Subset Games Forum

... right or wrong? Let's discuss. Now that the update will arrive soonish, I thought I can break my self-imposed abstinence a bit earlier and heat up one of the always recurring topics.

I think you are wrong if you claim all games can be won, and we are not yet talking about specific ship types. I think games can be lost with all ships regardless of your actions. I conclude this from my own experience as a super good player *cough*, and I devised a clever empirical simulation approach to show that the statement that not all games can be won is true (see further below).

What is the evidence so far? From all players world-wide, as far as we could obtain the data, the longest published winning streak is 18 times in a row, using different ships. With the same ship type, there might be longer streaks known, please come forward and report your number if you are reading this. So we have n=1 (a single experiment) that shows that one can win with all ships in a row without losing once. That is what we can gather from this single experiment. It can be done (unless cheating was involved, which we don't presume, because who would want to cheat in a single player game?). Does this single experiment show that all games can be won if you are playing sufficiently well? Of course not, if you claim that you are seriously misrepresenting the available data. It is similar to claiming that because there are people that have driven for 20 years without having an accident, everyone could drive for 20 years without having an accident, if they were only driving well enough. This completely disregards that there are events where you have an accident which is not your fault, but the accident happens nonetheless.

So instead of verifying the claim that all games can be won (if you are playing well enough), I will try to falsify that claim. How can this be done?

Everyone who thinks that he or she can win all games can engage in a virtual FTL game with me. I will play the computer side, and the claimer can be the player. I will only use legal moves, as will the player, and our turns will be 10 seconds of game time (but that time frame can be adjusted as needed). Some time values will be estimated instead of measured (i.e. how long does a Rock take in a certain ship to move from a certain spot in a certain room to another room? - I'm too lazy to do that), but for many things we have accurate values (loading and recharge time etc.), so I guess it will be close enough to show a good approximation of a real game. Every move and time estimation can be disputed with arguments and evidence, of course, any correction of wrong timing is appreciated.

Who wants to play?

I think the fact that no ship save one can kill the Boss out of hangar and that the RNG could be manipulated to give you nothing of interest, the proof is right there in front of our eyes.

So one player joins the crows saying that "If you think one can win all games", you are wrong. Still hoping for people to disagree, so I can test my ingenious virtual FTL simulator. By the way, a daring player can of course get advice from other people during his or her move.

By the way, the RNG is not being "manipulated" here, at least not in the bad sense of the word. If the game could be won regardless of what numbers the RNG calculates, the game would not be luck-based (and some people used to claim that the game is not luck-based in any way if you play well enough). If the game actually is luck-based (as I claim), then the RNG does matter. However, I will (try) not to let you miss every shot, I think (and can attempt to show) a pretty standard roll of dice can kill a player.

Also, you can play the Basilisk, if you think that's your only chance for a sure win against bad luck.

You may be hard-pressed to find someone who thinks you can win ALL games with perfect play. Upwards of 95%, though? Probably.

It is also my opinion that most games can be won (unless you are really, really, really unlucky), maybe more than 95%, maybe less, I think all these percentages are numbers pulled out of ... nowhere, unless one can show some meaningful statistics.

However, there certainly are people stating that all games can be won (if you're good enough), and I'm addressing this extreme and in my opinion outlandish claim here. I may dig out such claims of the depths of this forum or the whole Internet sometime.

The game is random. All of your shots could randomly miss and your enemy's shots could all randomly hit. Of course you can't win every single game in the universe of all games.

I don't think anybody has ever soberly argued that you can win every single game every single time with no exceptions whatsoever.

As to the actual argued fact that "you can win 95% or more of all games if you're good enough," that's true as well. Some may even be able to win 98%, or even 99%.

shadowcrust wrote:It is also my opinion that most games can be won (unless you are really, really, really unlucky), maybe more than 95%, maybe less, I think all these percentages are numbers pulled out of ... nowhere, unless one can show some meaningful statistics.

However, there certainly are people stating that all games can be won (if you're good enough), and I'm addressing this extreme and in my opinion outlandish claim here. I may dig out such claims of the depths of this forum or the whole Internet sometime.

I have recorded every game I've set out to record, so that can be taken as a good, fair sample. I've played 36 total runs (2 each with every ship, I think) and have lost 5 of those runs (never twice with the same ship). 2 of those losses were when I was newer at the game, and 3 of them were after a long break and I was obviously off my game. You could point out mistakes I made in all of those runs (and all the wins, too) and in the latter runs I actually point them out myself, sometimes mere moments after making them.

So with what I consider sub-par play on my own part, I won 86% of my runs. I don't think it's untoward to assume that if I'd played better, I'd have won 2-3 if not all 5 of those losses. Nothing in them says "Oh you got RNG screwed!"

In fact, I don't think any one encounter took me out. It was a slow beat down over multiple encounters that did it each time, exacerbated by my poor decision making skills.

Numbers! That's awesome!

If I look at my first profile, of my last 28 runs, where I was unlocking the last achievements and obtaining victories worth more than 5000 points for all ships, I won 21, which is a nice round 75%

In the new profile, where I tried not to lose, I won 10 out of 11 runs (each win scoring more than 5000 points, by the way ), which gives a better percentage, but from a smaller sample size: 90.91%

We record:

86.11% wins, sample size 36
75% wins, sample size 28
90.91% wins, sample size 11

overall win percentage of two relatively experienced players:

82.67% wins, sample size 75

all games on normal

If we get more numbers, it will make the conclusion better, but I think if we say we can expect to win 4 out of 5 games (regardless of ship) it would be relatively close to the mark. If we really try very hard not to lose, possibly sacrificing joyful game play, the percentage would be probably higher, maybe around the 9 out of 10 games I experienced with my (quite small) sample.

Now we've got numbers, more are welcome.

I should mention, I won all 4 of my runs with the stealth ships, which means yes, I won twice with the DA-SR 12.

I don't remember all the losses but the first 2 (in season 1) were with the two Engi ships, and 1 (in season 3) was with the Kestrel. I think I also lost with the Stormwalker (again season 3) and then one more ship. I think the B Zoltan ship. I should watch those losses again to see exactly what happened but I think I'm going to take a nap instead