How can you be sure that your model is correct?

How can you be sure that your model is correct?

It's not, it's ignoring forward prices, cycle losses, wrong spot price model etc etc

The worst thing is the claim that he's here to rescue dumb companies who have been operating these $100smln storages for decades. What an arrogance!

 wrong spot price model etc etc

OP seems to be suggesting that there is a unique, robust model that always work. 

 wrong spot price model etc etc

OP seems to be suggesting that there is a unique, robust model that always work. 

"Remarkably" he does

How can you be sure that your model is correct?

It's not, it's ignoring forward prices, cycle losses, wrong spot price model etc etc

The worst thing is the claim that he's here to rescue dumb companies who have been operating these $100smln storages for decades. What an arrogance!

1) Forward prices are not ignored, as I mention, users can calibrate the model themselves (so that the spot converges in expectation under Q-measure to the forwards). 
2) What do you mean by cycle losses?
3) Spot model is not too much wrong. Interestingly, for its simplicity it gives sufficiently good results.
And yes, I have created a better model (here is QQ-Plot: http://yetanotherquant.com/fig5b.pdf ) but it is proprietary.

>here to rescue dumb companies who have been operating these $100smln storages for decades. What an arrogance! 
That shows that you have keine Ahnung (no idea) about German energy market (what an ignorance!).
First of all, big players (E.On, RWE, EnBW) are indeed not so far from bankruptcy.  
Secondly, my model can be useful for medium and small companies like Stadtwerke, they may need (virtual) storage but (with a few exceptions) don't possess the quantitative know-how.    

Both big players and stadwerke use storage models that have been validated by me, I've validated storage models of various vendors and energy companies throughout europe.

What you have is not a real storage model but a very simple toy virtual storage model.

I can't help you improve because I don't want to mis (conflict of interest) having the option to validate your proprietary model. (If it ever gets that far, there are some very good vendor models). Should be an easy job to give a judgment on this

Both big players and stadwerke use storage models that have been validated by me, I've validated storage models of various vendors and energy companies throughout europe.

Most likely, you merely lie.
First of all, if a Stadtwerke (at all) uses something for storage evaluation than it is usually the software from KYOS (which is also, inter alia, used by big companies).
Secondly, though I cannot speak for the whole Europe, in Germany there are not so many vendors; namely KYOS (definite market leader), Decision Trees and probably one or two more.
Thirdly, big companies have their own experts (e.g.  Klaus Spanderen works for E.On, Sven-Olaf Stoll works for EnBW) and they hardly need facilities of the third parties for model development / validation.

And yes, virtual storage is what the most of market players need. Only few of them have physical storages.
   

Yes, KYOS, and there are other players, and larger companies develop their own models and hire external validators.
If you're making a virtual storage model then you should at least have a good gas price model. If you're not even doing that then I don't see how you're doing something even the smallest stadtwerke -which has interest in virtual storage- can't do themselves.

The need for external validation by model owners is not about quant capabilities but about acting responsible and complying to client quality standards. You can't just put up a webbased calculator without any documentation, explanation, validation and expect people to make trading decisions on that?! In the real word there is accountability, operational risk and costs associated with making mistakes.

You can't just put up a webbased calculator without any documentation, explanation, validation and expect people to make trading decisions on that?! I

The Tool is available under common disclaimer: "All content and functionality on letYourMoneyGrow.com is provides (so far) completely free of charge and without any warranty."
Apropos, the best documentation is the program code and I make no secret that I have taken the model from QuantLib.

And I don't expect anything from people, they are welcome to use it as is and if they want more they can contact me.
BTW, I got much more positive than negative feedback

For me the issue was your attitude.

You claim that professionals in the energy sector can't even values storage, but hey, here is the master who comes to their rescue!


However, I know the sector, and I know how professional people are and what models they use, and your model turned out to be complete useless model if you would compare it to the range of professional models people are using all over the place. ...This is for physical storage.. untill you mentioned it's for virtual storage contracts.

I don't have strong opinions on the model as it is. I can imagine you get positive feedback since you give something for free which is nice. But to be useful in practice like in the case "shall we do this €5mln virtual storage deal?" Then I don't expect people to use it (because it's real money and they can't asses if you model is any good or not)

Then I don't expect people to use it (because it's real money and they can't asses if you model is any good or not)

Sure. But what about using the model as a proof-of-concept instead of killing it off without giving it a fighting chance?

In fairness, finanzmaster has produced a model from Quantlib. People are free to apply the model with their own data and report on its applicability. 

Your analysis ideally includes where it works and where it does not  work. Then it is possible to have a decent discussion.

What's "speed"? (where are the docs)
Where is the seasonality? (where are the tests that show how good the model fits reality)

Suppose the model said "the virtual storage with these parameters is worth $1mln". What would it take so that you would sell your house and buy that for $750k?

For me the issue was your attitude.

Critics is welcome as long as it remains sachlich.
Next time please care about wording: "what an arrogance!" is not the best way to start a discussion and discussion in this way makes no honor to both of us.

You claim that professionals in the energy sector can't even values storage, but hey, here is the master who comes to their rescue!

I don't have strong opinions on the model as it is. I can imagine you get positive feedback since you give something for free which is nice. But to be useful in practice like in the case "shall we do this €5mln virtual storage deal?" Then I don't expect people to use it (because it's real money and they can't asses if you model is any good or not)

I cannot (so far) say whether people use it or not.
I can only say that we of course use more sophisticated model when we price the storage but we did also used this model as a cross-check.
And surprisingly, it gives the price, which is pretty close to that from a sophisticated model. Even if we don't make a risk-adjustment (i.e. assume that the historical and risk-neutral measures are the same).