The stats you see for the mulligan cards are dumb; they were never very useful; they are often a huge handicap and a distraction and they are better off turned off to avoid the distraction and keep your peace and quiet to think.
The main reason they are bad is that they look at the average game and not your own game; maybe they look at the opponent class and only show related stats to that; do they know their archetype though (no they don’t)?
Wait for it though: it’s even MUCH worse than that in reality; I doubt they check what else is in your mulligan currently since it’s only an average; do they do stats on the interactions inside the mulligan(no they don’t)?
They are BAD guides. Not only they refer to the average game which is already bad because even if they filter by different opponent classes: they definitely don’t filter by different opponent archetypes (because they don’t know the cards of the opponent no matter what (during the mulligan process)), worst of all: they do not produce stats based on the cards of the rest of the mulligan (the cards that are currently also in the mulligan) (we know because we know the formula those stats come from (they do not take into account the other cards currently on offer in the mulligan)).
The problem with seeing those stats is that 90% of the population can’t interpret data even if their lives depended on that.
And on top of that there is something called nuance that you have to be a real gambler to have on those scenarios.
Like…
How much you should Just accept to play a non optimal opening hand instead of Mulligan it?
That type of question requires a answer Far deep than Just Mulligan winrate and one that usually can’t be contabilized in a “good enough” way.
I agree with you here for the most part. I think the stats are mainly for newer players to the deck for a general idea. They don’t always align with what’s best to play.
A prime example is in DK Starship, it will tell you not to keep the 2 mana 3/2 Spellburst minion that triggers a DR. Sure, you want to use the minion when you can trigger the Spellburst, but you also need a 2 drop and this isn’t bad. If you have high cost cards in hand, you definitely want to keep this. If you already have a nice 3 mana curve, sure, ditch it.
But overall, it wants you to ditch it and it’s not really accurate.
I fall victim to this A LOT in BGs. BGs tells me what hero is better, but often times it’s only better IF you know how to use the hero really well and build a certain build, it might not really be better overall. Needless to say, I suck at BGs. I’m at 7k for a reason.
Yep and if anyone is interested in the mathematical reason this is true, check the formulas on how those things are calculated over at the help pages of hs guru. They work on the average wins or losses of the entire deck based on SINGULAR cards mulliganned (kept+random) or kept(manually) or drawn(in the entire match).
We can derive two major conclusions from the formulas: 1) they ignore completely interactions inside the mulligan itself 2) they ignore (as an extension to the 1st conclusion): ANY interaction that is unrelated to the blunt “did it win or lose?”,
and we didn’t even mention they don’t even track the type of opponent… (even if they try they will at best do Class-based and not Archetype-based (they will never know the archetype of the opponent (that data is barred))).
It’s a combination of best. It tells you % picked AND what the impact is overall. In BGs, it tells you what tribe is best that nets the most chance for a higher rank.
It’s not only what was played, but all the impact of said playing has.
So if you have a high cost minion in hand, chances are that keeping it in mulligan has a bad impact. It also tells you how many people have kept it in mulligan.
It also does this VS a class, so it changes all the time.
The problem is that classes have more than 1 deck, so it might be a good idea to keep it in mulligan vs deck 1 but not deck 2 and it only shows you the average. That’s why you can’t rely on it. You have to know.
Prime example is if you have Reska in hand at mulligan vs Warrior. It might say ditch it, but keeping it in hand against a Hydration Station Warrior is not a bad play. Keeping it in hand vs Brann Warrior likely is very bad.
The tracker in BG’s has no way to quantify a hero that a player is so good with it hardly matters what the lobby is. I have 2 or 3 heroes i can play in any lobby and get top 4 or 1st quite frequently. Unless you know how to play the build it is pointing to in the data below the hero it’s pointless IMO.
Yeah. this is why I have to ignore BGs a lot on hero selection and trinket selection.
I can’t play Elementals or Naga. I can never get the key pieces to cycle and do not know how to maximize them. So any hero that is rated high based on those is automatically a no go for me.
Don’t feel bad. I think Naga might be the one tribe most players either don’t know how to play well or don’t understand how they work. The pure build is such a no brainer and easy to counter. When you have a player that can play Nalaa correctly they can make a scary Menagerie build in a turn or two from basically nothing.
Elementals are quite weak ATM compared to other tribes as their scaling is stuck at last years BG scaling level for the most part. Unless you can get a double dip scaling working in your favor. I generally play a pseudo Elemental tribe build and pick up piece for menagerie along the way. When i get Rock Rock is when i dump everything except Wraith and cycle like made to scale my new Menagerie Board. I think i have stayed mostly Elemental in a whole game once or twice. You can make an amazing late game scam build from elementals as well with how easy it is to search for the pieces.
Personally agreed with just about everyone in here so far. Mulligan stats are great if you’re unfamiliar with the deck but ultimately the data can’t predict what you’re going to do after the match starts; it doesn’t matter how good of a keeper card X was, if you die on turn 6 because you colossally misplayed on turn 4.
Regarding BG’s, I ignore the hero stats and tribe place/win rates entirely, and often pay little attention to the trinket ratings too. That being said, the ability to see which cards are from what tiers really helps since I don’t have it all memorized (and don’t have to thanks to the tracker), and being able to preview golden versions helps me a lot.
I don’t know which tracker stats you’re using, but premium Firestone does take into account interactions between the cards
For example, if you have a card which is good in combination with another card, and both are in your mulligan, you’re told to keep both, but if you only have one of them, it will show you to drop both of them.
It also takes into account if you’re going first or second and adjusts the mulligan based on that.
Generally, it IS problematic if the class has multiple archetypes. In that case, your experience is needed, but apart from that and the periods right after the new patch, the stats are much more useful than anything else.
Don’t feel like reading the whole topic thoroughly, sorry if it’s actually a problem for anyone… But as for the subject, lemme do my usual stuff and reply, of course, with a quote from myself:
Do you have any evidence of this? I’m trying to find examples of what you mean and I only see that the Firestone tracker shows the kept percentage and kept impact; I wonder if internal card interactions is what you believe they do and you never checked if it’s true; what is certain is that at least D0nkey does not do internal interactions between cards on the formulas he has on his help pages.
In any case IT DOES NOT MATTER EVEN IF THEY TRY to do internal interactions because the trackers do not know the cards of the opponent; they exclusively know the class and nothing else on the mulligan stage; a card may have the opposite priority on a mulligan on different opponent archetypes.
And you can never claim “some classes have only 1 archetype”; what if the opponent uses a tech card or what if they do something else unexpected; and the more the opponent is smart the more likely to take advantage of that handicap of those stats apps.
The main purpose of those mulligan stats services is to sell a product; to hell with technical integrity; it’s profitable.
In the topicstarter’s case, it’s most likely a skill issue and a cognitive bias, the player perceiving him/herself much better than in reality.
Besides, I think I’ve already mentioned that those ‘stats’ from trackers aren’t trivial to interpret, for example, regarding
My suggestion for such players would be to just stop pretending to be smarter than they are and just ‘faceroll’ as they normally would (without a tracker, that is, which they can’t utilise properly anyway), according to their very own IQ… ‘The Algorithm’ will likely reward them with roughly 50% win rate soon enough by providing appropriate match-making, ‘rng’ outcomes etc. Trying to use ‘di stats’ would probably only confuse the system and interfere with the aforementioned ‘Algorithm’.
I asked the developer of Firestone btw. You’re imagining things. The stats are on singular cards against the total win rate (no relations between individual cards are taken into account).
I think those stats are more accurate the faster your deck.
If you’re playing something that wants to grab the agency and aggressively push damage, the stats are pretty straight up right.
This is a really good point. There are heroes that are objectively very good that I personally hate/don’t play well and they get skipped.
Having tidemistress on five makes staying on four and farming shakers/crooners pretty good. If you can pick up a trinket or two to support that it’s good enough to get to the end game.
I usually don’t want to do all that apm and pass, lol.
This is true; however, I take cyclone frequently as ten damage minions that early are good tempo while you level up looking for a better comp, ha.
Shocked you haven’t already received an explanation about why the dev is wrong about their product because our friend says so. lol.
and reply, of course, with a quote from myself: