Mine Modelling

[doctorj]

For a little while I've been playing around with developing excel models that value single-mine companies using monte carlo analysis. I'm wondering if anyone else has experience in this field or if anyone else has models they would like to share.

For me, I got into it thinking in ground value wasn't a sufficient measure. 'The Mining Valuation Handbook' by Dr Victor Rudenno has been a massive help. Also, I recently discovered an excel model (attached) by Prof. Richard de Neufville from MIT that's been an absolute revolution for me.

So, does anyone here try to build these models? If so, how do you go about developing your assumptions and what do your models look like?

 

[Sean K]

This looks great doctorj. I'm trying to make heads and tails out of it now, but it's certainly something I would like to work on to improve my pitiful attempts at valuation. Cheers, kennas

 

[cjfisher]

Interesting Model - thanks for sharing it.

My issue with this type of DCF model is that it looks too far ahead, and doesn't take into account further mineral discoveries, the cyclicality of metal prices, and lots of other unknowns and indeterminables.

I would like something a bit simpler and more flexible.

I haven't come up with anything yet - other than simply valuing the resource and subtracting the cost of getting it out of the ground.



Any one else got any ideas?

 

[prawn_86]

Very interesting Doc.

Once im finished exams i'll have a good crack at understanding it.

At the moment my valuations pretty much just include in situ values, or if they already are into the mine building phase i look at Enterprise Value, but this seems a hell of a lot more complex.

I usually add some extra 'value' on for the potential of resource upgrades as CJFisher has mentioned, although it is prob not too bad a thing if that is not included in the model, as then you are getting a minimum valuation, with more upside potential if any upgrades are found.

Personally i always try to work 'low' on my valuations. Ie - use much higher costs and much lower metal prices in order to see if the risk/reward is worth it.

 

[tech/a]

Doc.

In the end are'nt you just modelling opinion?

MonteCarlo analysis is in its purest form considering every known variable.
In the case of Fundamental modelling you have to place opinion as a model at some point.

The difference I see with this and a technical model is that with a technical model we can test a completed model with static variables over massive data bases,which have known endings.
The result is an expected return going forward.

I dont think you can have enough data to produce a reliable statistical significance.One variable is percieved valuation,your and mine can be vastly different regardless of what the Numbers say it SHOULD be. How often have you seen this disparity occur? Ducati's exercise proved this clearly I thought.

But happy to see evidence to the contrary.

 

[prawn_86]

Tech,

As you know, my TA knowledge is limited, and systems knowledge even more so. But dont opinions come into it there also?

If not your opinions, the opinions of those who designed the software you used? I understand that they work on probabilities and historical data but somewhere along the line a decision, hence opinion would have had to have been made.

I think that if you can get an accurate FA model it would be very handy, and much quicker than the way i carry out my valuations. Perhaps one way of testing this would be to put in historical variables of currently operating mines, before they were operating and see how close the result gets.

Have you tried this Doc?

 

[doctorj]

In the end are'nt you just modelling opinion?

In the end - sort of. I really don't want to turn this into a fundamental vs technical arguement. In a sense you're trying to make the inputs as objective as possible and examine the impact of reasonable changes in known variables on the final result. For me, this kind of analysis is most useful as a company nears production - many of the costs are already known and many of the risks in setting up a mine are taken out.

I can't say I know of Ducati's experience (i'd say it's probably outside the scope of this thread) - but the goal of this has always been to model the value of a single project to a company, not to model the value of the company or to model a group of projects.

 

[doctorj]

Perhaps one way of testing this would be to put in historical variables of currently operating mines, before they were operating and see how close the result gets.

Haven't done this - I'm far too early along the rude in determining what elements contribute to costs to start testing models. Things like the cost/availability of water & acids & transportation and who knows what else. Also, it's important to understand how a company takes a chunk of ore to get the contained metal out.

The either side of the equation (the revenue) is much easier once you know how fast they extract the ore, provided you can guess how they plan to model the pit or the decline. Drill intercepts tend to help this - they like to get the shallow, high grade out first to help early cash flow.

For anyone who decides to go down this path, if you can, I'd recommend going on a tour of an operating mine. I've been a couple of times and it's a massive eye opener.

 

[prawn_86]

Im really interested in this, i think it could be great to get another valuation method aside from the one i use, which has really just been thrown together as i increase my knowledge.

If you need any help with it feel free to let me know.

Once exams are finished...

 

[MS+Tradesim]

DoctorJ,

I think what you are trying to do is very interesting. Hope you work it out.

My real question is, will it improve your profitability? The reality of market sentiment and the greed/fear factor will ultimately see SP moving around fair value but rarely at it.

I love getting into the nitty gritty but apart from a sort of broad fair value range for targets I haven't found it to be overly useful to my bottom line. If you find a way to practically implement this sort of depth to analysis...fantastic! Go for it.

 

[doctorj]

My real question is, will it improve your profitability? The reality of market sentiment and the greed/fear factor will ultimately see SP moving around fair value but rarely at it.

Sentiment plays a massive part in price, there's no doubt. And who knows whether or not it'll be more or less profitable than what I'm currently doing. One thing's for certan though, I'm likely to learn a lot about mining along the way.

 

[MS+Tradesim]

One thing's for certan though, I'm likely to learn a lot about mining along the way.

Which will have a positive effect in itself...you'll be better prepared to pick from among the miners who has the better growth prospects based on project feasibility and economics. So, I suppose that will improve your profitability by allowing you to steer clear of Co. with projects that are not cost-efficient, etc.

 

[Pommiegranite]

DoctorJ,

I think what you are trying to do is very interesting. Hope you work it out.

My real question is, will it improve your profitability? The reality of market sentiment and the greed/fear factor will ultimately see SP moving around fair value but rarely at it.

I love getting into the nitty gritty but apart from a sort of broad fair value range for targets I haven't found it to be overly useful to my bottom line. If you find a way to practically implement this sort of depth to analysis...fantastic! Go for it.

I have come to realise that, in the resources sector, nearly every stock is undervalued according to the investors in that stock.

Sure this may be the case when taking inground resource valuation versus market capitalization. However, not enough comparative valuation between stocks is done. I believe that an analytical tool like the above spreadsheet is pretty much invaluable when deciding which stock to invest in.

In the end, once a company is receiving revenue and making profits, even market sentiment will not be able to deny a valuation range that a stock deserves to be in, especially if that stock's financial model is very robust.

 

[tech/a]

but the goal of this has always been to model the value of a single project to a company, not to model the value of the company or to model a group of projects.

Here in lies the crux of the matter.

Doc had no intention of the Fundie/Techie comparison turning into an argument.Both are valid forms of analysis.

Basically I see you doing a "Layman's" project valuation,which the company no doubt would have its production team working on from within.
Difference being your doing it to assign "Fair value" where as they would be doing it for feasibility.

In the end you'll no doubt have a high low and median range of "fair value".
The only problem I see is then having fellow investors agreeing with you to the degree of putting their money where your valuation is.


https://www.aussiestockforums.com/forums/showthread.php?t=2829&page=9

This is ducati's thread now seemingly defunct.
Although Duc was un able to prove his point and the thread is now defunct,I still think the exercise gave valuable insights into some of the misconceptions about Fundamental valuations and Duc's views on Stops.
Worth reading for all those trading whether fundamental OR Technical.There is a very strong message in the final result.
An invaluable thread.For those who are interested in re opening the thread and discussion I'm happy to participate there.
As I followed with interest from inception and ducs aim was to prove he could trade a 12% a year profit/compounding Fundamentally without stops over a 3 yr period,unfortunately it stopped a year short.

 

[cuttlefish]

I think when doing these valuations its also very important to consider some of the more qualitative factors such as the shape, depth, general geology of the ore body(s), the type and quality of the ore. Another factor is location which can create various macro risks such as political risk, operational risks (availability/cost of labour, transport logistics, weather/climate etc.)
And the third factor is management - how well have they done pre-mining analysis and due dilligence, how tight a project/operational management ship do they run, what is their experience etc.

I think these factors, as well as the overall project numbers vs market cap, are important in picking stocks that will make the distance and thus have the corresponding price tag allocated to them when the time comes.

The comment above that there are many junior mining/exploration companies that appear undervalued is true, which is why the above factors are important.

I see in-ground value alone as almost useless in valuing mining companies.

The spreadsheet you've got isn't dissimilar to ones I've put together in the past. I think metal price sensitivity and longevity are two big factors that differentiate between good quality and poor quality projects. Ultimately metal price sensitivity of a project tends to be most heavily influenced by recoverable grade (i.e.grade as tested by metallurgy) and depth/shape, though of course operational logistics are also a big factor. Longevity allows for potential capex overruns on startup to not kill the project (assuming the operating margin is good enough).

 

[doctorj]

For me - this is the stock that started my interest in modelling. I think it's a good example how a better understanding of the numbers produced at the Feasibility Studies can be used to identify opportunities (or even to create a universe for technical entry).

I'm not sure of a better way of analysing the data released in feasibility studies than feeding them into a model. Thoughts?

 

[cuttlefish]

I had a quick look through the equinox reports from the pre-mining feasability time period.

Positive qualitative factors along the lines I was talking about in my previous post:

* size/longevity - 350 million tonnes is huge meaning decades of mine life.
* grade - 0.76% is relatively high grade over that size of resource and compares well to mines like escondida etc.
* ore body shape/geology/depth - consistent simple shaped orebody, long shallow lode dipping gently from surface, most ore less than 200m below surface. Its also nearly all sulphide resource. This is a perfect open pitting scenario.
* Level of detail in the planning and feasability documents as well as the quality of capital backing they managed to achieve gives a leve of confidence that management seem to know what they are doing.

though I only skimmed them, the reports also seemed to address issues in relation to the political/title risks of the location, the weather and the logistical issues very well.

I'm still of the belief that as well as the raw feasability numbers adding up, these qualitative factors are also very important in sorting the wheat from the chaff when selecting upcoming mining companies.

 

[stoxclimber]

I'm not sure how appropriate Monte Carlo analysis would be. To me it seems like you should be using real options. The issue there is that at a cyclical high in metal prices you'll need to adjust your perception of in-ground value because if you assume that your in ground value is lognormally distributed around current spot prices you would probably significantly overstate value. Real options can also address mineral discovery and so on, like another poster alluded to.

Another factor to consider for valuing resource companies is that sector wide they will generally universally trade at a discount/premium to what one would consider intrinsic value depending on the pessimism/optimism of the sector as a whole, especially as it's tougher to model resource companies vis-a-vis general industrials etc. That's why multiples valuation is emphasised in the resource sector to a significantly greater extent than other sectors where modeling is more accurate, e.g. infrastructure.

PM me with any questions because I might not read the thread again, it's tucked away.