Gauntlet Round Compilation
Peachtree Rex
✭✭✭✭✭
in The Bridge
Once again, the gauntlet has started to inspire a new set of frustration. I would like to actually investigate if the gauntlet rolls are truly fair.
The first, and most time consuming, step of this process is data collection. I would like a minimum of 1000 (and, ideally, more like 10,000) individual rounds to compare. Fortunately, a relatively new tool has made this collection significantly easier:
https://iampicard.github.io/
The gauntlet section for this tool now has an "Export" functionality that will write the results of all of your rounds to a CSV file. This includes data like which characters were used, what the expected roll range and crits were, the streak prior to the round happening, etc.
I have started uploading my results here:
https://docs.google.com/spreadsheets/d/1sYlf7A6M-kaENSsxoHhiUrpNuv71zNRz50Izi-ARB6E/edit#gid=0
I plan on sharing ALL results I get from everyone here. I won't allow people to upload directly because I don't want them to mess up the data, but you can always make a copy of it and work with it yourself (similar to my Lvl100 spreadsheet).
If you want to contribute your rounds to the compilation, please send them to sttgauntlet@gmail.com (NOTE: I will not open anything that isn't a *.CSV file. It'll just get deleted)
I don't plan on doing any analysis on this data until it has grown significantly, but, I am hoping, that this data can do one of the following:
I don't expect I will be the only one analyzing this information. In fact, I hope there are others who are interested in it as well. The data is pretty exhaustive and should cover most, if not all, of the various gauntlet theories out there.
The first, and most time consuming, step of this process is data collection. I would like a minimum of 1000 (and, ideally, more like 10,000) individual rounds to compare. Fortunately, a relatively new tool has made this collection significantly easier:
https://iampicard.github.io/
The gauntlet section for this tool now has an "Export" functionality that will write the results of all of your rounds to a CSV file. This includes data like which characters were used, what the expected roll range and crits were, the streak prior to the round happening, etc.
I have started uploading my results here:
https://docs.google.com/spreadsheets/d/1sYlf7A6M-kaENSsxoHhiUrpNuv71zNRz50Izi-ARB6E/edit#gid=0
I plan on sharing ALL results I get from everyone here. I won't allow people to upload directly because I don't want them to mess up the data, but you can always make a copy of it and work with it yourself (similar to my Lvl100 spreadsheet).
If you want to contribute your rounds to the compilation, please send them to sttgauntlet@gmail.com (NOTE: I will not open anything that isn't a *.CSV file. It'll just get deleted)
I don't plan on doing any analysis on this data until it has grown significantly, but, I am hoping, that this data can do one of the following:
- If the gauntlet is fair, add confidence that it actually is in all cases
- If the gauntlet is not fair, hone in on exactly how it is unfair (like was done with the shuttle failure testing)
I don't expect I will be the only one analyzing this information. In fact, I hope there are others who are interested in it as well. The data is pretty exhaustive and should cover most, if not all, of the various gauntlet theories out there.
3
Comments
Create a request on the iampicard tool. They could also keep track of the presented matches in some fashion. All you need to do is ask and have good rules for what you're looking for.
I, also, believe that the gauntlet is a fair system. I don't have any proof of that. While this certainly won't rise to the point of proof, it can at least provide confidence.
Yes, bad beats happen, and they are supposed to happen. Sometimes. The question is if there is a bias in the system.
Of course it's not fair, it's a Gauntlet, its supposed to be painful.
Totally. Besides some players are notorious for ensuring victory by using dil to elevate themselves literally 10-15,000 points above the “lesser players”
Could you please continue the petty bickering? I find it most intriguing.
~ Data, ST:TNG "Haven"
They don't need to be equal in quantities, just sufficient in numbers.
I am a lot more interested in if crit percentages are accurate and displayed ranges are evenly distributed. For those tests, in particular, it doesn't matter what the expected win rate is only that the rolls are following expected behavior.
If the rolls are following expected crit and range behavior, then the system is be necessarily fair and operating as expected.
They fixed the AND shuttle error in the face of incontrovertible statistics. Even if they haven't publicly admitted it, the community work finally exposed it.
The problem with the gauntlet is that there hasn't been a transparent set of data to show any specific nefarious behavior.
Could you please continue the petty bickering? I find it most intriguing.
~ Data, ST:TNG "Haven"
My gut tells me the defender will have a higher overall % of maximum on their rolls, on both crits and non-crits than the attacker. What makes you so dead certain it isn't biased? Are you sitting on a pile of data that disproves my hypothesis?
"5% player" = 43 crits / 834 rolls = 5.10%
"5% opponet" = 50/1227 = 4.07%
"25% player" = 451/1788 = 25.22%
"25% opponet" = 297/1236 = 24.03%
"45% player" = 344/555 = 43.96%
"45% opponet" = 293/642 = 45.64%
"65% player" = 113/171 = 66.08%
"65% opponet" = 62/102 = 60.78%
I've been tracking for a few weeks since the last time gauntlet fairness was debated. Still need more 65% tier datapoints but the trend is clear. Crit percentages at the very least are fair. Not speaking for regular rolls or opponet selection or if specific situations will lead to certain results or if the freemasons and the trilateral comission have their hands on the scale.
initially, I was disappointed I missed the Guinan fun of a few weeks ago to get more 65% sampling. But now I have the 65% Kahless gauntlet, so yay?
Busted, lol.
Could you please continue the petty bickering? I find it most intriguing.
~ Data, ST:TNG "Haven"
I had one user who submitted 8000 rounds. After some conversation with them, neither of us were 100% sure how the iampicard tool managed to spit out that many. So I have decided to isolate those rounds to the "Gauntlet Dump" sheet and exclude them from the results.
To celebrate hitting 500 rounds, I have created two analysis sheets. The "General Analysis" looks at observed Crit% and Percentile range of individual rolls. Below that is a section comparing iampicard's Expected Win% measured against Observed Win%.
The "Reward Round Analysis" sheet executes the same exact test as the "General Analysis" sheet, but it only looks at "Reward Rounds", that is, rounds where the player will receive an extra reward if they win.
I'll stress that all results are still very preliminary. I'm going to wait until we're at 1000 rounds to actually crow about anything.
https://docs.google.com/spreadsheets/d/1sYlf7A6M-kaENSsxoHhiUrpNuv71zNRz50Izi-ARB6E/edit#gid=1506601349
Edit: if there are other analyses/breakdowns you'd like to see, let me know and I can do my best to accommodate.
IT'S HOSTED BY "Q"! Isn't Q the most annoying being in the universe? Do you find Q a resonable, logic character? Well, good luck beating the odds vs Q's game!
Yep, soon as i adopted this 3 caretakers dropped over one weekend.
Before that one guinan no locutus.
i adopted this 'i don't care non-strategy' as a means to get 20k rounds, they count win or lose, for my first locutus
Edit: after 16k rounds i finally rank #1 for the achievement
The crits sometimes are even lower than non crits which is mindboggling why it is even a crits in the first place
Not quite. It literally doubles the roll result. This is readily provable because there are no odd crit rolls.