EtherOrcs Orctober Recap: Poof’s Game Theory vs Actuals
It’s hard to believe we only minted the first EtherOrc three weeks ago. What an awesome Orctober we had. It’s been a wild ride, and we are only at the beginning!
As a data nerd, I wanted to give a quick retrospective about what went down “behind the scenes”. I am going to cover what we had initially modeled vs what actually happened within two key areas: minting and pillaging.
Key Question 1: How long would it take to mint all 5050 Orcs?
What we expected: 9 to 10 days
What happened: ~8 days
This was an interesting one. We’d gathered a lot of valuable insights from Anonymice on how fast people could mint (and burn) through a utility token based process. So we knew a few things going in…
1) The ZUG-to-mint-cost ratio needed to be exponential in order to keep up with the curve of new Orcs minting
2) The % of staked Orcs at any given time would be a key factor
3) The ratio of Orcs farming vs training could totally change the outcome of the game
We wanted minting to take roughly 9–10 days. We established the number of Orcs in each minting threshold so that the minting would last roughly in line with our expected timeline. We modeled assuming the extreme scenario where everyone works together and transfers ZUG at the optimal rates whenever possible.
Here’s an example of what the “average” curve looked like in our model to get to roughly a 9–10 day minting period with some simple middle of the road assumptions.
Example shown below is: 80% of all Orcs staked. 50% of Orcs staked are farming (vs leveling). All new Orcs minted either go straight to farming or level up, pillage, and then farm.
Interestingly… my estimate of ~50% of Orcs farming throughout the mint process was accurate. So what was different that led to all Orcs being minted a full day early? Well there was one thing I didn’t account for (that blew us all away)…
Virtually every single one of you lovable Orc herders staked your Orc to farming or leveling. We had an average of 98% Orcs staked during the minting phase. Who has the time to sell or idle your Orc when there’s ZUGGING to do?
It’s unbelievable how so many smart Orc owners stayed engaged for the entire time with their Orcs. And this is with a complex contract to interact with! A whopping total of 58,720 transactions have been processed in the game to date! Amazing!
Even more impressive… we still have over 97% Orcs staked at the moment I’m writing this…
Key Question 2: How would people engage with the high tier loot pools?
What we expected: Lots of burned gear at the higher level pools and a slower burn on the early pools
What happened: Less gear burned at the higher level pools and a faster burn of gear in the earlier pools
This was an interesting one. We knew that people would burn for new traits, but we weren’t sure the amount… And since gamifying rarity was a critical element of EtherOrcs, we had to be tactful to get this one right otherwise we could end up with too much gear in one place or another.
And just to make it spicy, we built in the secret pay-ZUG-to-use loot pools in advance. This could make planning ahead difficult for those who were not astute readers of the contract.
We tried A LOT of ways to model this. And I mean A LOT. Many hours debating small details of the pools, how many of each item, if it was the right leveling curve, etc.
Ultimately, what proved most helpful was to run a variety of simulations on what I call “the Orc wave.” Orcs focusing on leveling created a giant wave that went directly to the highest level loot pools (namely tier 8 since the others were time locked). A certain percentage of them fell off along the way — going for the safe option of looting a lower pool. Once the wave hit the final available pool they would loot it until it’s empty and then move on to the next most rare pool and so on (essentially the wave splashes against the wall of the final pool and then works its way backwards).
Once the looting begins, Orcs can have one of two behaviors: polite or degen. Polite Orcs roll once and then call it a day, no matter the outcome. Degen Orcs roll every slot until they achieve the best possible outcome for a given pool.
What it showed us was the following: We needed a cooldown long enough to slow down nonstop pillaging. The Degens always dominated otherwise. Also we couldn’t ever eliminate the most extreme edge cases. As an example, if everyone only looted each pool once (polite Orcs) the rarity curve would be heavily weighted towards top tier items.
So what did we prepare for? Everyone to be a “degen orc.” What actually happened? A mix of a few different things…
1) We were prepared for 300 Orcs to go all the way to the end and race for the Ether — in reality, it was closer to 80–100 Orcs. We also found that these Orcs were a little bit more polite there than we expected, only rolling the later pools a few times.
2) Early pools that were limited (e.g. the Crypt) emptied several days sooner than expected as the community built some pretty thoughtful strategies on leveraging Zug bonuses to help their farmers at key moments
3) There was about as much pillaging / re-rolling as expected — the median Orc who pillaged did so about 7 times in total…
Overall: what does it all mean? The Horde is incredible and of course managed to outsmart our planning.
What I’m not mentioning is that just about everyone in the community had developed their own home cooked strategy and management method. You are all the smartest damn Orcs I’ve seen! And that’s part of how you all managed to beat our expectations on just about every turn. Big brain moves!
Now… back to testing out some more models for RAIDS… Orcvember is here…