Better Life for Supports?
When the patch notes for 7.00 were released, many players thought life for supports was going to get significantly easier. Support heroes received interesting, new talents, some providing either a choice between extra Gold or XP.
Moreover, lower level requirements, as well as Tome of Knowledge changes have made getting to these crucial levels a much simpler task. In turn, it should have created a more balanced support/core environment—with cores still being item dependant and no significant changes to the gold gains for the majority of core heroes, this meant that the period of relevance, or even dominance, for supports has increased.
The reality, however, was not as definitive. Supports did not suddenly become more popular among the players and neither do they dominate the pubs. It certainly doesn’t mean that the role is less important, compared to cores, but the ability to reliably raise MMR with support heroes is still well within the domain of the selected few exceptional players.
That said, the economy of support players has definitely improved. Compared to 6.88, the spread of the average GPM has tightened across heroes—on average, the highest earning heroes now earn 3% less Gold, while the poorest earn almost 4% more. The same, to an even higher extent, can be said about the XPM.
The new flat GPM increase and XP gain talents are definitely helping with the support item and level progression, and in this case the stats definitely do not lie. But do these talents actually help the heroes win more games?
GPM and Respawn Time Reduction (RTR)
Looking at the table, it becomes almost obvious that in most cases the GPM talent loses to its counterpart. In fact, in almost 65% of all cases, taking GPM over the other benefit is a game-losing choice.
Furthermore, in the majority of cases where taking GPM is a better alternative, it might not be the impact of the GPM increase itself, but rather the extremely low or even negative utility of the alternative, with the biggest culprit being the RTR, which consistently results in a higher chance of a loss.
On paper, the RTR sounds pretty good—spending less time dead means the hero has more time providing utility for the team or farming. In practice, especially in pubs, it frequently means that the hero simply has more chances to feed extra gold to the enemy team.
In a typical game of Dota, the life expectancy of a support is rather low, especially if the hero respawns much earlier than the rest of his team after an unsuccessful fight and has no backup. Certain heroes with RTR might be able to delay the high ground push on their own, but once again, it puts their life at risk and you generally should not expect too much of a single support or utility hero in the later stages of the game.
The RTR talents might shine in the competitive scene, however—higher degree of coordination and lower propensity to feed should be extremely helpful in this regard.
Returning to GPM, the only cases when these talents are a game winning choice, which is not supported by the negative RTR utility, is when they come early enough to matter. Getting a smaller amount of extra gold early in the game is always better than a larger increase in the later stages, because the talent is actually doing something for your hero for longer.
However, the impact of these GPM talents is still rather small, with the highest comparative value of +1.2% for Ogre Magi, with the alternative being +100 Cast Range.
Overall, the data for the GPM talents was certainly surprising—most players expected them to be significantly more impactful. However, as it stands, taking extra health, cooldown reduction, extra XP gain and even raw damage seems to be working better for the majority of players on the majority of heroes.
XPM
Taking the extra experience gain talent means sacrificing an immediate gain in favor of a long-term benefit. Turns out—Dota really doesn’t work like that, with these talents showing even worse performance than the GPM ones. Only in 5 out of 24 cases is picking +XP gain justified and two of these cases are when the talent is compared to the RTR.
The game has definitely sped up since the introduction of the patch. Snowballing has also become more of a problem, and trying to counter it with a late game investment is not an option, at least for pub players.
In professional games, the “comeback Midas” was a concept that occasionally worked, and when you are losing, the +XP talents feel exactly like the “comeback Midas” pick-ups. Only in the competitive scene, the choice is generally made by a much more capable player and is often backed by a team composition, which can hold the high ground for the item to at least pay for itself.
When you are winning, it is generally a “win more” option—you are not going to win more games by making this choice, but you are going to win them harder. In Dota, it doesn’t matter how you win—the only thing that matters is the destruction of the enemy Ancient. When trying to play reliably, it is almost always better to get the immediate advantage, however small, over the illusory late game gains.
The lesser of two evils
Visage and Omniknight are unique, in a sense that they have mutually exclusive GPM and XPM talents. Interestingly, for Visage, taking the XPM is a better deal, whereas Omniknight benefits more from extra gold.
Seeing how the Visage gameplay has been developing this last week-and-a-half, it is understandable—more and more frequently the hero has been picked as a core, rather than a support, which allows him to get to his Aghanim’s Scepter in a timely manner, regardless of the talent selection.
Omniknight, on the other hand, really doesn’t need a lot of extra levels past level 10, since it allows the hero to have both of his defensive spells maxed out. The extra duration on the Guardian Angel does come into play, and levels are still important for the hero, but definitely not as important as getting a timely Mekansm, potential manapool increasing items and Aether Lens.
Closing Thoughts
Talent trees are a major addition to Dota. They improve gameplay, making it more complex, and level the playing field in the later stages of the game, by offering unique upgrades. While, naturally, there is a certain amount of imbalance still present, it is still an overall positive change.
Understanding which talents are a better fit for the current situation will probably take the same amount of practice, as was once required for learning all the items and their appropriate use. The aggregate data should certainly help facilitate the process of learning and highlight the definitive outliers, however, as with everything aggregate, some degree of caution is required with the evaluation.
The information discussed above was based on data from the first week of the patch—a relatively small sample size, especially for game like Dota. While the trends are likely to continue, some changes to the balance, even the ones not directly related to the talents in question, might have a significant impact on their utility.
first!
Missed!! 2nd still though!!
3rd wew
This chat for discuss, why you so stupid /_\
Yeah, supports can catch up a bit on levels and gold now this patch because of the bounty runes and talents. Still feels uneasy with the map and wards though.
4th!!!!!
WOW WOW WOW
7th hahahahahahahhaha
eighth
I'd like to report those fucking first place claimers and other shits that are even speaking for spamming.
Lol because of this supports are trying to carry
i love reading these posts but then you scroll down and you'd think you're in one of pewdiepie's old videos comment section
9!!!! i played dota 1!
interesting.. i believe that IN ANY CASE, exp and gold augment would be better, how wrong i was hahaha now i know why u lost games in 7.00
11th? WTF!
Still map is so huge,as support I have twice more job to cover my team with wards :(
У кого не заходит в стим +
I think causation and correlation got mixed up in this article. Most likely, players especially pick the GPM and XPM talents when they are behind and are having a bad game (which causes them to lose). The analysis should be repeated while conditioning on the GPM/XPM lead at the time the player chose the talent.
Tho I cant still climb like I would, I have to say that I am now getting to some more expensive items during the game when before it would be a tad bit more difficult.
Still, I think a better job can be done.
XiaoZhuUu, Exactly. I completely agree with your point. Though we have to consider other factors like when heroes pick GPM or XPM buff, they feed more gold and experience when they die.
Cannot login Valve?
Please fix eror
I do hope to see more posts about the support meta in this patch, because its interesting to look at since supports had an easier time in patch 6.88. I main support and especially in the lower mmr brackets I find that supporting and support players have become very confused, myself included the first few days of 7.00. my first reaction was " well, shit, i think support is dead. we have an impossible task of warding huge areas with too many juke spots and far too few "flying vision" high ground spots to ward. im constantly behind in xp and gold and cant buy the essential teamfight items to remain relevant to the team forcing me to pick the xpm and gpm buffs to catch up. and oh no my 20 min starve strat for the carries doesnt work anymore."
I still think some of these factors are still an issue but i no longer believe support is dead. what really confused me at first and someone else mentioned is that it feels like supports are forced into a more carry like attitude because remaining relevant throughout the game is hard, and you will be forced to accelerate your economy for warding, items for the team and levels. but actually aside from having to experiment with new ward placement, things have become better for supports. in my opinion its mostly the talents and surprisingly the rework on root, its so much better now.
personally I dont ever consider cooldown reduction especially the heroes who get it early on, because thats not when its going to matter even if support early game is better. as for gold and exp buffs, i think that especially where I play, the choices are so situational and feel so unbalanced that you will always always prefer to improve your stats or spells where possible, working your way to remain relevant, but be forced to take xpm/gpm increase only because its more useful that cooldown reduction.
would love to read what more people think about the buffs and the way supports have to play in the new economy.
Possible complete misunderstanding of something that could be distorting respawn time winrates. If you're winning and dying less, you're less likely to take the respawn time talent, but more likely to win no matter what you take. Correlation /= causation.
There are definitely some valuable points in this article. So I would rather point out something different as I think.
At first, RTR maybe do NOT help (support) players win the game, but RTR does help us ENJOY the game. U can't do anything with death in fountain, right?
Secondly, questions about the choice between more GPM and other buffs in early game have no certain answer for sure. But again, more Gold in support players may NOT help them to win the game, but more gold ensures them more space to make their own choice, in other words, it helps them to ENJOY the game!
So, I am pretty happy to see these new choices in patch 7.00. Do not forget the phase after "Good Luck", it is " Have FUN!!!"~
A little surprised by treant gpm one not being as good....once treant has lvl 4 passive he moves pretty fast anyway....this patch i think you want your support to be a lane dominator
I think a lot of commentors have a good point that breaking this data down further would be helpful, but seeing as we are so early in the meta any further breakdown of the info will be useless within a week. Give the meta a little more time to settle before you dive into it deeper.
The info is good because it gives us a heading on what is generally going to be a better choice. In all games played, 90% of the time it is better to get stats, which I will assume means even when you are losing you are more likely to succeed by not going for the RTR or GPM gains. When we have more info it will help us know when it is a good idea to get those stats, but I think looking at the alternatives to the gold and xp gains, I would almost always go for the stats instead. I mean, how long would it take to recoup the benefits of an extra 40 dmg? That's an entire items worth of damage at a single level gain, rather than slight advantages growing over time. The extra 40 dmg will likely give you the opportunities to get the kills that will boost your xp anyway.
But the bottom line is that we are seeing that the gains are having at most a 10% difference in the outcome of a game. That's nothing to lose our heads over.
I would say the gpm and xpm are probably not as bad as this data suggests but that they aren't necessary becuause of runes.
As a support/offlaner, I'm very happy about the talent tree helping with gpm/xpm depending on the hero. I'd say heroes that have a level 10 boost instead of 15 or 20 have an even bigger advantage right now, too.
"Larger" map, more rune opportunities, harder stacking... Sounds like the skillcap for supports is higher in general.
One more thing only, I see that now and then people are mentioning how the map got larger and it is harder to ward. Doesn't it encourage better warding? I mean, you have to manage your wards, you only have a limited amount of yellow ones and you have to make them count. Just plan ahead, predict whether you are going to be on the defensive or offensive and provide vision accordingly, and remember that in the end it is not going to be perfect all the time.
As support player, lvl 10 talent that provide gpm or xpm bonus is actually neat, since the pace of the game seems to be faster and with that talent, I get to keep up with my core. In my case, higher level of skill for jakiro seems far better than 8% damage amp
as a safelane support, it's harder to shutdown an offlaner now thanks to the every 2min spawn time of neuts.
Looking through the comments again I think @FightMachine has touched on an interesting point that could explain the numbers we are seeing. I think what is happening is that people are either getting higher levels without the items they are used to having at that level, or have too much gold and subconsciously think that having the items means they have the levels. Both sides of the coin lead to people committing to a fight they aren't prepared for. Your power comes from both items and levels, and so the gains on either side aren't dragging you ahead as far as you think they are.
I'd venture to guess that people commit more when they have more items, but feed more gold during pickoffs when they have higher levels.
the gpm talent does indeed come too late for it to matter. If you're behind, its usually game over by the time the 120gpm comes. If it's a back and forth game, it will of course matter. Then again, if you are a position 4 who already has some good farm, it still won't matter =/. There's only so much wards you can buy in 6 minutes
these stats are pretty dumb, you reason low respawn time has a higher w/e is because if you're stomping you just wanna respawn fast, and if you're behind you want gold, simple
I describe this blog by 2 words, "decision making"
изи патч
Ахахахахахаха зайдите на русский паб и посмотрите там всегда по 5 кери лучше бы вы зделали реализм баланса что можно пикать только 2 кери 1 танка и 2 сапорта и жизнь стала бы лучше а сейчас ни о каком балансе говорить не возможно по этому не смешите нас и зделайте реализм баланса и все будет хорошо!!!
Челик, ты на нормал скилле играешь, зачем ты что-то пишешь?
Wards only last 6 minutes, its not enough.
I didn't count, but i think im 25th! or 50th! i dunno! but i commented!
Been maining support for quite a while now and honestly I'm lacking words to describe how bullshit the role now is. And I can't even pin point the what's exactly real problem, it just feels like we're a lot more powerless. Yeah we can get items faster but what does that matter if the enemy cores get ridiculous damage buffs and we die even faster? I'm having a hard time harassing the enemy offlane and keeping an eye on the 3/5/7 min timer to stack camps because the core roles get their lvl 6 by 5/7min and i have to either help my mid with a gank or save some other lane from the enemy mid gank. Another issue is obviously ward placement. I can't feel any other spot besides the cliff being a good ward spot anymore - there are way too many scattered trees that block a lot of vision when wards are placed on the ground, as well as the enemy monkey king 80% of my games - It feels like a big waste to not put a ward on a cliff because of the potential 4sec stun on MK. But since everyone is putting wards on cliffs it becomes pretty obvious where to deward, leaving you depleted of obs with basically 1 ward at all times. Also probably my biggest issue with 7.00 are the players that are coming back after a long break and play core roles while being super rusty. You'd think that a player at 4.5k knows picking a void in a 3 melee team and rushing BF doesn't work but nope. Also slark pickers
@ Child Labour
I think Valve was pretty intentional about their new jungle. A lot of the frustrations you have I think will iron themselves out in time, but the jungle was fairly intentional. They seem to want the laners to rely less on jungle stacks and they want the supports to be concentrating on making intelligent moves rather than afk stacking and pulling. Valve wants the laners to stay in lane, and they have added camps to make jungling viable while still trying to slow down the farm and snowball potential of carries. If they hadn't done this, the carries would probably hit their timings even faster. The jungle is a diferent shape, and I think you'll get used to that over time, but I think you will also have to worry less about the stacks and just concentrate on warding and harassing. It might seem hard now, but with time I think you'll learn what works and what doesn't.
@ Child Labour
some really good points and I feel your pain with warding with MK in every game, you basically had to cliff ward if you did not have a hero with flying vision on the team, and even then you still had too :(. However I feel I am getting a better sense of what I need to be doing at specific times, now that some time has passed and I have a few more games in the new patch(stacking priority/securing runes/ ganking)
got harder since you can stack at odd minutes and rune spawns at every even minutes rotations got hard since map got too many low ground tho wards got more vision cos of the map changes
and thanks we got backpack now we can store wards, tp's and consumable for supports
Wait... Who said omni is a support?