Which Gauntlet Would You Find More Interesting?
5000 Quatloos
✭✭✭✭✭
The current frequency of featured skill in the gauntlet appears to be
1.5 1.6 occurrences of the featured vs 1.0 for any other skill. Increasing the featured frequency further than it already was would yield more diverse sets of Gauntlet candidates. The table below shows experiments for which characters the current Gauntlet I'm in (DIP + Hunter, Astrophysicist, Bajoran) would have most "utility" : sum( (probability of skill pair) * (value of skill pair) * (crit bonus) ) in the Gauntlet. Which mix of favored opponents would you find most interesting?
[Edit - table slightly adjusted by adding a bunch more of the second-string candidates into the calculation including MIrror Inquisitor Troi, Mambo Picard, Damar and Leland]
[Edit 2- table adjusted again with better estimate of featured skill frequency and fixing a couple of trait bugs for some crew]
1.5 1.6 occurrences of the featured vs 1.0 for any other skill. Increasing the featured frequency further than it already was would yield more diverse sets of Gauntlet candidates. The table below shows experiments for which characters the current Gauntlet I'm in (DIP + Hunter, Astrophysicist, Bajoran) would have most "utility" : sum( (probability of skill pair) * (value of skill pair) * (crit bonus) ) in the Gauntlet. Which mix of favored opponents would you find most interesting?
[Edit - table slightly adjusted by adding a bunch more of the second-string candidates into the calculation including MIrror Inquisitor Troi, Mambo Picard, Damar and Leland]
[Edit 2- table adjusted again with better estimate of featured skill frequency and fixing a couple of trait bugs for some crew]
Accepted. Mark them, Galt.
Which Gauntlet Would You Find More Interesting? 50 votes
1:1 - I like my walls.
0%
Current 1.5:1 - Mixing it up a bit is interesting but I don't want to stress too much.
8%
4 votes
1.75:1 - Give the dark horses a fighting chance
16%
8 votes
2:1 - The heavies should shine only when appropriate
26%
![[QH] Oxmyx](https://us.v-cdn.net/6030345/uploads/userpics/547/n2W6EFWXI2VWP.JPG)
![[ISA] Big McLargeHuge](https://us.v-cdn.net/6030345/uploads/userpics/898/nXKVS62J0RWS7.png)
13 votes
3:1 - Super Rares are people too
32%
16 votes
I'd rather have a root canal than play another Gauntlet round
18%
![(CSF) Jesus Is Lord [••••]](https://us.v-cdn.net/6030345/uploads/userpics/839/nQHMW3V3VSZSB.jpg)
![[AR-UNI] General Abre](https://us.v-cdn.net/6030345/uploads/userpics/168/nZ7SNTTE4KXLY.jpg)
9 votes
4
Comments
you not right
DIP appearances: 67 91
CMD : 38 54
SEC: 39 58
ENG: 39 55
MED: 43 51
SCI: 45 53
The average of the last 5 non-featured is 40.8 54.2, so a ratio of 1.64:1 1.679:1.
That would mean, I believe that any of the 5 skill pairs that include DIP occur 9.1% of the time, while any of the 10 skill pairs that don't occur 5.4% of the time. Equivalently, the featured skill will show up in a sub-round 45.5% of the time.
Notable Super-Rares:
Lieutenant Valeris (451)
Admiral Owen Paris (411)
Formal Dress Wesley Crusher (398)
Commander Donatra (380)
B-4 (376)
Disguised Tuvok (357)
Waitress Ezri (357)
Blood Oath Jadzia (356)
Dr Gillian Taylor (354)
Romulan Kirk (353)
Note - the first few posts above had some gaps as I hadn't yet added all possible competitors. I think this image should now have all of the top 25 contenders for this gauntlet.
Notable Super-Rares:
Commander Donatra (473)
Alidar Jarok (447)
Marshal of France Q (391)
Admiral Picard (380)
Admiral Owen Paris (368)
Bajoran Dukat (366)
Lieutenant Valeris (364)
Defiant Commander Worf (359)
Increasing the frequency of the featured skill would undermine the need to cover all 6 skills. If that's a strategy anyone prefers to use (and I've seen a few of you do it!), then fine. I'll happily use my 3x fatigued crew to keep a streak going against a "no matching skills" opponent.
To me, un-doing some of the walls involves a more creative selection of traits. Caretaker is great, but less-so when he's in a gauntlet with a bunch of 25-45% ENG opponents. He might still be the better choice, or he might not. I might pick Assimilated La Forge as my ENG representative if he's 45% and take my chances with RNG against a Caretaker. Same argument for any of the other gauntlet wall superstars. So find the non-intersection of traits between the 600-800 max rollers and the 400-600 max rollers and force us to choose 5% w/ best rolls or 25/45/65% with weaker rolls. Some gauntlets do offer these kinds of choices, others don't.
Continuing the experiment of how much the featured skill might matter, here are the possible MVPs for the Gauntlet begun last night featuring MED + Gambler, Hologram, Civilian, accounting for featured skill occurrence:
Notable Super-Rares:
Mirror Phlox (403)
Nurse Faith Garland (403)
Dr. Toby Russell (381)
Doc Crusher (374)
Anij (371)
Edith Keeler (360)
Dr. Gillian Taylor (356)
The President of Earth (352)
That would make it all but impossible to compete in. The players with an advanced star base and all the bonus collections would win every gauntlet (instead of just 90% of them).
Let me start by saying i have an advanced crew, starbase, and many collection bonuses. Does not give me a 90% advantage. I loose aprox 50% of my rounds which is why i simply asked for proper numbers.
I realize that new players do not even compete in the gauntlet for lack of crew and bonuses.
Maybe a nice compromise would be to create separate sections for different lvl players similar to arena. And then make the numbers actually work.
As i understand the need for lower lvl players to be able to play, i also dont think that veteran players should be punished for all the hard work they have put into earning the % bonuses.
So, there's actually an easy way DB could split the difference and meet in the middle - adding more "dice rolls" to each match-up. This would still leave some uncertainty / element of chance, but yield more cases where Jedzrilanna would actually get to "see" the expected effects of skill bonuses or better proficiency characters. More dice rolls would reduce the variance in the distribution of scores, which would make better skilled crew advantages more noticeable in intermittent play.
What's going on is that most strong gauntlet characters' proficiencies have very broad ranges. Consider Defensive Phlox- his max MED proficiency of 913 is breathtaking, but his min MED is 261. So, put two Defensive Phlox's head to head, one with a 10% skill bonus and one with a 15%, and what you have is two very very broad bell curves, slightly offset, but massively overlapping. Better-bonused Phlox does indeed win an extra 5% of the time, the problem is that our brains aren't wired to accumulate and notice that evidence over the weeks and weeks of Gauntlet play it would take to show up statistically. Which leads to a perception of basic randomness, which is unnecessarily unsatisfying and seems to not reward pursuing or bonusing good crew.
More rolls of the dice going into the match-up total would tighten up the width of the two bell curves, so they would overlap less. (It would also give more chances for the crit bonus to kick in and be noticed and appreciated). It wouldn't guarantee success, but it would lead to a much more satisfying experience by having the wins be more noticeable in the here and now. The more rolls added, the closer it would get to Jedzrilanna's "deterministic win", and the more impact the skill bonuses would have on performance. Out better-bonused Phlox's "loaded dice" have a bigger influence on success rate the more times the dice are rolled.
PS - I need someone to think through and validate my math and logic here. I'm pretty sure it's right but peer review would be welcome!
OK, here are some examples / experiments on how many times "rolling the dice" in a match-up matters. For convenience I didn't include 5% crit chance in these experiments.
Experiment 1: 0% Bonus Defensive Phlox vs +15% Bonus Defensive Phlox
Taking his MED skill only, and running 20,000 match-ups, here is the distribution of final scores. Horizontal axis score, vertical axis is # of occurrences (i.e., these are histograms). Unbonused Phlox is in orange, Bonused Phlox in blue. (I realized after I ran these that the maximum proficiency bonus is 13% not 15%, but this will suffice). First, using the standard 3 rolls:
Bonused Phlox actually wins the match 70% of the time but by golly those distributions overlap alot. Anyone other than a Soong Android would be forgiven for seeing this as essentially 50/50 when strung out over many many weeks.
Now, let's say we had a Gauntlet which rolled the dice 12 times, not 3:
Bonused Phlox wins the match here 85% of the time. YMMV, but honestly to me that looks more like what I would have "expected" for all of the blood sweat and tears put into bumping up proficiency bonuses. More importantly, I think it would actually be a discernible, not just statistical advantage.
The problem is this same tweak would make it even harder for underdogs to break through the big walls.
Experiment 2: Defensive Phlox vs Reverend Phlox
Here are 20,000 match-ups using the standard 3 rolls (Reverend Phlox is blue, Defensive Phlox is orange):
That seems pretty reasonable. Defensive Phlox wins 90% of the time, but the clearly weaker Reverend has some chance. Though I can see why a 10% chance is a little sporty given the skill gap.
Now, in a gauntlet with 12 rolls:
Defensive Phlox wins 99.5% of the time, which seems a little deterministic for me and pretty discouraging to entry-level/medium-level players (hence the good idea above about tiering the Gauntlet...).
Maybe something in the middle. 6 rolls seems like it would take a good bit of the "randomness" out but still allow underdogs a fighting chance...
I do not disagree
FWIW, 25-60 is kind of all you need to cash in on a reasonable number of vaults/cases/whatever and merit farm every day. Most of my gauntleteers are in the 2nd-4th rows of the images I posted above, and that's about where I end up, maybe higher if I don't miss gauntlet cycles and spend much more time thinking about individual matchups. When I do place higher I don't notice any appreciable difference in the reward outcomes.
I 100% agree that Gauntlet tiering would be a great idea. (As a side benefit, could keep the trainers as rewards in the lower tiers ... I don't need them anymore but there was a time when they were definitely my limiting resource).
Per the original post I'd like to see the featured skill importance enhanced a little bit, to have more Super Rares show up higher in the list and significantly increase the variability in the lineups, giving more people a chance and limiting the blunt-force value of the monsters to gauntlets in which you would expect them to shine. And from the simulation above I'd like to see the number of rolls increased slightly, to reduce the "completely random results" perception and properly reward skill bonus collection and crew development.
Keeping the experiment going, here are MVPs for the MED + Civilian/Desperate/Tactician Gauntlet started last night:
I submit that winning (or even a top-10 finish) is not necessary for “success” in the gauntlet. There are maybe only two things you can get from the gauntlet that matter: merits and the legendary crew.
Merits come from putting together wins regardless of the number of trophies and you don’t need a top-5 gauntlet crew to get wins. Even with a top-5 crew, trying to go for regular streaks of more than 3 wins is dependent on extreme luck or spending dilithium. And if you’re faced with a wall of unbeatable opponents, redshirt someone rather than refreshing. Merits have more value for thawing frozen crew, buying Faction-only items, and/or buying merit crew packs than they do for refreshing gauntlet opponent lists.
For the legendary crew, it is true that the Gauntlet Vaults have a higher drop rate than other reward boxes. However, that rate is still abysmally low. I try to maximize the number of reward boxes I get through short streaks rather than trying to win every gauntlet. Another way to maximize the number of boxes is to aim for ranks that grant seven boxes without worrying as much about whether they are Vaults or something else. If you can’t hit 1st place, aim for 11-15, 26-30, or 41-50, as each of these will drop seven chances at a legendary rather than 1, 3, or 5 that come from other ranks.
Interesting...I went in a slightly different direction: in addition to Guinan and The Caretaker, I went with Antaak, Enabran Tain, and Gangster Spock over several others with higher calculated utilities. I’m actually all the way up in third place right now, largely due to a solid set of crits from Enabran Tain during this morning’s battles.
Interesting ... that's consistent with my experimentation with these over the last few days : not having specific heavy hitters in the first several rows doesn't seem to have much of an impact on final rank (which I agree really doesn't matter much past a certain point). I think anyone in the top 25 above would be successfully playable, and probably (now that you've shown with Antaak and Enabrian) many in the next set of 25 down. I think the actual individual matchup selection and crew-exhaustion strategy probably dominates over the purely statistical optimization above.
A rule of thumb I'm coming up with during the experiments is if I choose a weaker skill but skill-featured and well-crit'ed crew to field, then only use them when both of their skills are available, since that's what I'm counting on (higher likelihood of them appearing and having the good crits rolled to boost their scores); using them single skill would be counter to the strategy. It seems to be working well...
Once in these experiments I had two crew in the overall priority list within a couple of points of each other and I loaded the higher one, leaving a 2-skill hole in my lineup. The slightly lower crew (Caretaker) had both of the needed skills. I've played 1-skill holes before but never 2-skill, so I tried it, and it hurt. I should have sacrificed a smidge of theoretical advantage for having the Banjo available to use when I needed him, rather than having to throw my better crew against the wall.
This one's a decent one for the Super-Rares.
Observations thus far:
Using this rubric, the variety of strong gauntlet crew from day to day is a lot deeper than I would have guessed just looking at the lineup of opponents that show up. The ones that show up are driven by (a) likelihood that players have that crew, and then (b) likelihood that they field that crew given that they have it. The availability is high for gauntlet crew that were mega threshold rewards, and then I think there's a tendency to overplay brute-force power crew, if for no other reason than gauntlet isn't worth too much daily time and energy thinking about :-)