Tim Stolte

Quantitative Researcher at @AmdaxGroup. Master's degree in Econometrics / Quantitative Finance.

[Amdax @AmdaxNL]

𝗔𝗺𝗱𝗮𝘅 𝗶𝗻𝗶𝘁𝗶𝗮𝘁𝗲𝘀 𝗔𝗺𝘀𝘁𝗲𝗿𝗱𝗮𝗺 𝗕𝗶𝘁𝗰𝗼𝗶𝗻 𝗧𝗿𝗲𝗮𝘀𝘂𝗿𝘆 𝗦𝘁𝗿𝗮𝘁𝗲𝗴𝘆 (𝗔𝗠𝗕𝗧𝗦)

We are proud to announce that Amdax initiated AMBTS B.V. (‘AMBTS’), with the intention to list a Bitcoin treasury company on Euronext Amsterdam. The long-term ambition of AMBTS is to become a 1% Bitcoin treasury company over time and unlock a compelling investment opportunity for institutional investors.

💬 @LucasWensing explains:
“While Bitcoin has been the best performing major asset in the past 10 years with fast adoption as digital capital, it is still relatively small in investment portfolios. With now over 10% of bitcoin supply held by corporations, governments and institutions, we think the time is right to establish a Bitcoin treasury company with the aim to obtain a listing on Euronext Amsterdam, as one of the leading exchanges in Europe. This will provide institutional investors with new financial instruments to get exposure to the performance of bitcoin.”

Read the whole press release: https://t.co/BkGNxzgsTn

[Amdax @AmdaxNL]

𝐀𝐦𝐝𝐚𝐱 𝐛𝐞𝐡𝐚𝐚𝐥𝐭 𝐌𝐢𝐂𝐀𝐑-𝐥𝐢𝐜𝐞𝐧𝐭𝐢𝐞 🇪🇺

Met trots maken we bekend dat Amdax de MiCAR-vergunning heeft ontvangen. Met deze vergunning voldoen wij aan de nieuwe Europese wetgeving voor cryptomarkten.

💬 𝐀𝐦𝐝𝐚𝐱 𝐂𝐄𝐎 @LucasWensing :

"Deze vergunning is een erkenning van onze langetermijnvisie, “We hebben vanaf het begin gebouwd aan een bedrijf dat niet alleen technologisch vooruitstrevend is, maar ook regelgeving, integriteit en risicobeheersing centraal stelt in zijn bedrijfsvoering. Die strategie werpt nu zijn vruchten af."

Met MiCAR bieden we vier gereguleerde cryptodiensten aan onder één vergunning:
● Orderuitvoering
● Bewaring
● Overdracht
● Vermogensbeheer namens cliënten

Deze combinatie van diensten onder één vergunning is uitzonderlijk in Nederland.

Een groot compliment aan het team dat dit mogelijk heeft gemaakt. Samen blijven we bouwen aan een professionele, transparante en betrouwbare cryptomarkt.

Lees het volledige persbericht hier: https://t.co/mRLTa9uAcE

[Amdax @AmdaxNL]

BlackRock lanceert een nieuw fonds op de blockchain. Hoe dat werkt? We leggen het je uit in de Marktupdate van deze week.

Ook bespreken we de geruchten over de aanklacht tegen de Ethereum Foundation, en de rol van ethereum staking bij spot ether ETF’s. 👇
https://t.co/qtTIySUsb6

[Tim Stolte @TimStolte98]

Discussion on the #Bitcoin power law model stirred up again.

@hcburger1, @Giovann35084111 and @peter_vijn published articles in response to my critique on the model.

I'd like to point my attention towards these gentlemen and share my view on some of their ideas/findings. 🧵👇

[Tim Stolte @TimStolte98]

@Wojball @hcburger1 @Giovann35084111 @peter_vijn Either one is statistical evidence that the model is likely to 'fail', which is important to know now before it happens.

Also, we'll probably not realize when it fails because very long periods of over/underestimation are quite common.

[Tim Stolte @TimStolte98]

Discussion on the #Bitcoin power law model stirred up again.

@hcburger1, @Giovann35084111 and @peter_vijn published articles in response to my critique on the model.

I'd like to point my attention towards these gentlemen and share my view on some of their ideas/findings. 🧵👇

[Tim Stolte @TimStolte98]

@Wojball @hcburger1 @Giovann35084111 @peter_vijn Sure! Despite their defense, I still think the model is invalid due to:
1) Autocorrelation; over/underestimation today means over/underestimation in the coming period (weeks/months)
2) Non-stationarity residuals; there will likely form constant over/underestimation in the future

[Tim Stolte @TimStolte98]

@hcburger1 @Giovann35084111 If you're curious, these are the out-of-sample forecasts since the start of 2022.

Unfortunately, I can't expand the out-of-sample window due to the nature of the method (it requires two full cycles).

[Tim Stolte @TimStolte98]

@hcburger1 @Giovann35084111 It's therefore that I strongly believe there is no proper way to forecast asset prices in the medium/long-term.

The only thing I can come up with is a forecast without any statistical inference. Feel free to read this article. https://t.co/xIPNhw2kKx

[Tim Stolte @TimStolte98]

@hcburger1 @Giovann35084111 @peter_vijn ... and here https://t.co/6RGCaW6gnh

[Tim Stolte @TimStolte98]

@hcburger1 @Giovann35084111 @peter_vijn For those who are interested in the original articles, you can find those here https://t.co/ox6nijMMGN

[Tim Stolte @TimStolte98]

@hcburger1 @Giovann35084111 @peter_vijn But it turns out the 'lower bound forecasts are even better than the original ones (based on the RMSE).

This shows that it’s very easy to create ‘incredible out-of-sample forecasts’ in a log-log setting, even if we mess up the slope deliberately.

20/20

[Tim Stolte @TimStolte98]

@hcburger1 @Giovann35084111 @peter_vijn Let's play around with the coefficients and assign values to the exponent that are outside the 95% confidence interval of its ‘true value’.

If the power law model actually produces incredible forecasts, my contaminated forecasts should be nowhere near as accurate.

19/20