3-5 year plan isn't feasible....there is no accounting for the hunter density...I buy a LT...or you buy a Sportsman...does that mean you and I small game hunt?...of coarse not...weather trends and averages...Have you been living in this state the last few years?...even the meteorologist can't get it right half the time...I check the weather several times a day and it changes a few times a day. You speak of trends...but if they are making such comments on just the last few years trends...being weather...this weather has been unprecedented in it's extremes...I would love to know how they could possible enter predation into the figures...I haven't seen any numbers nor know how they would even come up with the numbers on ferrel cat kills ...let alone all the other predators...

I bet you they can say within a certain acceptable margin of error the number of hunters who hunt small game. They only need a minute sample to correlate that info...it's basic math. That's why you see CNN predicting a political winner in a state with 3% of polls reporting - and much to the grimace of the average joe...they're right 99.99% of the time.

Agreed, this debate can rage. Suffice it to say, I've seen what can be done with analytics. I assure that whatever the DEC attributes to be a variable, they have some model to account for it as best as possible - save for the unknown - such as predicting a population of X and then a biblical winter occurs - but even then they can predict recovery rates with variables on winter mortality.People who are not around predictive analytics reasonably don't understand the power harnessed in that information. It's forest for the trees thing, really. You are getting caught up in the trap of "what if" and not in the year to year consistency in collection, variable accounting, and reliability. Paralysis by analysis...Basically, you account for it in your model, and then move on.And, trust me, I can safely say the DEC has weather and predation variables accounted for in their models since they would be two of the major ones. I bet you could ask their group about the impact of population on a particular game, such as rabbits, with an average winter degree change of .5 degrees, 1 degree, etc. or forecast predation by coyote with an expanded population of 3% 5%, etc....and get a realistic response. Just because you don't know how they do it, doesn't mean it is impossible. I don't know how to fly to the moon...but NASA does - think about the variations in that model before they pushed the "go" button.

Edited August 22, 2012 by phade

[growalot]

Just one more observation...I admit I have no idea what they actually use in their models...but.. I just read in the NYON...that it looks like turkeys had a good hatching dry spring and all

Now normally I see 6-9 poults with a hen around our place...have posted pics of this..in years when they said numbers were down across the state...this year..2 poults...that's it...now we very well may have had a good year...but the wild card factor in this was...the main nesting area for the turkeys in our area was logged at the same time the chicks were due to hatch...I ask you ...Do you think that the DEC would ever consider logging operations in turkey poult numbers? and if so how? for no one needs to report a property being logged nor when it was logged...and as to my knowledge there isn't any law against logging when turkeys are on nests....just an example of why I have no confidence in their numbers

[phade]

Just one more observation...I admit I have no idea what they actually use in their models...but.. I just read in the NYON...that it looks like turkeys had a good hatching dry spring and all

Now normally I see 6-9 poults with a hen around our place...have posted pics of this..in years when they said numbers were down across the state...this year..2 poults...that's it...now we very well may have had a good year...but the wild card factor in this was...the main nesting area for the turkeys in our area was logged at the same time the chicks were due to hatch...I ask you ...Do you think that the DEC would ever consider logging operations in turkey poult numbers? and if so how? for no one needs to report a property being logged nor when it was logged...and as to my knowledge there isn't any law against logging when turkeys are on nests....just an example of why I have no confidence in their numbers

Microcosm vs. Macrocosm.

You've also not considered that many species are renumerating.

It may very well be that there are a good number of turkey already in the area, and because of that, fewer eggs were laid and hatched. Instead, you attribute that to logging - which may be accurate...or not.

It's hard, but you have to train your brain when it comes to this type of data. I'm barely able to hold conversations with some of my team...they are so socially inept it is comical. But if you asked them to extrapolate a random scenario...they can whip it out of their heads.

What I can say is that you are focused on the number being accurate to the Nth degree, and really, that's where you're missing the dots to connect. The DEC takes into considerations all of the variables it can, crunches numbers, and uses the data changes in BETWEEN THOSE NUMBERS, to guide decision making. It doesn't matter if they are accurate to within 0.3% or 1.2% or 3%....as long as everything stays the same in how those numbers are gathered, considered, and processed, the changes are where the value comes from.

Edited August 22, 2012 by phade

[growalot]

That's why you see CNN predicting a political winner in a state with 3% of polls reporting - and much to the grimace of the average joe...they're right 99.99% of the time.

Sorry are you not noticing a trend?...CNN knows the "hard" numbers of Democrats...Republicans ...Independents......They have Countable numbers from past elections to go on to make such a prediction...

Such an example used in comparing data input for that group over a wild life population doesn't work to show how analysis...data input can be accurate

Edited August 22, 2012 by growalot

[phade]

Sorry are you not noticing a trend?...CNN knows the "hard" numbers of Democrats...Republicans ...Independents......They have Countable numbers from past elections to go on to make such a prediction...

Such an example used in comparing data input for that group over a wild life population doesn't work to show how analysis...data input can be accurate

Again, no offense, but it seems like this is going over your head. The reference to CNN was related to the number of hunters who target small game.

The DEC does the same with hunters - they have availability to a variety of polls, studies, samples, surveys, etc. and those hard numbers from a minute portion of the hunting community can be extrapolated to a certain degree of accuracy.

FYI CNN uses exit polls as the leading input to predicting winners. Previously, they used past elections and party information; however, it had proven to be less accurate than exit polls. Exit polls are scary accurate.

Edited August 22, 2012 by phade

[growalot]

This is the close...if the DEC published the exact criteria they use in how they predict deer...small game and fish populations then I would read and see if my opinion on how they correlate the data would change...

[phade]

This is the close...if the DEC published the exact criteria they use in how they predict deer...small game and fish populations then I would read and see if my opinion on how they correlate the data would change...

There is a reason that info is not posted - much of it is likely creative property on contract. I doubt they have a "traditional" bio or predictive analytics staff per say...its probably done via contract work so they don't have to pay benefits and (ridiculous) salaries that go along with it.

It also goes back to the fact that most people won't understand it altogether, reduces their authority, and is what many call a no-win proposition.

Edited August 22, 2012 by phade

[steve863]

Yeah, it is sometimes hard to believe, but Phade is 100% correct in what he is saying. Take exit polls for instance. How many people do they question out of all the people who vote? Mighty few, yet the predictions they make from these exit polls are hardly ever wrong, especially these days. They have got it down to a very precise science.

[phade]

And trust me, I'm not a DEC leghumper. I think most of the guys/gals there are good, but are hamstrung by regs, money, and politics. I bet if we let those there now run their own ship with no constraints, we'd all be happy. But, suffice it to say, they are shackled and because of that, make some dumb moves.

Because I am immersed in analytics and stats daily for a publicly traded company that relies on it...I have seen first hand how things like variable are taken into context, and how decision making can be steered by such data collection and interpretation.

My team is now mining facebook and twitter to determine clients who are more likely to buy a certain product/service based on things they say, what's in their business name, how many friends/followers, they have, etc. Now that is something to behold...taking text and turning it into numerical data.

[k burke]

No method is perfect, but bottom line for me is that these are trained professionals with advanced degrees (probably) in an scientific area, probably working with a somewhat consistent and accepted method. I don't see how or what other method(s) would be better.

[growalot]

Why you bring up some very good points...here's a question speak of constraints....Do you believe all analytical teams are equal in accuracy...is your team as good as the next guys ? You know where I'm going with this right.....Really have a good day...lol

Edited August 22, 2012 by growalot

[phade]

Why you bring up some very good points...here's a question speak of constraints....Do you believe all analytical teams are equal in accuracy...is your team as good as the next guys ? You know where I'm going with this right.....Really have a good day...lol

Oy.

You're killing me Smalls.

And, you're grasping for straws. Although my group is one of the best in country according to most professional associations, I am confident the professionals behind the DEC info are capable and accurate. I'm even more confident if they are hired guns...analytic companies who have government contracts are almost always some of the most highly skilled stats geeks out there. That's where the real money is...even more than traditional for-profit companies. Its guranteed money in most cases.

Edited August 22, 2012 by phade

[steve863]

Phade, I think you need to give Growalot a statistical analysis of exactly how many deer she will see this year from all of her 224 tree stands and 43 food plots and maybe she will start believing some of things you are saying. LOL Of course you have to take into account the trespassers that rampage her land and might be scaring some of the deer off of it. LOL

[Doewhacker]

And don't forget the attack Labs that roam there too!