Rebalancing of $Play

There has been some discussion in discord of rebalancing $Play, as it currently has >40% of the token weight (down from 50’s yesterday). To put this in perspective, AXS original allocation was ~8%.
Just wanted to start a discussion on here about if/when a rebalance should happen. Personally think AXS as a current ‘winner’, shouldn’t be entirely sold off to buy current cheaper coins, as it is likely priced higher for a reason (high adoption and high revenue). That said, having $Play be >40% weighted to one token is not great either. I would move for a rebalance to have AXS reduced to <20%, but happy for someone more knowledgably to run some numbers here.

3 Likes

Agree, rebalancing seems reasonable. Maybe we should distribute between MANA, SAND, ENJ, and UOS, with a small boost on the last one as it seems to be slowly incresing for the last month.
Yet, I would get someone moreknowledgeable to drive a better informed decision

1 Like

Hey thanks for proposing this. I personally see both pros and cons of performing a rebalancing during a sort of hype period (relaying here an interesting piece on AXS previously shared by @Cass), but yet agree that the current uneven allocations look too skewed towards recent runners-up.

By applying the same methodology followed to define the initial $PLAY allocation (30days adjusted Market Cap) we would get the numbers below, where the last column highlights the rebalancing effect vs current $PLAY allocation.

Data below, waiting for community feedback.
Definitely open to discuss different weights for this rebalancing.

3 Likes

I’ve long been a proponent of declaring some principles for rebalancing.

  • AMMs rebalance constantly
  • Some traditional index funds do a version of this by rebalancing whenever allocations drift a certain % away from the index amount
  • Some traditional index funds rebalance once a quarter or once a year regardless of whether the allocations have large or small drifts
  • Many traditional indexes that define the target allocations for index funds only change their target allocations once a year
  • Some traditional funds do not follow indexes and instead update positions to take advantage of internal ideas & research in search of alpha

What are we going for?

2 Likes

I think applying the full delta rebalancing in the table @gabo shared would be too extreme a change in weights. Using an exponentially weighted moving average (EWMA) would weight recent prices more heavily and lower the delta.

Alternatively, weight the current price with the 30-day average by some ratio (eg, 60%/40%). So for example for AXS, using 60% AS-IS and 40% 30-days adjusted market cap: results in a target weight of 31.68% using the values in the table.

My main concern is a kind of thrashing where the index sells off winners, but then needs to rebalance back into them later if they continue seeing exponential growth. I expect the PLAY pie will eventually be dominated by a couple winners that nail the tokenomics and “catch lightning in a bottle” like Fortnite or World of Warcraft. I’ll echo that the article @Cass shared seems particularly well-researched and makes a compelling bullish case for Axie Infinity.

I also think 40% is not too unreasonable an allocation. It’s not (yet) as bad as SUSHI weighted as 60% of YPIE. But perhaps there should be some rules for rebalancing when a single asset becomes 40%+ of the index.

2 Likes

agree with @fairshare . we want the winners to run so don’t want to cut too much–especially after the drawdown of the last few days. over 40% sounds like a good time to rebalance. maybe take anything that hits the 50% mark down to 40% and distribute to the other allocations. that way you’re selling pumps.

Reading this yesterday I agreed with the thinking posted above. But I also checked the Axie chart, and today I see that it made a new high, now at $30.31. Based on that I’d suggest we wait and see how it runs. This could be another big wave and afaik rebalancing is like cashing out. Better to do that when it looks like Axie has topped or plateaued.

2 Likes

Very much in line with your reasoning :ok_hand:

1 Like

This is a very bad idea for an index product IMO. What you are describing is timing the market and it’s a losing strategy over the long run.

It’s essential for an index to have a rules-based rebalancing strategy. E.g. 40% max per asset is reasonable (perhaps too high).

I’m amazed there’s no policy in place on this already, frankly.

2 Likes

Agree that waiting for the high or a plateau is just an attempt to time the market and should be actively discouraged. It rarely has good outcomes and is not the point of an index.

That said, I dont see that the ‘lack of a policy’ for rebalancing is an issue. It is stated by the Dao that anyone can call a rebalance at any time via governance. Which I think, is what we are doing here…
Once the terms are agreed we can move forward with rebalancing based upon that. The alternatives would either be a more traditional non-dao model, which kind of defeats the point, or to have had this discussion before $Play was launched, which would have delayed things and I dont think was required at the time.

In general, I am not overly concerned with the ‘extreme’ change proposed by @gabo, but would be interested in seeing the full run of EWMA weights proposed by @fairshare, purely as a comparison?

More interesting to me is laying out the principles for a rebalance proposed by @jnova. My own views based upon that list I would be criteria like:

  • No more that 1 rebalance every 6 months.

  • Only if a rebalance hasn’t taken place within 6 months, and weights are skewed since the last rebalance

  • Skewed defined by 1 token constituting >40%, or 2 tokens a make a combined 75%

  • After rebalance agreed on given date, actual rebalance to occur after 15 day holding period (as long as it still meets the above criteria)

  • Rebalance to conform to standard model (to be decided)

Although I get the idea of a few tokens that contain ‘lightning in a bottle’ and letting them run. That is not much of an index and will need rebalancing to some degree either way in the longer term (i.e. World of Warcraft or Fortnite have or will decline and others will undoubtedly take their spot). I feel that Crypto moves somewhat faster than tradfi, and so would propose a rebalance be looked at more often, but still with limits (i.e. no more often than every 6 months?).
I dont see a rebalance as selling off the winners, more as taking profits on the way up and diversifying that profit into other potentials (that may later take the crown…). As long as it is done infrequently and without bias. The delay to rebalance after agreement might act to avoid thrashing, bias or attempts to time the market.

Only some ideas, would welcome further discussion!

4 Likes

100%. When nearing ATH is a good time to rebalance so you take advantage of the upside, but protect from the downside. If continues to run, you still have a 20% allocation and in theory the other coins play catch up. @gabo how quickly can we get rule in place to rebalance once anything passes over 40-50%?

I’d be happy with any plan that we can apply again in the future to avoid market timing.

To me PLAY is an emerging market fund and a bag of hopefuls and while I’m not particular, I don’t want any holdings to starve and not have the chance that Axie had. Maybe a maximum threshold for a single asset and limiting the (currently 3) large cap holdings to 70% or 80% of the pie or something like that? Deciding on any method we will repeat with PLAY in the future would be a good outcome for me and I don’t mind Axie’s allocation while we work it out.

2 Likes

Here is a PLAY input file for Tiramisu: https://api.npoint.io/dbeb0fa6adb72a726f89 in case anyone would like it.

market caps:

3 Likes

Any thoughts on adding ILV to the index? I think Illuvium is an exciting project with a great team, and it’d be awesome to get some exposure via PLAY.

1 Like

Following up to the previously shared rebalancing proposal, please find here some updated numbers based on the actual allocation of $PLAY (AS-IS), which evolved quite significantly from last week.

Based on the multiple feedback collected here and on discord, this new rebalancing proposal includes the following:

1. Increase of the max allocation threshold from 20% to 30% since almost unanimously considered as quite restrictive for a proposed rebalancing (see my previous proposal)

2. Modelling of rebalancing allocations according to 2 options
- 30days Adjusted Market Cap (OPTION A - current methodology)
- 15days Adjusted Market Cap (OPTION B)

As reference, both options bring the cumulative allocation for the 3 larger cap holdings to ~70% (down from the current 80%).

I would personally consider the OPTION B as a good compromise for rebalancing, since redistributing the upside from runners without excessively affecting their momentum, while reallocating gains to smaller caps as an indexed product should be doing.

NOTE: From simulations performed, the 15 days Adjusted Market Cap (OPTION B) basically reflects an allocation based on a “Blended Market Cap” resulting by weighting in the 30d Market Cap and a lower range Market Cap (I did consider the 7days). Props to @fairshare for suggesting this perspective.

I’d propose to move this to vote, in order to expedite the rebalancing and avoiding further skewness induced by $AXS.

Shall some consensus be found around one of the proposed options, rebalancing could be ideally taking place as early as on this week, in conjunction with a required maintenance to be performed on the $PLAY index in order to reflect the migration of both $PUNK-BASIC & $MASK assets to the vault v2 recently introduced by NFTX.

5 Likes

The introduction of additional assets to $PLAY wasn’t here discussed yet, but I’d be happy to take this topic on after this initial rebalancing takes place.

$ILV definitely represents a good potential candidate for inclusion, as already requested by multiple community members.

3 Likes

PLAY started as a product to provide Gaming and NFTs exposure for investors. It did that pretty well and I understand there is now a desire to further narrow down the product and establish a policy for rebalancing, I’m not surprise to see different takes about that and I think it’s an issue with definitions.

Coming to a better definition of the product will in my opinion make it easy to define the rebalancing policy.

Is PLAY really an index?
I never liked the word index to describe Pies (even if I used it), I think it’s an incorrect term to describe tokenised portfolios which are Pies. Index investing have their tokens weighted based on their market capitalization, while we used that as a starting point to decide the initial allocations for Pies, as a community there is a clear desire to implement more criterias for weights and rebalancing than simple market capitalization making products like PLAY more similar to a smart-beta strategy.

The community input adds an element of active investing to passive investing.

Smart beta is also rules-based but the strategy can consider things like momentum, volatility and quality of an asset. I find super important to crack this point first, as it naturally guides the decision on rebalancing, for instance @kitblake makes a good observation about the state of the market and while I don’t believe trying to time it is advisable, factoring momentum into the rebalancing strategy is a good idea if properly measured.

What do you rather have?
Option A: Simple market-cap based weights, fixed rebalancing schedule (every N months), better suited to representing the market, requires little to no community input.

Option B: Rules-based approach taking in consideration multiple factors, dynamic rebalancing approach (when value X overn N%), aims of of lowering the costs (no uncessary rebalance) and generate alpha, requires community input.

  • Option A
  • Option B

0 voters

5 Likes

Yeah, naming… We probably shouldn’t refer to these tokenized portfolios as indexes. Actually the word ‘Pie’ works really well. It captures the combinatory but selective aspect.

Of course it would be possible to setup a Pie that’s a straight index, e.g. composed of the Top 10 coins on CoinMarketCap. The criteria would be crystal clear and the ‘Pie-index’ wouldn’t need any management or community input. It would just run and adjust whenever there was a change in the Top 10.

But for typical Pies I appreciate hearing from the community (and especially in this forum). When @ChrisSavadge posted his constructive critique it really made me ruminate. Others responded with various ideas for criteria and I thought we could work out a list.

This could be taken to a logical extreme and the criteria could be coded. So IF the rules are quantifiable THEN we could put them in a smart contract. This would work fine for a Pie-index like the Top Ten. But for $PLAY I think it falls squarely into the “really bad idea” category. :slight_smile:

As @alexintosh stated, there are other factors in play that we’d be wise to consider. Some of these are intangible, like buzz, sentiment, or bullish patterns in a chart. For these non-quantifiable factors I depend on the community. Personally I don’t have the knowledge or experience to invest in this subsector. But with the wisdom of the group it’s possible. It becomes a managed risk.

3 Likes

Option B works. Let’s do it.

4 Likes

Yes $ILV and $YGG would be very nice additions.

On a personal side. I don’t really see the point of having tokens with less than 1% allocation. I would even say 5% should be a minimum. Having 9 tokens with less than 2% allocation, looks more a lot of work in terms of rebalancing than a real advantage for the Index (Pie)

Then between A and B i don’t see a huge difference. So …

1 Like