IIP-34: dVIX - Decentralised Volatility Index

VIX products in TradFi are huge and there is nothing like this in DeFi !!

dVIX will create a new market DeFi and will unlock an entire new way for traders to express themselves. Like FLI, dVIX will further extend Index Coop’s moat as the premium index product provider. DeFi needs a VIX product and dVIX is just that.

dVIX can be massive. Hats off to @george & @ampodgorski for coming up with this product. :slight_smile: Simply amazing.

3 Likes

Thanks @Matthew_Graham! Can’t wait to get it rolling.

2 Likes

Thank you @Matthew_Graham ! we think it would benefit the crypto ecosystem!

3 Likes

Welcome @ampodgorski and @george!

You’re landing with a :boom: and I really like the sound of this index product. I think the use cases, target market (huge), and benefits to the Index Coop are logical and compelling. I also think first mover advantage and liquidity are vitally important here. And, the branding advantages of the Coop owning THE Vix product in DeFi are significant.

Given your comments about the early stage development of Hegic and Opyn, what would you suggest an MVP might look like today or soon?

What are the main risks you foresee?

Semi related, as the options infra in DeFi matures and becomes more viable, what do you think of a Variance Risk Premium product in the Coop? (Background: The Variance Risk Premium is Pervasive - and FactorResearch)

1 Like

Thanks for the positivity, @DevOnDeFi!

An MVP would still be viable now, but the price would be more stable with more options. Currently, Hegic has certainly has sufficient market depth of options for calls, but perhaps not puts. Therefore, smoothing functions will be required for the MVP of the dVIX currently. This is what Deribit uses in their methodology for DVOL, which is a centralised exchange volatility index of BTC.

By using this query on Hegic’s subgraph ,

query_str = ‘’’{
options(
where: {
status:“ACTIVE”
symbol:"<selected_symbol>"
type: “<option_type>”
}) {
symbol
strike
period
expiration
type
premium
settlementFee
totalFee
}
}
‘’’

where selected_symbol can be either ETH or WBTC, and option_type is CALL or PUT, yields the following.

For ETH Calls (79 contracts):

For

  • ETH Puts: 19 active contracts,
  • WBTC Calls: 13 active contracts
  • WBTC Puts: only 3 active contracts

Therefore, the BTC element of the volatility index might be uninformative if included at this stage. However, the ETH element appears to have sufficient market depth. Given that the equation used in the whitepaper notes that only the term for options expiring in one week should be used, the equation can be tweaked by including weighted terms for results of options contracts with different expiries.

As for your note on a dVRP, yes this absolutely would be a secondary product to the dVIX, but would require more of a liquidity injection due to the risk born by the pool if it were a pool-to-contract protocol. I think that this could also be partly addressed by a dynamic transaction fee model. On first thought it could be an index, if it were a form of weighted index on the dVIX with a leverage ratio like FLI.

4 Likes

Thanks for the informative reply @ampodgorski - awesome.

dVIX and dVRP seem like high value vol related products for the Coop to focus on - and they’re so complimentary, yet polar opposite in nature.

Strongly FOR both.

1 Like

Sorry it took a while for me to get around to this but great to see it on the forum! Thanks @ampodgorski and @george for sharing this with us.

Given the level of novelty and differentiation I’m really excited to see the community discussing it, I think it’s an important next step in our product offering. Felix laid it out in his post a while back about branching out from thematic indexes to explore the more complex products that are currently missing in DeFi, dVIX certainly meets that criteria.

When I looked into Hegic before it seemed pretty much unusable unless you were writing options (LPing) as the strike prices were absurd. My level of understanding around options is basic so I’d like to know if that affects what you are looking at, and have you considered how the data will improve as DeFi options improve? Is Hegic just the best place to get the data from right now?

Hi @DarkForestCapital. Thanks for the input.

I agree that Hegic’s pricing function is quite flawed. In fact, I read yesterday on their V1 whitepaper that they calculate it via the implied volatility from skew.com. When you use their subgraph their implied volatility values make no sense. It’s usually a stable value of 9500 for BTC.

That being said, we can obtain the implied volatility from centralised sources. This is achievable and perhaps necessary for an MVP. One main decentralised options protocol whitepaper I’d like to properly look through the next few days is FinNexus. Their OptionsFormulas.sol looks very promising, which can be found here https://www.docs.finnexus.io/options/contracts/.

1 Like

Hey @ampodgorski, my understanding is that you need a very deep and liquid secondary options market for a vol index. And that currently, on-chain options pricing is not good. How confident are you that under current conditions the proposed index would actually closely track the volatility of ETH?

Also, I see that you are proposing using 7-day options to construct the index. I assume you will have to roll them forward every week? At what cost? Also, given the potential that these options are mispriced, how would you manage that?

Hi @verto0912

Yes, it needs to be very deep, but I’m not entirely clear on what you mean by liquid. The secondary market for options is certainly in its infancy. Adding a weighted measure from a centralised exchange like Deribit might provide a sanity check on the measure until the options market reaches maturity.

The index does not propose to write options, but only look at the option premiums of existing options or options that could be written according to decentralised options protocols. This has been noted in the above replies. This is a Solidity getter function and would not cost.

In short, agglomerating across several options protocols would provide an appropriate volatility measure, particularly because one can read the options pricing functions according to smart contracts. However, it would be improved greatly and would require less non-TradFi/CeFi maths and metrics once the secondary options market develops on DeFi.

1 Like

I just meant you need a wide range of strike prices and maturities for ETH and BTC, with significant secondary market liquidity.

Would it be possible to run a simulation for the next 30 days and see what how dVIX would perform?

1 Like

Hi @verto0912, it is correct that you would need a wide range of strike prices. Decentralised option protocols have smart contract pricing functions that are functions of strike price, i.e. pseudo-continuous. These will provide the required market depth to begin with.

It is incorrect that you need a wide range of maturities. You need two: one either side of 7 days from today (or when the calculations are run). If you read the referred DVOL methodology linked above you will see that they use two maturities for their 30 day measure.

In the proposal above, there is a screenshot from a backtest for the month of March using centralised option data, calculated at 30 minute intervals. I have attached a screenshot of a simulation using Hegic on-chain data can be seen here, calculating the 30 day dVIX for ETH 10 days into the future (Figure 1). The bespoke smoothing function for the integral that accounts for the sparsity of options contracts uses, where its peak is a function of the median option premium for the relevant expirations (Figure 2). Comparing this to Deribit results (Figures 3 and 4), which does the regular integration method to calculate the implied volatility, yields the same results. We can see that the smart contract method generates the same results, but requires a bit more code / math than just the VIX formula quoted in the whitepaper. I have tried to be brief, and can go into a lot more detail. I’m not sure if I should post my Jupyter notebooks here or not.

My main concern is that the math can be transferred to Solidity (I believe they could, although I am no Solidity guru), or there are next to no options contracts for a certain expiration. This would then have to infer from the contracts that do exist at that point in time.

1 Like

Hey Verto, a 30 day annualized expectation of volatility can be created using the data from Derbit, just wanted to showcase a 7 day due to how quick crypto moves.

1 Like

I love this proposal. I would also love to see a similar follow up that is this same thing but tracks VIX in traditional equity markets rather than BTC/ETH. I think there is likely a lot of demand for people to use DeFi to get exposure to traditional financial vehicles.

1 Like

Hi @rshipp,

That’d be great to do, but wouldn’t be via on-chain data until options exist for that, like wTSLA or something like that.

1 Like

Yeah, for now maybe it makes more sense to leave it to a project like SNX to create sVIX with an oracle.

@ampodgorski & @george

Just flagging the announcement from UMA titled “Introducing the dVIX protocol volatility token”

Here’s the dVIX methodology paper as well.

Would love to hear your thoughts on it.

Hi @verto0912, I’m at a loss to be honest. Their methodology is slightly different and uses a different formula, not on-chain data: at the moment they’re no different from Deribit’s ethVOL or the CVI.

We were in discussions with Set about building the dVIX with UMA. Now they come out with a Medium article 12 days after this post and backdate their whitepaper to be before our 2 April post. The whitepaper even says ethVIX but their Medium article doesn’t match this and seems to plagiarize our name.

[EDIT]: seems they’ve been established earlier, but just under the radar.

I’ve just been introduced to this thread by a community member, so forgive the delay in responding here.

You are right that our ethVIX methodology is using Deribit’s data. The reason we’re doing this is because there are zero robust options platforms with any significant volume, including Hedgic and Opyn. Deribit represents (for better or worse) more than 90% of ETH options volume so we found it indefensible to use anything other than their data. The Cboe’s VIX methodology is not possible to employ in any useful way without robust options markets.

FWIW, we’ve actually been working directly with UMA’s core and founding team (e.g. Hart, Kevin) for months. Our ethVIX index has also been approved as a pricing feed for UMA synths.

(On a side note, I feel a little weird about you using the word “plagiarize” here – if you’d done any basic research at all, you’d have seen that we presented work under this name in January @ ETHGlobal’s MarketMake event.)

Hi Chris,

Thanks for your post.

First off, I apologise for using the word plagiarise. I had an initial emotional response to finding out about your project, and was blind-sided. I editted the above post to state that I was wrong to rectify for anyone who had read that I claimed plagiarism. But perhaps I wasn’t explicit enough. I’ll state it again here. I’ve seen the timestamps of your Tweets etc. and you by no means plagiarised. Again, we are sorry for the mistake.

That’s great that ethVIX has been approved as a pricing feed for UMA synths. What is proposed in this thread is that we have a modified CBOE VIX methodology to account for the otherwise insufficient pricing mechanisms within DeFi for options contracts. Using this algorithm, we can get a target of the VIX with a reasonable amount of accuracy for all assets with a certain threshold number of options contracts.

2 Likes