New Mexico Big Game Draw Simulator

[Bigeaux20]

I didn't realize it would require downloading which you can. But here is a snippet. Seriously, anyone suggest your hunt codes, and it can be for any animal, any weapon, and resident, non-resident, or outfitter.

 

[wapitibob]

So this database (of which this is a snippet), that is a public document? They use this one? Or a different one? This is what I used.

That is a draw report spreadsheet, not the NM Elk draw database. You need to file a public info request to get the database.
To accurately simulate the NM draw specifically, you need to know the choices, by order, of every applicant, and use the same random number generator, under the same parameters the NMDGF.

As I said before, Gohunt buys the database; which has all choices, in order, for every applicant. Then they run a monte carlo simulation. You may not like the results, but they're not "wrong". The difference is NM (likely) assigns random numbers differently than Gohunt does, and only runs thru the database one time.

This is an example of the WY draw database.

 

[Maverick1]

"Mental Masturbation" is a great line. I'm not sure you haven't nailed it. Feels like a question I am driven to answer more than the joy at the end. Tell me if this is the wrong way to look at it, but the question I keep asking and tell me if you have a feel for the answer, but what do the odds mean. The question I am trying to answer is not what do the odds mean in general terms, but what does it specifically mean. Does it mean if there is 3% odds I can expect to draw 3 times in 100 draws? Does it mean if there are 100 people, 3 should expect to draw. Does that make sense?

Not certain what you are asking in terms of “what do the odds mean”???.

But to put it into terms for understanding and provide some perspective: If you applied for a tag every year from the age of 20 to the age of 70, over that 50 year period the odds would indicate you might draw one tag. Or zero tags. Or two tags.

On any individual year: you and 49 of your best friends apply for tags. Out of you and those 49 friends, one of the fifty of you would get a tag. Or two.

As stated above, creating a tool to get to that answer is kind of an exercise in mental masturbation.

Start key-stroking. Knock yourself out!

 

[Maverick1]

I pulled some random hunt codes for non-residents and ran it 100 times. Like I said if anyone else wants to see the simulation on three hunt codes, post em up and I will put up the report. I already know im crazy, but maybe this can help someone.

 

[wapitibob]

......

As stated above, creating a tool to get to that answer is kind of an exercise in mental masturbation.

Start key-stroking. Knock yourself out!

And all these "odds" are past tense; created after said drawing is complete. No different than looking at the odds of winning 5 games of solitaire, after you finished playing.

Predictive odds is where were headed.

 

[Bigeaux20]

Not certain what you are asking in terms of “what do the odds mean”???.

But to put it into terms for understanding and provide some perspective: If you applied for a tag every year from the age of 20 to the age of 70, over that 50 year period the odds would indicate you might draw one tag. Or zero tags. Or two tags.

On any individual year: you and 49 of your best friends apply for tags. Out of you and those 49 friends, one of the fifty of you would get a tag. Or two.

As stated above, creating a tool to get to that answer is kind of an exercise in mental masturbation.

Start key-stroking. Knock yourself out!

Well I guess we will see how obtainable it really is. See you didn't even know you were helping all along. Nice job. I appreciate you!

 

[swavescatter]

I finally got access to draw odds with free onyx premium... Meh.

I've written Monte Carlo simulations before, but with things much more straightforward: nuclear physics. People aren't predictable like physics.

Unless you're modeling sentiment based on previous year weather for each unit, but also changes in neighboring states (based on species) and predictive draw pool population, you're likely missing half of the story.

In the end, I pick two dream tags for choices 1 & 2, followed by a slump buster third choice. Either I pull a low draw ranking or I don't. For non residents, I doubt it matters as much.

 

[Bigeaux20]

I finally got access to draw odds with free onyx premium... Meh.

I've written Monte Carlo simulations before, but with things much more straightforward: nuclear physics. People aren't predictable like physics.

Unless you're modeling sentiment based on previous year weather for each unit, but also changes in neighboring states (based on species) and predictive draw pool population, you're likely missing half of the story.

In the end, I pick two dream tags for choices 1 & 2, followed by a slump buster third choice. Either I pull a low draw ranking or I don't. For non residents, I doubt it matters as much.

The "people aren't predictable like physics" is saying a lot. Seems like that's about all anyone can extrapolate is some variance based on how they acted this time, then that time, and the last time, etc... It seems like it's a known unknown in that it's impossible to predict what people will do. But what feels like is possible, is to draw comparisons from what they have done overtime and those patterns seem to converge. And it seems to me the dynamics of a single draw, is dependent on the random sequencing. Human choice could literally change 25% in one direction, and the draw will produce 25% in the other. Hyperbole of course. It's been an interesting endeavor for certain.

With OnX premium or elite you should be able to get Huntin Fool odds as well, in case you weren't aware. And why are they so much different than GoHunt? That's what got me going down this rabbit hole in the first place. Interesting perspective for certain.

 

[Mighty Mouse]

The question I am trying to answer is not what do the odds mean in general terms, but what does it specifically mean. Does it mean if there is 3% odds I can expect to draw 3 times in 100 draws? Does it mean if there are 100 people, 3 should expect to draw. Does that make sense?

In common parlance, I would say 3% odds means the latter (i.e., 3 out of 100 applicants drew). That’s certainly how New Mexico (and all other states to my knowledge) reports their data: they tell you there were X number of applicants and Y number of them drew. Knowing the odds of success in a single year, you can calculate the odds of drawing any number of times over any number of years using the binomial distribution formula:


Below is a graph of this formula for every value of x with n=100 and p=3%. There is no single answer to “How many times can I expect to draw?” The answer is a distribution of probabilities.

 

[Bigeaux20]

In common parlance, I would say 3% odds means the latter (i.e., 3 out of 100 applicants drew). That’s certainly how New Mexico (and all other states to my knowledge) reports their data: they tell you there were X number of applicants and Y number of them drew. Knowing the odds of success in a single year, you can calculate the odds of drawing any number of times over any number of years using the binomial distribution formula:
View attachment 852642

Below is a graph of this formula for every value of x with n=100 and p=3%. There is no single answer to “How many times can I expect to draw?” The answer is a distribution of probabilities.
View attachment 852643

Yeah now the trick is getting that to translate to a cohort of people who lack the chops to absorb all that. I can follow along, but I can't explain it back to you. Tell me if this sounds bananas. To me it's easier to digest that if you want one of the 9 late archery non-resident tags in 16A, you have to be sequenced, on average, in the top 2000 people when the music stops. If you change to the early archery instead, once again on average, you have to be in the top 6000. And that is an extrapolation of thousands of simulations that eventually converge to something that resembles a value. And as always, I think this is true. How do you feel about that information. Thanks a ton for your input. It's awesome and the math aspects take me way back to when my brain was younger.

 

[Bigeaux20]

Mighty Mouse, this is going to sound like weak sauce, but by all means I don't have the skillset to differentiate your binomial distribution against the simulation approach. I had to ask AI what the differences were. When I started this, the question was, what is the absolute most dependable and accurate as possible approach to determining the likelihood of being drawn. Don't quote that but it was some mumbo jumbo like that. A Monte Carlo simulation was the answer provided. What came after that was just an exercise where I kept asking what if. Just wondering how this lands with you?

Binomial Distribution Approach:
Strengths:
Simple and analytical: Easy to compute exact probabilities without running simulations.
Provides a clear theoretical distribution of outcomes over many trials.
Useful for understanding long-term odds with a fixed success rate.

Weaknesses:
Oversimplified: Ignores real-world complexities like multiple choices, quotas, and competition.
Assumes a constant success rate, which may not reflect reality (e.g., the simulation showed ELK-2-257 had a 5% draw rate, not 3%).
Less practical for short-term planning (e.g., “Will I draw in 2025?”).

Simulation-Based Approach:
Strengths:
Realistic: Models the actual draw process, including choice preferences, quotas, and competition.

Detailed: Provides per-hunt-code, per-category, and per-choice success rates, plus additional metrics like exhaustion points and “Bigeaux” success rates.
Practical: Focuses on a single year (2025) with projected data, making it directly applicable to planning.
Flexible: Can incorporate historical data, variability, and custom rules (e.g., the “Bigeaux” applicant strategy).

Weaknesses:
Computationally intensive: Requires running many simulations (100 in your code) to get stable results.
Relies on assumptions: The ±X% variation in applicant numbers and the rules for assigning 2nd/3rd choices are based on assumptions that may not perfectly match reality.
Single-year focus: Doesn’t directly address long-term odds over 100 years (though you could extend the simulation to multiple years).