* cheating?

[JohnnyCyclops]

Good grief. I just looked at the developers own site. Create a challenge by plugging in a set of 4-5-6 cards, click the ‘solution’ button and bingo. Could be done easily within a four minute real time game, let alone turn based. Then wait to see if the opponent makes 1-2-3 test themselves and then use one fewer tests. Auto-win.

I do understand some players are super smart and can guess with no or few tests BUT this dev provided tool completely undermines any sense this is a fair competitive online game. Would be suitable for a physical game so long as devices are kept off the table.

[Romain672]

The website only give one possible answer, not the real answer. So cheating with it is really hard.

However, there is other ways I will not said here.

[TheFeeling]

They are not necessarily cheating. One can deduce a lot of information from just looking at the verifiers, even without a computer.

And there are only 125 possible codes, if you manage to spend 30 seconds on each of them, you can find out the answer (or a very small number of possible answers) in an hour. This is only cheating if you have a computer that does the work for you, but the numbers are such that it's possible even without a computer.

Here is an example of a non-obvious way you can cross out a code:
1. assume that a given code is the answer
2. find all the answers from all the verifiers for that code
3. find all the codes that would lead the the exact same answers
4. if there is more than one, you can cross out all of them, because no amount of turns would allow the players to find it, and the rules say that this can't happen.

Note that this kind of deduction is implied in the "Advanced strategy" chapter in the rule book.

I understand what you say but if you have a look at their games, both players win ALL of them in 1 turn regardless the level of difficulty. I've watched dozens of replays, I couldn't find one where they needed more than 1 try even with 6 verifiers at most difficult level.

[Jellby]

It's too easy to cheat in this game, and not at all easy to prove it.

[NachitoRus]

I just saw this and I feel I should write as I suppose it originally targeted me as well as other top players.

I think @poptasticboy does a fantastic work explaining it and he has done the same in other posts so I won't take much time for it, but you can solve most Hard puzzles with 0-2 questions. Some require more, but even some of those also could end up only taking 2 questions if you are lucky with your proposal. I get how this can look like cheating for a person who doesn't know how that is done, but it is perfectly possible and the game was designed like that.

Having said that, if you have two players against each other and one of them knows how to do "deep deduction" and the other doesn't, it's like having a beginner playing chess against Magnus Carlsen. Nothing will make sense for the beginner and they will lose before they know what happened. Every time a person asked me how I got the solution so quick, I always try to help them and explain part of my process, but I see how the final experience is not enjoyable for beginners and for that reason I've recommended many players who played against me repeatedly and showed their frustration to red thumb me so the website won't pair us up together again. I've even give red thumbs to some of those players myself because I felt bad. We could blame the matchmaking algorithm for this but unfortunately there are not so many people on the playing queue for the game so I won't blame the website for letting people play instead of keeping everyone waiting. Maybe letting every player finish the game would at least leave a better taste, rather than abruptly stopping the game because someone else solved it. I don't know how this is handled in similar games in this website but could be a good implementation.

Regarding how I solve the puzzles, I will answer a few quick questions: No, I don't cheat. The reason I have such a high rate of wins is because most people I play against don't know how far you can go with deductions. If you check my games against other pros, you will see I tie a lot. And no, I don't use the official Turing Machine website. I wasn't even aware you could check solutions there and from the quick look I got, it seems you just get one of the possible solutions without knowing if there are more, so not very useful in my opinion but maybe I'm missing something. Also I don't use any specific software for this game. I've seen some people have uploaded programs that solve every puzzle but a) I'm pretty bad with computers and without an instruction book I don't think I'd be able to use them and b) they would just get the same result as I get (if they work properly) with my own method. I help myself with an Excel sheet as the space in the playing board is not enough for everything I need to write down, but no brute force calculations are done by Excel. I've considered sharing my method here but I've decided not to at least for the rest of this season. But basically involves solving the puzzle analytically to a great extent, and then testing a bunch of possibilities one by one and in a very optimised way. I've studied physics and mathematics and my main reason to play the game was the challenge of solving it with pen and paper. I have manually derived excluding rules for every combination of 2, 3 and 4 verificators together, which generally leaves me less than 20 possible solutions once I apply the corresponding ones. After that I test each of them with a quick test that involves just a few algebraic calculations. The whole process takes me less than 5 minutes if I do it relaxed, and I've tested myself to do it in live games and I can do it in less than 3 minutes for most games. I don't think i will share my results as I don't think it's fair that people who don't understand it use it, but whoever is curious might find it a challenge to develop a similar method by themselves.

Additionally, on my personal quest to be the #1 on the leaderboard I have analysed almost every other top player in order to know who does "deep deduction" and who doesn't. I understand that some people will think I'm taking it way too far and it's not fun to play against someone like that. Again, I can only blame the matchmaking, because at some point players who take it serious should not be paired with beginners. Also I try to avoid casual games and mostly play Arena to avoid players who just want to have fun and not compete. I have talked with a few other pro players and they all used different methods including specific software or manually checking the 125 possibilities in an Excel file. They were nice enough to answer me so I won't call names.

Now, regarding CHEATING: I've actually very recently found out that there is indeed a way to cheat beyond using brute force software. I started investigating after I saw truly absurd guessing chances (I have myself gone for guessing quite a few times based on probability to avoid ties while trying to get the 50 wins in a row achievement, unfortunately failing many times) and with help from a friend we realized there is a way to instantly know the solution. I have written to the BGA game developer about this directly and I hope he will fix it soon. I have tested it against people who i know are using it (to avoid losing unfairly) and also analysed already finished games and it totally works, although it would require some time to have a complete dictionary of solutions. You are very welcome to accuse me of using it, but if I was using I would already have my 50 wins in a row. I just lost a streak of around 45 wins because I had to guess to avoid a tie and I wouldn't have to guess if I knew the answer. I also won't call names here but I have a list of people who 100% are doing it (or another cheating method that I don't know of).

I hope this post clarifies some things. I will be very happy to answer any questions that don't involve revealing too many details of my method. I hope that is understandable. I want the best for the community here and I will be very happy when the cheating issue gets resolved. Also I would be happy if the Arena games didn't show which verificators the opponents are using so games are not about outwaiting the opponents. This is something I do myself because you can get info from what others do if you know they have done the "deep deduction" but I agree it makes the whole experience less enjoyable. I would sacrifice some bits of information if everyone has more fun (and also so i don't give away information).

[JohnnyCyclops]

Hi Nachito, thank you for your detailed and nuanced post. I've stayed away from this game in recent weeks because of the massive imbalance between beginner and experienced players, much as you explain. The chess analogy is useful. I do think allowing all players to complete their own play would help, even if the winner is 'excused' from the table to move on to their next game while the slower player completes their game/test.

At the back of the official rules there's a Handicap System*, which states:

If an experienced player is playing against newer players, we strongly recommend using this system to level the playing field.

In the first round (and only the first), the experienced player fills in a number of answer squares. These squares count as questions the player would have asked (leaving fewer for the player to ask in the first round).

Fill in 1 square if the difference in experience/skill is slight, and 2 squares if the difference is greater.

For example, in the latter case, the player could only ask one question in the first round, because there are 2 squares already filled in!

If that was implemented in BGA, likely using ELO to determine both experience and inexperience, it would go some way to levelling out what the designers clearly understand to be a material advantage for experienced players.

* I'm not a fan of the term 'handicap' given its negative disability connotations, but it's what the designer chose, so I'll go with it.

[poptasticboy]

At the back of the official rules there's a Handicap System*...

I'm not sure this would make much difference to players who don't use deep deduction, as the deep analyser would still frequently solve the puzzle in 1 round, but it would hugely (and unfairly) impact players who *do* use deep deduction against each other. There is a minimal number of questions required to solve a puzzle (barring the horrible "wait and see" tactics that I can't stand!) and anyone legitimately able to find that optimal solution should be credited for it. This system would cause the higher ranked player to get a loss through no fault of their own.

The handicap system may be useful in a real life, casual setting, but I don't think it's useful online when deep deduction is in the mix.

For the same reason, I feel it's unfortunate that players lose ELO when they tie someone lower ranked; it's not that type of game. It's impossible to "play better" when both players play optimally. But I digress

[poptasticboy]

Maybe letting every player finish the game would at least leave a better taste, rather than abruptly stopping the game because someone else solved it.

Agreed.

And no, I don't use the official Turing Machine website. I wasn't even aware you could check solutions there and from the quick look I got, it seems you just get one of the possible solutions without knowing if there are more, so not very useful in my opinion but maybe I'm missing something.

Most of the time it's as you describe, so it's more of a "sanity check" that you haven't done something silly. Of course when a set of criteria cards only has 1 possible answer (or even only a few) it would allow someone to get it without going through the deductions themselves, which is unfortunate, but it's not foolproof, and either way it couldn't *beat* someone deducing optimally.

However, there is occasionally a situation where not every possible valid solution to a set of criteria cards has actually been included as a possible puzzle and assigned a puzzle number. In these cases, someone who abuses Create a Challenge will indeed find the solution more quickly than someone deducing fairly.

I will only ever submit a solution that I can deduce fairly. At least twice now I have known what the answer would be because of Create a Challenge but refused to use it because it is not deducible. My opponent however has chosen to "guess" in fewer questions... and of course got the right answer, because it was no "guess" at all.

Also I would be happy if the Arena games didn't show which verificators the opponents are using so games are not about outwaiting the opponents. This is something I do myself because you can get info from what others do if you know they have done the "deep deduction" but I agree it makes the whole experience less enjoyable. I would sacrifice some bits of information if everyone has more fun (and also so i don't give away information).

This is my number one issue at the moment, to the point where I'm actually currently taking a break from playing this at all on here and reassessing if I wish to continue.

I have zero interest in playing the "wait and see" game and learning things from what opponents do; it's just a different game and not one that I have any interest in playing. Ultimately in a solo setting there is an optimal solution to each puzzle; I am not interested in playing opponents who either "guess" or "learn from opponent's plays" to solve faster than the optimal solution. It sucks all the fun out of it for me, to the point that I have reached now where I'm no longer sure if I wish to continue playing unless this is dealt with.

[JohnnyCyclops]

At the back of the official rules there's a Handicap System*...

I'm not sure this would make much difference to players who don't use deep deduction, as the deep analyser would still frequently solve the puzzle in 1 round, but it would hugely (and unfairly) impact players who *do* use deep deduction against each other. There is a minimal number of questions required to solve a puzzle (barring the horrible "wait and see" tactics that I can't stand!) and anyone legitimately able to find that optimal solution should be credited for it. This system would cause the higher ranked player to get a loss through no fault of their own.

The handicap system may be useful in a real life, casual setting, but I don't think it's useful online when deep deduction is in the mix.

This system is part of the rules irrespective of venue (in-person or virtual).

I don't see the issue of two (or more) players using deep deduction being hampered by this system as anyone capable of deep deduction is by definition experienced/expert so no scoring handicap would apply.

For non-Arena games players can limit ELO level of opponents when opening a new table and this can help a less experienced player coming up against someone using more advanced techniques. That's not possible in the Arena where it's just raw matching.

A two-point handicap of the experienced player would create a draw/tie on what otherwise would be a 1-3 score (bringing it to 3-3). This gives the less experienced player a better chance to win but is no guarantee. This system levels it up a bit, as the game designers clearly intended. For those concerned with the intellectual puzzle rather than with rankings then the solo game fully meets their needs.

All of this is to say, and as I felt Nachito alluded, that with a large skill imbalance I foresee the number of people wanting to play competitively will shrink, providing fewer and fewer opponents for those still wanting to play competitively. In the long-run it serves no one to have a large skill imbalance. The designers provided a rule option to address this.

[poptasticboy]

All of this is to say, and as I felt Nachito alluded, that with a large skill imbalance I foresee the number of people wanting to play competitively will shrink, providing fewer and fewer opponents for those still wanting to play competitively. In the long-run it serves no one to have a large skill imbalance. The designers provided a rule option to address this.

I get the motivation, but I can't imagine any kind of handicap system that would sit well with me; if one can solve a puzzle in the minimum number of questions and yet still lose, then it's a flawed system. Handicaps just aren't really appropriate for this type of game.

Yes the designers provided the option, but they didn't design the game and rulebook for an online competitive setting. They made a physical game and wrote a rulebook appropriate for that. The whole "wait & see" issue is proof enough that online implementations are not always directly comparable to original physical games, and sometimes need tweaks to make them more appropriate to the new setting.

In a physical real-life setting people aren't typically going to be getting optimal solutions via deep deduction; they'll be testing and making deductions based on the tests. In this setting it may well be appropriate to use the kind of handicap system suggested to help players at the table have more fun, but I don't believe it works in a turn-based setting where people have as long as they need to use deep deduction. I for one certainly wouldn't want to play under those conditions.

To be honest there's just a fundamental difference between players who enjoy using deep deduction techniques to solve puzzles, and those who want to play a game about testing and then making deductions from the results; ultimately the two types of players are looking for different experiences from the game and I'm not sure there's a satisfactory way to reconcile the two, other than limiting people only to play opponents who are looking for the same experience.