What the Heck Happened with Bots
First and foremost we should address the elephant in the room. Recently, monsterveritas published a critical report about our platform, for which we are grateful. Let’s discuss the bots first. What happened? Our primary intention was to prevent bot farms from taking advantage of our platform, whether or not Splinterlands added decks to onboard new players. We’ve been largely successful at preventing outright mass bot manipulation and been able to ban multiple users attempting to exploit our platform for large multi-account bot abuse. This leaves the question, how were people able to individually bot accounts or maybe run just a few and were not discovered until now? Simply stated, our bot detection mechanisms weren’t defined well enough to encapsulate large multi account abuse and individual account-by-account or a few accounts abuse. It was a large multi account behavior driven algorithm that did not take tokens into account, nor some very specific individual account behaviors. In short, we were using behavior detection algorithms to detect bots en masse and it was successfully detecting bots, but unfortunately not all. Distinguishing between a player and a bot on a broad spectrum on a behavioral level is incredibly complex especially on a one-off basis.
That changes today. As of today, thanks to the Splinterlands community identifying the issues we still face. We’ve banned 19 accounts abusing our platform with Xbot, Archmage and a variety of other botting services being used to exploit our system. Here are the addresses we banned:
Address
zedorak2
zedorak3
zedorak4
zedorak5
bleublanc
sertinytictac
peakmonstah1
chaosss-agent
bronze-4d
tuzia2
madrax
kaato-kun
ganmik
tub3r0
ryu07
ryuuuuu
ding3284
acd2006
latigo-amigo
We’ve beefed up our bot detection and are ready for round 2! Make no mistake, fighting bots is an incredibly difficult endeavor, as they can adapt incredibly quickly if they figure out how we target them. Our commitment to the community is we will continue to improve our bot detection and will never allow bots once detected to continue using our platform.
Another issue discussed was multi accounting. Our platform policy allows people to borrow multiple decks if they have multiple Splinterlands addresses. Some people want to play on multiple accounts, and we are okay with that! There truly are people who are ACTUALLY playing on more than one deck at a time. Heck, I’ve done it myself before! That said, in the spirit of keeping the community onboard and reducing exploitation we are going to limit the number of Splinterlands provided decks that can be played by a single person to one. This will only apply to decks provided by Splinterlands for player onboarding, and no one else. Here are all the accounts that were multi accounting and have been removed from the Splinterlands sponsored decks:
Multi Accounting
Mayan5
Mayan50
Mayan51
Mayan52
Mayan53
Mayan54
dml0028
dml028
flick02
flick04
flick05
isko010
isko25
jersonps
jersonps0927
kumaliwa26
kumaliwa27
oddish.shroom01
oddish.shroom02
siann28
sianndaniel28
This means that even with botting bans and multi account removals, there are still many decks being played! In summary, not all of these players were botting. There was some multi accounting and limited botting occurring that we have addressed. We’ve already done mass bot bans, and will automate the enhanced detection process as well starting Monday. (We are manually running scripts for now)
Player Onboarding
Next, let’s talk about onboarding new players. I’d like to share our strategy for onboarding new players over the next few weeks and how we are doing so far:
Banging down the doors of ALL player communities and guilds we can get our hands on
Why? There are player communities and guilds out there with thousands of players just waiting to try out Splinterlands. Splinterlands is incredibly difficult to play without assets and the goal is to get players in, get them to try the game and convert a percentage of them to paying players. Well Tyler, why would they pay if they get to play for free? Well, that’s exactly what Free-to-Play is. The idea is to give them JUST A TASTE, enough so that they are then determined to buy their own assets and get into the game fully.On this point, we’ve already established partnerships with two player communities from which we expect to yield a few hundred new players as we onboard them. Here are the communities we are bringing into Splinterlands and the corresponding partnership announcements. This is just the beginning.
Morfyus:
https://twitter.com/morfyusofficial/status/1629048949322440704?s=46&t=YL6h5vdjldwHHYS9lyHYtA
- morfyusofficial
Metanome:
https://twitter.com/metanomy_gg/status/1626112026849460224?s=46&t=YL6h5vdjldwHHYS9lyHYtA
- metanomy_gg
- Pinging all of our influencer contacts to see how they can help us access new players
Why? We have influencer contacts within and out of Splinterlands that are excited about the idea of our platform. They are happy to ideate with us and lend their influence and time to helping onboard new players. - Currently petitioning Splinterlands for an email campaign to all emails that have created an account but not purchased a spellbook to get them to buy a spellbook in order to borrow a deck for free for 30 days.
Why? This one is entirely up to Splinterlands but we believe if we could get a swell of people purchasing spellbooks and renting decks, we can convert a percentage of them to paying Splinterlands customers. - Petitioning our community to refer new AND RETURNING players to the platform so they can start off playing competitively in bronze
Why? This one is very important. There is an uproar about these decks not ALL going to new players, but returning players is a very important metric and here’s why. A 2017 report by Facebook Audience Network found that players who return to a game after a long hiatus are 10 times more likely to make an in-app purchase compared to new users. Another study published in 2016 by SuperData Research found that free-to-play game users who made an initial in-game purchase were more likely to make additional purchases compared to those who had not made any purchases. The report also found that players who had previously made purchases had a higher retention rate and higher lifetime value compared to those who had not made any purchases. Getting returning players is arguably a more lucrative endeavor short term for Splinterlands as they are more likely to purchase, and as a free-to-play user who makes a purchase they are more likely to make an additional purchase and stick around for longer.
Here are some anecdotes of players able to return due to our platform🙂
Why Bother?
There seems to be a misconception that there is a ton of money at stake here or we are profiting heavily from this so let’s look at the numbers. The Splinterlands sponsored decks went live on our platform on Tuesday, February 14th. The experiment has been running for 10 days. In total, the listings have earned roughly 1400 SPS as of yesterday for a total of roughly $39.20. This is across 156 decks all being played. The split for the players is 50% which means they get 40% because NFTy takes a 10% transaction fee. This means that those decks received $15.68 worth of SPS spread across 156 different addresses. This means that NFTy has earned a total of $3.90 from this.
Let’s have some estimation fun now. Let’s say that number was doubled just for funzies and we made $8. That means roughly $.051 per deck every 10 days. Let’s also say there were 1000 active decks provided by Splinterlands. That means that NFTy would earn around $150 a month! WOW! Now we can retire to Fiji 🙂 I can take my wife on that vacation we’ve always wanted!
Joking aside, why build this? Why take so much flak from the community over just under $4 of compensation? It’s because we believe in the future. We know that if we can solve problems for a game developer, our platform becomes infinitely valuable despite the tiny transaction numbers from this experiment. It’s because between the beginning of November and to February borrowers from our platform had purchased 44,951 chaos legion packs, 1,925 rift watcher packs, 10,262,640 credits and 143,137 potions. They had also rented 20,702 cards and engaged in 38,369 P2P marketplace purchases. This is an experiment and we are still accumulating data, but the numbers are promising. If you study any game development best practices, you will find time and time again that free-to-play is one of the best mechanisms for onboarding players into a game.
Summary
We’ve improved our bot detection, because our behavioral detection algorithm was missing some bots. We have banned all exploiters we identified. We’ve shared our plans for player onboarding, some stats of what we’ve already accomplished and talked about why even bother!
I implore the Splinterlands community to help us, not fight against us. We are here for the long haul and we are here to grow the ecosystem. If we are successful in bringing on new and returning players everybody wins. Our platform literally lives or dies by the success of Splinterlands! Instead of finding every reason to throw mud for the pack delegations to our platform, let’s try to see if we can be successful in bringing in more players returning or new and get them to be loyal Splinterlands fans just like the rest of us are!
We are aware of problems, we are addressing them, but come talk to me! If you have additional ideas or thoughts on what we can do to fight bots please share with us. We are in this together, and if we work together, there is no reason we can’t build incredible solutions that have a lasting impact.