2022 odds

[The Kentucky Boys]

I tried to do that for decades with spreadsheets and it was never possible to get accurate odds from the prior year. I could get a general feel for demand, but not for actual odds that reflected the applicant pool of the prior year(s). Everything I tried with spreadsheets was oversimplifying the true historical odds when the system looks at all three choices before it goes on to the next applicant.

As to what goHUNT does for NM, and any state with multiple choices before going to the next persons (AZ, NV, NM), they get the historical applicant data, tag numbers by unit, applicant residency, applicant choices by species, and rerun the draw for that state over 500,000 times to get the most likely probabilities based on the historical applicant data. That process, compared to my spreadsheets, have some pretty big variations, even though general demand trends are pretty easy to see in any methodology.

Randy thanks for all the information and videos. I'm learning. My problem I was trying to learn Wyoming, Colorado, Nebraska, and South Dakota. I finally had to just look at one at a time. I had never even traveled out West. Everyone has alot of great opinions. But as a newbie to hunting Western game I realized I just had to go. Put boots on the ground and learn alot. I'm looking forward to many more trips. Thanks again. I really enjoy the videos

 

[Chama Grande]

I tried to do that for decades with spreadsheets and it was never possible to get accurate odds from the prior year. I could get a general feel for demand, but not for actual odds that reflected the applicant pool of the prior year(s). Everything I tried with spreadsheets was oversimplifying the true historical odds when the system looks at all three choices before it goes on to the next applicant.

As to what goHUNT does for NM, and any state with multiple choices before going to the next persons (AZ, NV, NM), they get the historical applicant data, tag numbers by unit, applicant residency, applicant choices by species, and rerun the draw for that state over 500,000 times to get the most likely probabilities based on the historical applicant data. That process, compared to my spreadsheets, have some pretty big variations, even though general demand trends are pretty easy to see in any methodology.

Wait so gohunt is getting the individual application choices from a particular year? I'm assuming they're working with NMDGF on data gathering and the like? That would definitely be further down the data rabbit hole than my spreadsheet goes. Especially if they run a Monte Carlo simulation, like I think you're alluding to.

For this draw, I'm probably ok with a rough estimate of demand but maybe next year I'll take a run at the gohunt and compare. Or if all my choices get me thrown in the trash maybe I'll do it this summer

 

[ImBillT]

Wait so gohunt is getting the individual application choices from a particular year? I'm assuming they're working with NMDGF on data gathering and the like? That would definitely be further down the data rabbit hole than my spreadsheet goes. Especially if they run a Monte Carlo simulation, like I think you're alluding to.

For this draw, I'm probably ok with a rough estimate of demand but maybe next year I'll take a run at the gohunt and compare. Or if all my choices get me thrown in the trash maybe I'll do it this summer

I absolutely believe goHunt does indeed get more detailed data than the “complete report” on the NMGF website. What hoops they have to jump through to get that information I’m not sure.

NM is probably the state whose draw odds most depend on who places what codes in what order. To calculate accurate odds without that information is not possible and to calculate odds directly rather than calculate the distribution of simulated results would be a horrendous nightmare rife with opportunity for error. I don’t know what a Monte Carlo simulation is, but goHunt runs the draw exactly the way the state runs it, tabulates the results and runs it again, millions of times, and lists the percentage of times that each applicant for a specific hunt code drew as the draw odds. For instance if they ran 5,000,000 simulations, for a hunt with 1 tag and 10 applicants and the number of times that each applicant drew a tag averaged to 600,000, then the draw odds are 12%. The reason that they are better than 10% is that some of the 10 applicants listed the code in question as a second or third choice, but drew their first or second choice instead, and thus did not count against the odds. I believe it would be possible to calculate the odds rather than simulate, but the odds that someone would draw their first or second choice would be dependent on the odds of another hunt code for which odds where dependent on other hunt codes, so you’d either have an utterly insane system of equations(that I’m not even sure would be solvable), or you would have to interpolate over a large number of iterations, which would be a nightmare far worse than a simulation.

For the easiest to draw NR rifle elk hunt in NM two years ago there were 8,17,23 NR applicants and 8 available tags. If choice order didn’t matter, odds would be 16.67%. Without knowing the other choices of individual applicants the only way to take choice order into account is to use the average probability that some drew a first choice and second choice. Roughly 98% of first choice applicants did not draw their first choice and roughly 97% of second choice applicants did not draw their second choice, improving odds to roughly 17.3%. goHunt listed odds at 27%. Again, this is because goHunt has more detailed information, and that information suggests that most people putting the easiest to draw hunt in the state as their second or third choice had first and second choices with better than average draw odds as well.

If goHunt can run a draw millions of times, AZGF should be able to run it once.