This is a very long post that I have wanted to write for a long time, but laziness, lack of motivation and my little interest in the economy of Splinterlands were preventing it from being written. So for those who don't like to read anything long, here I give you a very brief summary of what I want to celebrate with this post:
In the latest Splinterlands update it is no longer possible to submit your monster lineup after the official time to create your team has expired. This can change the entire dynamic of the game by making it a little more fair. Even if it doesn't end up solving the main problems, it definitely points in the right direction.
That's all! That is the reason that made my day when for the first time I no longer had to wait because the back door that was open for too long, being exploited by many bots, was finally closed. Now, if you want to understand why I am so happy, you will have to read the entire post, understanding all the paranoia that led me to see patterns and clues in each of the battles, and create strategies that work very well for me in leveling my battles against the bots.
Disclaimer: This entire story that I am going to tell you is based exclusively on my extensive experience battling in Splinterlands, almost everything is based on patterns, clues and speculations fueled by my emerging paranoia after losing thousands of battles that are no longer mere statistical deviations. To a rational person this should be nothing more than baseless speculation, but after reading this post, let me know if you really can't see Bigfoot's silhouette clearly in the background.
When paranoia starts to arise
If you're an experienced Splinterlands player like me, with over 20,000 battles under your belt over several years, without having used bots or helpers, then you clearly understand that feeling that something is wrong with the outcome of most of those battles. That feeling that I think is quite generalized is our first clue. Of course, it's completely unfounded but I usually go by my instincts and clearly something doesn't add up.
The most recurring question in my mind every time I lose a battle is, for example: How could my opponent know that I was going to use magic? That team is perfectly structured to counter my monster lineup. Of course, if you've been using magic or the same lineup of monsters in the last few battles, it's easy to predict what you'll use in the next battle. If that's the case, the mystery is solved because you are totally predictable.
But the problem arises when you decide not to try to win but to send out a completely random lineup of monsters, with the goal of seeing what kind of counteroffensive your opponent launches. Surprise! Again, if this is one of the countless bot accounts that play Splinterlands, it turns out that the monster lineup they use is like they know what monster lineup you'll be playing with. This is where my doubts started, but in the end I thought no, it's not possible that they were reading my monster lineup beforehand. How is that possible?
Now, playing Splinterlands wasn't about winning or losing, because I almost always lost to bots no matter how good my monster lineup was, but about solving the riddle of whether or not these bot accounts could read my monster lineup beforehand.Here are the first symptoms of paranoia, the conviction that my opponents are reading my monster lineup in advance.
Of course, losing one battle after another takes a toll on your self-esteem. If you are a rational and fair-minded person, you begin by analyzing all possible alternatives that can explain the outcome of those battles. At first you begin to believe that maybe you are not as good a player as you think after all; that your opponents use better strategies than you because we already know that the entire structure of the game is unbalanced and a level I team can easily beat a team with higher levels. You even come to think that it is simply a matter of luck and that soon it will be your turn to win. But no, nothing changes and your thoughts become negative and increasingly accurate about the game's flaws, almost convinced that the best thing is to simply abandon the game. The boldest ones end up ranting about the Splinterlands ecosystem.
But despite all this, you continue trying with new strategies, renting cards, making changes that you consider great improvements that should lead to different results, but no, everything remains the same. The only thing that changes is that your paranoia has been considerably reinforced.
What was interesting was that this "reinforced" paranoia added a new element to my play in Splinterlands. Now my interest was to discover if my opponents could know my monster lineup beforehand, which I found much more entertaining than losing battles.
Thus, for months I dedicated myself not to fighting but to systematically searching for clues and patterns that would help me verify that my opponents were able to read my monster lineups.
In search of clues and patterns
In order to find clues and patterns, the first thing I had to do was commit to myself to play at least a couple of hours every day. And I did so for many months.
The first pattern I identified was the fact that the first two or three battles in each game session were the fairest. That is, I won if my team was stronger and better structured, both in soummoner and monsters, and I lost if my opponent clearly came with a team superior to mine. That's normal, how battles are supposed to be. But, after those first two or three battles, things change radically, and a string of battles lost one after another begins. Almost every day the same pattern repeats itself, without fail.
So, I adopted a strategy that seemed smart to me which was to only fight three battles at most per session and wait at least a couple of hours to play again. The results were surprising, my percentage of battles won once again reached 60% as it had historically been. But, it didn't turn out to be a good strategy because I was wasting a lot of time playing just to avoid them reading my cards, and the idea was not to win more battles but to figure out what they were doing to read my cards.
My first clue
I started doing more complex analyzes to figure out if certain opponents were pre-reading my cards. Logic dictates that if my opponent is reading my cards, he must necessarily be sending his monster lineup after I send mine. Therefore, those suspected of reading my cards were always in second position on the battle screen, because Splinterlands always places the team that submits its lineup last in that position. Now, I had a tool to identify the possible candidates who were reading my cards lineup.
When you have vast experience playing Splinterlands, just knowing the rules of the battle, you are able to put together a lineup of monsters for those conditions in less than 15 seconds. And that's just what I used to do, and I immediately clicked on battle. THAT'S THE MOST SERIOUS MISTAKE I WAS MAKING! First of all, because if they are reading my cards, giving them more time to do so guarantees a victory. Second, because this way almost all the players will send their lineups after me, and therefore I will be unable to identify those who are reading my cards.
So, I systematically started sending my monster lineups at 20 seconds, then at 30 seconds, and so on until the time was practically running out to just 1 second. With this methodology I found my first big clue. There were a large number of accounts, almost always related to bot farms, that submitted their lineups very accurately just 1-2 seconds after I submitted my monster lineup. After hundreds of battles I didn't have the slightest doubt, they definitely knew exactly when I was sending out my monster lineup.
But, the really important thing here is that if they know exactly when I am sending my lineup of monsters, it is because they have already fully identified me, I mean my user account, otherwise it would not be possible to know when I have clicked on battle. And if they know my username, then they already know all the monster units I have under my belt. Knowing my monsters, it is much easier to guess or predict what my alignment is going to be for that battle with those particular combat rules.
My paranoia continues to grow. At this point I was more convinced that those bot accounts were definitely reading my monster lineups. But I still didn't know how they managed to identify my username so they could do their tricks. I will talk more about this matter of how they manage to identify me later, because now I want to focus on one of my most terrifying discoveries that made me seriously think about stopping playing Splinterlands. In fact, for a long period of time I was convinced that this entire game was nothing more than a big, vile scam on players.
Were Splinterlands developers colluding with bot farms?
One of my most surprising discoveries was the fact that bots were capable of violating the game rules in Splinterlads. How is that? Well yes, in my search for which accounts were the ones that sent their monster lineups after me I found that the vast majority were able to send their team even after the time had expired. So, I started sending my lineups even after the legal time to send the team had passed. At first, I started with a couple of seconds after the time was up. Surprise, the bots were sending their team after me. I started sending my team 8 seconds after the time had run out; surprise again, the bots were sending their teams after me. I kept increasing the time, until I reached a limit of 15 seconds, just at this time, the bots were no longer sending their teams after me. Apparently I had reached the limit, because on more than one occasion I waited up to 18 seconds and had no problems.
I finally adopted the 15 second limit because it was the safest time to send my monster lineup without the bots sending their teams after me, and above all I didn't run the risk of losing because the battle was declared abandoned. Needless to say, this became my battle strategy over the last year in Splinterlands.
I never understood why Splinterlands had left this backdoor for bots, since apparently the sole purpose of its existence to me was to give the bots enough time to read all the players' battle lineups.
It didn't represent any additional advantage for me, but it did allow me to replay the game, since I was finally playing one-on-one against the bots, knowing that they were no longer able to read my battle lineups.
I suppose that the few who have read this far will already understand why I believe that the latest Splinterlands update is the most important in a long time, it clearly marks a change in direction and makes the game fairer for everyone.
This update has restored some faith in Splinterlands and their team, because I'm assuming this was all an inadvertent mistake that they had no knowledge of. Let's say giving them the benefit of the doubt.
How do bots read your monster lineups beforehand?
Everything I have told you up to this point only proves in my opinion that the bots are capable of identifying user accounts beforehand, and most likely are able to read your monster lineups. But it does not give any indication of how to identify a user.
This is where the first pattern I described at the beginning of this post becomes important. Every time you start a new game session, after a relatively long pause, the bots don't seem to be able to read your monster lineups beforehand. Therefore, in those conditions that are fairer and the bots do not have the advantage, the results of the battles are given by the strategy and skills of the players. After 4 or 5 battles the problems arise again, because it seems that they are already able to identify the user.
My hypothesis is that bot farms are using the last 5 battles that the user has participated in to identify them. I imagine that the procedure they use before each battle is to read all the last n battles in the battlechain, then they hit battle, and with this they have the last 5 battles of their opponents with the specific combat rules for those battles. With these very specific features it is possible to filter and quickly locate those 5 battles if they have occurred within the time range they have in the battle history. When the matches are found, the user is already lost, because he has been identified. With the username you can obtain all the information necessary to predict or guess the next fight.
In my opinion, that would explain how it is possible that some Splinterlands bots have percentages of up to 90% of won battles, a percentage that seems extremely excessive to me if all the possible combinations of cards, combat rules and strategies are taken into account. And if those percentages were real, that they are not cheating of any kind, then we should give them a Nobel Prize because there is nothing in AI that is as successful as these bots are.
This is how I believe these dishonest co-players are able to identify their opponent. That clearly explains the pattern I have observed, that after several battles in a row, a string of lost battles will inevitably overtake you.
Did this change dispel your paranoia?
Not at all, in fact I continue developing strategies to obfuscate bots. One of my favorite strategies right now, simply because I enjoy doing it, is to click on battle and after looking at the combat rules, quickly go and rent a card that significantly alters the development of the battle. This way I level the battlefield more because there are monsters that the bots don't know that I can use for that battle.
Conclusions and recommendations
Definitely, in my opinion, this latest Splinterlands update has been the most significant change in recent months for the Splinterlands ecosystem: A backdoor has been closed and it makes the battles more fair and even in the game. No one can send monster lineups once the time has run out. Now players only have to wait until one second left to click the battle button and bots are much less likely to read their opponent's monster lineup. Of course, they are still able to predict your possible lineup because it is very likely that they know your arsenal of monsters. But right now there's no way to prevent this, unless Splinterlands decides to obfuscate the opponent's battle history in some way.
Finally, a reminder. All of this that I have explained here is nothing more than an analysis of the results of my battles, which following certain basic logic and methodologies allowed me to reach certain conclusions over three years, which in the best of cases can only be consider them to be unfounded.
And I can't leave without asking; Did you get to see Bigfoot?