You are saying the same again. If you have the 3-head hydra, 2-head does nothing. But yeah, if you dont have 3-head hydra (as in your example here), then 2-head hydra works again. Still doesn’t make it work when you do have the 3-head hydra 
For someone who have spent a lot of time arguing that games need to be simple and easy to understand (which I disagree with), it is quite surprising to see you arguing for having skills do nothing.
It would be the optimal play if our re-spec is expensive (something you personally always wanted). 
I do, in a sense (actually I havent said respecs must be expensive, they can be completely free, if they have a long cooldown. A cooldown is better than a cost, since costs tend to become irrelevant over time. Not to mention, respecs should be full respecs. Paying per point you respec is a bad idea).
But the optimal play here would either be to:
Anyway, I wonder, which item X are you imagining that would work with 2-headed hydra and not 3-headed hydra? That likewise sound weird?
Unlike, if you pick a frost hydra instead of a fire hydra, it might make sense than an item that increases hydras frost dmg, only works with the former - although item affixes should not be that specialized regardless.
Why? It makes exactly such types of situations interesting - when you go sub-optimal until some point in time when you re-spec specific skills. It involves more skill than the full re-spec.
Re-specing is a huge topic. Besides this here (for which I am arguing just for the fun of it, which doesn’t make my argument not valid however 
) I’d go for lite re-specing in D4 so casuals are happy.
It isn’t weird at all - we should have items building upon all skills, not only the top tiers of the pyramid. For example, a property like “If your Hydra doesn’t have more than 2 heads, do this…”.
Ah yes you mean what I had been advocating for last week ago ? ^^
Well it could be simple if the “open” nodes are just for one category of skills, but it wouldn’t resolve the issue of too much stuff on the tree.
Anyway, I have a hard time imagining a satisfying tree with active skills. It should be as it was, with only passives but much more than the Blizzcon version, and more intricate.
Because it means you can do respecs cheaper.
Like have two builds where the only difference is fireball in build A and Frostbolt in build B. That is suddenly around 1/6 the cost of a full respec. Or maybe you just change a few points in hydra upgrades, and you now have different builds for a small cost compared to full respecs.
All respecs should have the cost of a full respec, no matter how little you change.
The upgrades are not skills, they are upgrades. The skill is Hydra, no matter if you pick many or few upgrades.
Terrible imo. Items should be generalized. Working with all kinds of skills at the same time. And certainly not only a single upgrade in a single skill. You would need to add tens of thousands of legendaries to do the same as a few hundred legendaries can do otherwise.
Which is both impossible for Blizzard to design, and would take RNG frustration to a whole other level.
You could pretty much remove the active skills from Lolli42s picture, and it wouldnt change anything (other than making the tree harder to read).
Though it might be easier to design and add the “cross-over” upgrades, as in upgrades that work for multiple skills, if you had no skills in there.
Honestly, the most realistic idea is the one Last Epoch uses. I think giant trees can be interesting, with all the cross-overs and branching.
But having a skill tree per skill just seems much easier to design and balance. While still offering a good amount of choices.
Yes, I said it in my post, it wouldn’t resolve the overstuffing of the tree. For that, you’d need to rethink the concept of “one-skill” modifiers into groups like “skill-category” and/or “damage-type” modifiers. And maybe have just a few “one-skill” modifiers in the highest branches.
EDIT : Well you edited while I was writing. ^^ We think alike.
Though I understand the issue in the Last Epoch system, in the end it lacks clarity. Maybe Diablo should go another way.
Yeah. I’d be fine with a mix (as mentioned much earlier in the thread, the one thing I would change in Lollis picture, would be adding way more upgrades that worked with more skills at the same time). but even then I’d say it should be >50% skill-specific upgrades and <50% generic upgrades (the generic upgrades are still in both Blizzards and Lollis versions, they are just in the roots - but since you get the passive points in the active tree, the cross-skill upgrades are still part of that tree, in a roundabout way).
My reason for having mostly skill specific upgrades is exactly that the more generic upgrades should be from items.
We dont really need +% generic dmg upgrades in there, since that is what items already delivers.
Skill upgrades should be all the stuff that is too narrow to be viable on items.
Both as normal affixes and legendaries). And of course, since nothing is black or white; there will always be some overlap between what skill mods and what items can offer. The difference should just be the general goal.
If you made an item that says “Fireball does X” then either make it into “Projectile skills does X”, or make it into a fireball modification/upgrade node instead.
If you have to kill some boss to gain a re-spec point, you would be forcing the player to do it as many times as skills he have just for a single re-spec. This won’t be productive. Re-specing involves highest skill when it’s properly balanced regarding how much time\currency it costs to the player.
Being able to re-spec each node you want would be too easy. Being able to re-spec the whole tree only would be too time demanding. Re-specing part of the points in each pyramid is the way to go (say all those above the one you need to re-spec).
That’s very hard to balance. They won’t do it. And they don’t need to do it.
That’s fun, no?
I agree. Many generic nodes that Lolli42 put in the base of branches can be taken out (I understand why he did though) and others can serve for several skills (“+ number of shards/bolts/spears”, “Burning targets take more damage”, “Projectiles moves faster” etc.) Of course there are still some very skill-specific modifiers that can’t be put together.
There are also many skill-enhancing modifiers in the roots (“Burning effects last longer”, “Casting a Cold spell applies Chill to nearby enemies” etc.) that I think should be in the branches. I do like the root concept but these should be more generic upgrades like “Damage increase when not moving for 3s” or “+Crackling energy”.
I am not saying are respec has to force you to place all your points again for no reason. Just that unlocking the ability to change even one skill, should be the full respec cost. After you “pay” that, you can change 1 point or 100 points.
Exactly why it shouldnt be respec points. Just one cost for a full respec.
Also, having to kill a boss to respec is not good. This is why cooldowns is kinda more realistic. What if you made a bad build that cant kill boss X? Then you can never respec? Wont work.
The best could be a combination imo: Every other week you get a free respec. But if you want to respec more often, it comes at a really high cost (and probably still a short cooldown, to prevent the system from breaking down even if the intended high cost turns out not to be so high after all).
I dont see why it would be. Balance the skills instead of the items primarily. Blizzards refusal to touch skills in D3 balancing is a major mistake.
Certainly not harder than balancing 10,000s of legendaries individually.
Just unrealistic to do. And also not very fun, when you cant find any useful items anymore.
The main purpose of making items more general is to get away from the RNG Hell of Diablo 3 (which then lead them to increasing droprates to silly rates, to get out of RNG Hell).
I like it.
While I’ve read a majority of your posts, I haven’t processed it all fully.
First thing I wanted to mention was the structure of the tree. I think Lolli’s picture is ideal, but, would probably be a bit difficult to navigate on console. I think if you were to take the tier’d structure (with point requirements) that would be next best thing. Passive tree included (for each element) would be cool.
The main problems with this I think would be the learning curve and point distribution. A tree like this would be intimidating to some, and there would have to be numerous skill points given per level to allow for full customization.
A good approach to this, in my opinion, would be to limit the number of skills you can obtain in the early game, thus forcing the player into learning the tree. Then maybe in the later acts of the story, there could be a limited or paid respec quest, allowing the player to re-distribute.
I still think everyone overemphasize how complexity/depth turns people off.
It should be solvable. Like the first time you play each class the game could ‘recommend’ skills to pick at the start. And when you have chosen a skill, which upgrades to choose each time you lvl.
As for multiple skill points per level, sure, that doesn’t seem like a problem.
However, it could also be a reason to reintroduce the tomes Blizzard showed at Blizzcon, where in end-game, even after you hit max lvl, you could still get some skill points by finding tomes in the world (but not unlimited!)
Yeah, that’s true. I think the idea of complexity is open to interpretation, depending on the audience.
I forgot about the tomes to be honest. Those would be a good way to scale skill power in the late game.
It is not what I personally would prefer, but I still gave you an upvote for the effort and because it is better than the current one.
Full re-specing serves for less skilled game since one player could have made a single mistake, the other many, and both would have to do a full re-spec for the same cost.
It has to be an activity (boss or farming some material), not a cooldown. Blizzard would never do only cooldowns for re-spec since many players would stop playing the moment they make a mistake until the cooldown of re-specing passes.
In Diablo 4 I’d make re-specing as:
1] Easy material farming for 80% re-spec of points total
2] Medium difficulty dungeon for 90%
3] Hard one for 100%
So, every player could farm some of the 80% re-spec at any time. If you are level 40 and you have 100 points to distribute, re-specing with material from 1] would leave you with 80 points to distribute.
It would be a full re-spec meaning that 20% of your points stay locked. You can’t use them until you don’t use another material or the daily/weekly re-spec option at vendor. The daily re-spec option would give you 90% of points, the weekly would give you 100%.
So, in short, full re-spec based both on materials and cooldowns. Every type of player would be satisfied with this.
Not possible if you have more than one general legendary effect (such affecting many skills) per skill/effect/passive from the tree. It’s unsolvable if we aim skill balance (if we don’t then it’s possible, but the game would be totally broken from OP combos). And we all want more than one item for say buffing the 3-headed Hydra effect.
That’s when crafting comes.
Besides, having an easy way to re-spec fits perfectly with having many legendary effects in the game. Say I am at 90% points with my build A, then I find some new very strong legendary, but I already used my daily re-spec. I still have the 80% re-spec materials in me. So, I am forced with a choice - keep my current build A, or make another build B based on the new legendary I just found, but using 80% of the skill tree points I have? That’s good design forcing players to make decisions.
I think that a simple orb or regret style item would do it^^
Farm it, use it for individual points
Orb of regret is fine when you don’t have to re-spec drastically. With current skill tree it would be totally enough, but with a way bigger skill tree the full re-spec would be more convenient.
That’s true but since blizz most likely won’t make a very big tree, these will work 100%
It is fine if they release the game with a small tree and gradually build upon it transforming it to something entirely different at some point.
What Diablo 4 should aim at start is to draw the casual and new players to the genre - easy to learn should be the biggest priority at launch. That’s another reason why a fully polished game isn’t needed at all at launch.