DH deck variance promo thread. My 5 rung story

Multiplayer Discussion
1

This thread is in support of custom variations over netdecks. Promoting it hopefully among those who have not because of winrate concern.

I seldom run netdecks. Losing more often due to my clever failures is not particularly exciting, per se, but it is entertaining and feels more rewarding than using someone else’s concoction. Nonetheless, I want a winrate that does not make me feel as though I participate to merely to be a step on the ladder.

Here is the impetus for the content of this thread.

Thank you @Bling for this. For some reason, reading that motivated me to adjust my lax deck building.

So, what is needed is a way of evaluating a deck to determine whether it is theoretically “faster.” This was more challenging than it first appeared. To begin with, the netdecks do portend to show “tempo” to be the most successful, so cards shared among slightly varied top netdecks needed special favoring in regard to possible cuts. How “faster” could be evaluated and achieved was also suggested by netdecks. Included were: one charge minion (always), one rush minion (not always), several draw cards [2xSkull of Gul’dan] (always). Although, discovering an effective variation that did more than swap one or two cards from the standard was clearly needed, the lack of variance in the top netdecks did nothing to show the way. I would need to guess the most favorable “faster” factors.

None of the netdecks viewed that day had a taunt minion. This is an obvious expectation for an aggressive deck, and the reason is that a taunt is perceived as slow. I expounded that. There are varying degrees of slowness for cards because of a player’s will to gain maximum value from play. There is definitely a psychological factor to it. Therefore, my analysis ranked cards by their inate or perceived “effective usage.” I called it a conditional factor. Cards were scored as conditionally soft, medium, or hard. The implication is that something causes a player to hold these cards, thereby slowing deck playability.

  • Soft - mostly perceived and strategic; a board state so desired that it feels unplayable prior.
  • Medium - mostly assessed and strategic; a board state that enables the full value to actualize, like with combo cards.
  • Hard - mostly mechanics; a board state that enables play. (ex. affects minion)

My guiding premise was that these conditionals were as important of a quantity consideration as cards per cost and rush|charge|deal damage abilities. Here is a table that showed the data gleaned from several top netdecks for tempo DH, and the one that I made in an attempt at improving these fastness factors.

{EDIT: Updated two missed deck changes that occurred prior to playing it. Revised conclusion.}

#@0#@1#@2#@3RUSH+DRAW+WEAPONSOFTMEDIUMHARD
netdecks27->84->6616->749->109->112
my deck #138554547140
my deck #22782454812

The netdecks data shows the range between several. The two hard conditionals in all referenced netdecks was [Eye Beam]. Weapons were regarded as medium conditional play cards, since each must wait its turn behind a currently equipped weapon, and strategically might lose value if used only for face damage. RUSH+ includes charge minion [Kayn Sunfury]. The custom added rush minions were [Furious Felfin] and [Faceless Corruptor]. DRAW+ would include “summon from deck” effects, but none were included.

This effort resulted in two custom decks. Only one did the work to move me from gold 10 → 5. The other lost twice. I ended the day with deck win rates at 75% and 50%. An improvement over custom decks constructed previous to this analysis.

CONCLUSION

The sample size was too small for sure, so this is tentative. The deck that did best had the most medium conditional cards, and least soft conditional cards. A counterintuitive result. Higher conditionals should cause more card play delay, but apparently the value generated by those cards was significant enough to overcome its greater slowing factor. Maybe the soft conditionals are more damaging to deck speed. One example would be holding onto [Twin Slice] for [Altruis the Outcast]. Not shown in the data is that the second deck included one taunt minion [Bonechewer Brawler]. That deck was made to give a greater variety of options each turn. The data shows it with the least number of conditionals, and one more low cost card, although weighted more toward cost two. I do not know whether the second deck was slower or less powerful. Again, too few games to tell. Both decks only used five draw cards, 1->2 less than netdecks. I should suspect that less draw would be less fast, but synergy and usefulness during actual play felt like the biggest factor contributing to the winning round.

These decks are being tracked by hearthstone deck tracker. I am not posting them here because, again, this thread is to promote more deck variance in ranked.

GLHF!

1 Like
2

I would like to introduce the concept of value per tempo, which optimal value
is dictated by the consistency of the archetype in question.

Example: Let’s say a Tempo DH’s VPT is 3/7 then comparatively a Rez Priest’s would be 8/2. A Rez Priest can’t have lower tempo (only 7-10 mana cards) or he would die before turn 6; he can’t have a higher tempo or he would lose in consistency. (Mass Rez on 1/2 minions is a lack of consitency).

The second concept I would like to introduce is tempo fluency.
Tempo describes the momentum in one turn, fluency in more than one turn,
and takes swings into account.

Now, by considering the paragraphes above, it seems obvious that what is consired to be “fast” may mean different things in different contexts.
Being fast may mean killing your opponent by turn 5 and not by turn 6, yet this has nothing to do with fluency, because of swings, and those swings are only possible if an archetype has enough value in hand.

The thing is, fluency isn’t necessarily undermined by deck playability in a short term or 1 turn (waiting for a big Altruis turn).
Another problem is that I have trouble classifying certain cards by your method, cuz you didn’t give enough examples, imo.

Look at the card Burn Mana,
the conditon for use is not your own but depends of your opponent, for you may want to wait until your opponent is about to play his win conditions. That’s the problem all cards of highly strategical value (all priest removals) have in common, thus although Mana Burn appears to fall under soft conditionals, in reality it has hard conditional factors, independent of you.

Now I would change weapons as soft, since they enable to play other combos, thus in fact increase the effect of playability of another card. This increased effect of playability should count to the enabler.
Example: Turn 2 Umberwing, turn 3 Stayr Overseer. (Without Umberwing, Satyr would only be played the next turn). Side note: 2 Warglaives can be used in the same turn.

1 Like
3

Wow! Thank you for your thoughts regarding the wall of text, and for giving fair consideration of the presented ideas.

I like how you broke down tempo, but we might have different views about how the term applies. Let me explain. I believe that Hearthstone is designed to emphasize only three general strength curves.

  • Face [proactive] - All tools included are to support maximum face time or damage growth per turn.
  • Fend [reactive] - All tools are played predictively to create a board advantage in a majority of turns, with kill damage resulting consequently.
  • Fatigue [exhaustive] - All tools work to reduce the wherewithal of the opponents deck.

Now, these general designs guide strategic deck builds that also meander into hybrids straddling two or all three. These emergent strategies may be classified as above or by the hybrids below.

  • Tempo [preemptive] - Combines 1 & 2 to generate favorable board state and swing potential when necessary.
  • Control [preparative] - Combines 2 & 3 to glance or absorb the offense by deploying survivability tools in an effort to outlast the opponent by exhausting their resources.

When you say “tempo fluency” I associate that with my understanding of the strategy premise, being that board advantage can be shifted. It is my purview that the purpose of tempo is to exploit game options that prevent or reacquire advantage. Tempo decks probably allow the broadest use of tools to adjust to any meta.

When you say “value per tempo” it does confuse me. That is a rate or ratio. How would value be assessed per card or play, and how would tempo be scored? It reads the same as if you said value per [insert any strategy].

As for the weapon, we agree. I did favor including weapons as a soft conditional for play, but seven mana does not allow a cost 5 and cost 2 weapon to be played without additional allowance by the board state. Only one weapon may be played to its utmost value in most cases. However, I would agree with anyone who claimed that one weapon ought to be excluded from being counted as having a conditional play hindrance, since the first weapon used may always be played on curve without losing value due to other weapons in hand.

I should have stated that some of these conditional assignments (soft or medium) will be different between people due to perspective. For instance, I had to edit it a second time to add that [Faceless Corruptor] is also a rush minion, and that its condition is medium, not hard. It can be played without a minion to transmute.

I am not asking you to elaborate about value per tempo, but the read would be gladly studied if you did. I do not want to ask you to devote more time, though, since you have done a lot already.

By the way, as expected, the 75% winrate is dropping fast. Gold 5 -> 1 is filled with priests, mages, and shamans. It does better against the other six classes, but they are maybe one-third or less of matches.

1 Like
4

For me tempo means having at one particular moment more stats on the board than your opponent (ofc that’s not true for example vs otk, DH, but let’s simplify for the next part).
Arguably I use VPT in 2 different, distinct instances. I use it for particular cards and then for archetypes. VPT is used more as a tool for classification, it’s not really used as a ratio (although one could, but the result is difficult to interpret)

If I play on turn 2 a card with 2/2 stats then it has a tempo of 5, cuz it is average for its mana cost. It would have a value of 2, because value is always bound to card advantage.

Dragonqueen Alex has 10 tempo and 10 value (at least on turn 9).

A card that destroys your whole hand, costs 1 and summons 3 Millhouse Manastorms has 10 tempo, but 1 value (at least on turn 1).

Plague of death, which silence and destroys the whole board, has 1 tempo but 10 value. It resets the board bu tdoesn’t provide additional tempo. It does though destroy many played cards, thus has a high value.

For archetypes you have to look at the first 10 turns. DH will very often perfectly curve thus let’s say use 70% of his mana crystals perfectly. Yet at
turn 10 he might no longer have any breath (no more cards), thus horrible value, if you average turn 1 to 10, perhaps 4 (arguably I summoned this numbe r out of my hat, it’s relative to other archetypes). Thus his VPT would be 4/7.

A Rez Priest on the other hand might not play any card at the beginning at all, but his archetypes cards have more value (Mass Rez) than let’s say any cards of a DH.
Thus from turn 8 on, the Priest will outvalue the DH.

I believe that all the non essential cards of an archetype must together form a theoretically best VPT. If the core cards of archetype would have a VPT of 7/4, then the core cards together with the non essential cards must have a VPT of 7/4 too, which would be perfect for that particular archetype.

One can argue that having low HP makes a difference for evaluating tempo etc., but a) I simplify, b) you have either enough hp or not; it is a yes-no question, which is always difficult to evaluate.

1 Like
5

Thank you! That helped me understand. This system you use still relies upon the devs imparting fair balance. For example, a cost 1 that does one million damage to face (your choice) would be limited to a value of 1 with a tempo of maximum - 10. It is a ridiculous example, but it does support the point that effects do not compare well without accounting for HP. You must have scored every card to put this into practice. You must have been playing CCGs for a long time.

I hope more chime in. This type of discussion is rare on this forum. Oddly.

1 Like
6

I think most people associate this thread with work, not fun^^

1 Like
7

lul
… 20 chars