I had a lot of fun with Magnus BEFORE they rolled his cleave and Overpower into the same skill. Now, early game, with a double nuke he's pretty nasty. Blink daggering into their team and charging 2-3 of them back into yours and ulting I would argue is on par with a Tide ult, but better imo. The stun is longer, and as long as it hits 2-3 of them with a strong initiate, you're going to win the team fight barring something like a smart Void hanging way back so he can leap in and Chrono.
Also, he just makes any melee carry you have smash face, and I usually build him semi-carry after the dagger, getting a Buriza then Satanic if you can get the money, at least the crit and Helm of the Dominator.
I've always liked that hero, and back in DotA1 had like a 100 minute game with him where the other team had gotten mega creeps but I was so farmed with Magnus that my crit cleaves were almost one-shotting entire waves with an Assault Cur and all. Fun hero, and his initiate power is amazing. Exactly what I wanted them to put into DotA2, it was a little frustrating that Tide and SandK were the only really amazing strength initiates they had put in the game.
Is it just me or does Puck fall off towards the later game? His spells do little damage compared to a lot of other heroes and Dream Coil just feels like it's overall worse than, say, Ravage nine times out of ten. 500 damage with scepter boost (and ONLY if they break the coil) compared to 450 right up front, and the stun isn't really much longer than Ravage's (3.5 total stun time compared to about 3.2.)
Because his spells do so little damage and have semi-long cooldowns he's not amazing at pushing or counterpushing. Seems like the only thing he's good at in the later phases of the game is just staying alive.
After a generic blink/guinsoo's throw in skadi/mjoll which is always the favorite caster combo and a buriza. If you can farm all of that gold you won't fall off at all, the ability to constantly chase someone with that much auto attack damage will keep you extremely relevant if you play properly. Puck can easily gun down some pretty tough heroes just by being able to constantly dodge and blink around. Imagine a fed SB who can't do anything because every time he gets next to Puck, Puck just phase/blinks away and keeps attacking him until he's dead.
Of course, you might say "anyone is good with those items," but again, if the game is over at blink I don't see why he needs to be relevant late game, the game is already over. Most games are over in 30 minutes and his spells are still pretty relevant. As much as any other int hero, anyway. Compare him to Lion who does waaaaaaaaaay less damage in a team fight. Two very reliable CCs, yes, but that's the trade you get to make as an int hero > CC for damage. If you have a Dark Seer to vac everyone together, Puck by himself can hit two thousand damage with his two basic spells, that's pretty neat. Lina does more, but Lina doesn't have Dream Coil or phase, either.
I personally don't like Puck because I just can't stand his early game. Cower at the tower and shoot off Illu orbs to try to scrape together enough last hits for energy boots so that you can start spamming creep waves for blink so that you can FINALLY be useful beyond ult>die. Talk about your crappy auto attacks from a hero with pretty much zero lane presence.
All of this could be relevant if this was a game where you pick your custom hero for whom you could pick whatever skills you want. Sure I'd make a hero with Ravage instead of Dream Coil and Chaotic Offering instead of Summon Familiars, but unfortunate as it is you can't separate heroes from their skills in Dota which makes comparing random parts of different packages moot. "These skills are both ultimates, they should be equally strong" and similar statements only cause headache for me.
If you aren't comparing abilities between heroes that share similar roles, what are you going to be drawing your data from? Just random "oh hey this feels like it'd be a good idea"?
So why is that? Where do you draw the data from? How would you make both ultimates - which occupy the same general niche - both be equally scary? Is Tide/Ravage too good, or is Dream Coil/Puck too weak? Is it just due to the way teams are currently playing and operating (the "metagame") or is it an issue that would be present regardless if things shifted so that they favored a hero like Puck?
How else are you supposed to obtain objective data to make decisions with?
If you can't do that, you might just as well debate on the differences of couriers.
Is it just due to the "metagame," that Tide's kit is just preferred over Puck's? What if both are acceptable within the current play environment? Does that just mean one is more of a niche pick than the other, or is one too strong or too weak?
That's the point, that's why I'm asking. Where does Icefrog pull his data from? How is he determining what needs improvements, what needs nerfs, what needs no changes at all? Knowing that would lead to a much better understanding of the game.
Puck does get picked, but not as often as Tide. Soon there will be 100 heroes in the game, that would mean that for every hero to be evenly picked you'd see every hero in one out of 10 games. That's not a lot, and with the amount of carries and hard carries in the game that isn't even remotely possible within the confines of the meta-game. Tidehunter can lane well in any position except jungle and solo mid, while Puck demands a solo lane. Flexible heroes will always get picked more often just because of drafting strategy alone. Niche heroes will always get picked less because there's more options for their position. There aren't a lot of good support heroes, so the good ones get picked a lot because most of the time a team wants to pick one or more of them.
If you don't buy that, take a look at the TI2 statistics. Puck gets picked/banned at 10% of the rate of Tidehunter but has 78% win rate compared to Tidehunters 52%. Obviously Ferrari_430 and Dendi along with small sample size have a lot to do with the high win rate, but the point still stands.
Yeah. The question is, should this just be accepted as part of the game, or would it be possible to try and make adjustments here and there to try and make every hero (or as many as possible) have multiple ways of playing them? Like you said, Tide can do a lot of different things, and so can heroes like Profit, Dark Seer, and some others, which would lead to them being used more frequently than other choices simply due to their versatility.Puck does get picked, but not as often as Tide. Soon there will be 100 heroes in the game, that would mean that for every hero to be evenly picked you'd see every hero in one out of 10 games. That's not a lot, and with the amount of carries and hard carries in the game that isn't even remotely possible within the confines of the meta-game. Tidehunter can lane well in any position except jungle and solo mid, while Puck demands a solo lane. Flexible heroes will always get picked more often just because of drafting strategy alone. Niche heroes will always get picked less because there's more options for their position. There aren't a lot of good support heroes, so the good ones get picked a lot because most of the time a team wants to pick one or more of them.
Do you think trying to have as many heroes as possible have this versatility is a good design goal, or do you think it's better to have a handful of all-rounders with most of the other heroes fill in specific roles and niches?
I don't really like the idea of enforcing change in meta any more than I like the idea of Riot enforcing no change in meta in lol. Dota pro scene has shown it's ability to transform and evolve on it's own many times before, and I prefer seeing the players discover and try out new things on their own to everyone rushing to pick the freshly buffed new fotm hero. Adjustments are good, as most in the latest patch are, but they should be made with time and consideration and not because people want to see that hero that hasn't been picked in a month. e.g. Morpling.