I don't know if you are familiar with DoTA, but can you comment on why DoTA doesn't seem to have the problem with games stalling as players fight for vision control? I believe the DoTA ward model is very similar to the original LoL ward model.
Here's a quick rundown of the three systems:
Dota:
Observer Wards: Invisible, last seven minutes and give vision of a fairly large area, but there are only four available to each team at a given time. Cost 150 gold for a stack of 2.
Sentry Wards: Invisible, last four minutes and reveal invisible things in a fairly large area, but don’t grant vision of an area (so you need to find some other way to get vision of the area where the invisible thing you want to see is). Cost 200 gold for a stack of 2. There is no limit to Sentry Wards.
Old League of Legends:
Sight Wards: Invisible, last three minutes, give vision of a fairly large area, take three hits to destroy. Cost 75 gold. Unlimited.
Vision Wards: Invisible, last three minutes, take three hits to destroy. Give vision of the same area as Sight Wards, and also reveal invisible things in that area. Cost 150 gold. Unlimited.
New League of Legends:
Sight Wards: Invisible, last 3 minutes, give vision of a fairly large area, take 3 hits to destroy. Each player can only place three Sight Wards at a time. Placing a fourth will destroy your oldest ward. Cost 75 gold.
Vision Wards: Visible, last forever until destroyed, give vision of the same area as Sight Wards, reveal invisible things, take 5 hits to destroy. You can only place one Vision Ward. Cost 150 gold.
My not-at-all-certain observations are that 1) DotA's jungle is more complex and perhaps a bit larger and 2) DotA's terrain behavior is more advanced [tree removal, cliffs overlooking the river]. These factors would probably make it more difficult in DotA to place wards in the jungle in a way that is highly likely to yield solid intel — you can get a fairly good amount of safe coverage with a defensive ward on the cliff near river, as well as along the river exit once you've advanced past the river. To the best of my knowledge, you're just not going to learn as much from jungle wards in DotA, because of the sight lines and the difficulty of enemy jungle invasion (and perhaps because of the random invisibility rune spawns, though I don't know how much of a factor that is). Contrast all of that with LoL: accessing the enemy jungle is comparatively easy, and key ward placements on the enemy side (especially tri-brush, blue buff brush, and river junction at wraith camp) are almost certain to be high-information investments.
Additionally, in LoL you have to teleport to a friendly creep, structure, or ward; the secondary nature of a ward as a rally point is something that increases the stakes of ward superiority. In DotA (if I'm not mistaken) the teleport limitation is not the same, so the pressure is not the same. I'm sure there are also itemization and ability factors that I have no knowledge of whatsoever. If I'm horribly wrong on this, I welcome a more experienced perspective!
EDIT: `chc has more specific intel on the wards themselves. The team-wide limit on observer wards is a distinct difference!
If anything, teleports are more limited in DotA. Teleport scrolls can only target allied buildings, or a small area around them. Boots of Travel can also target allied units, but are rather expensive, and thus typically only seen on certain heroes which particularly benefit from them (especially Tinker).
However, I would say that Dota still has a problem with stalling, maybe even more than League of Legends. There are often one-sided games that drag on for quite a long time before the winning team has a decisive enough advantage to actually finish the game. This comes from a few features, foremost 'buyback' which lets players come back from the dead before their timer expires for a fee, which means that even if you kill their entire team inside their base, unless you have the force to kill them all again, you usually cannot take any ground.
I am most curious about the ward aspect, specifically. Are you saying DoTA has less vision choices? And you are referencing smoke?
That's been in flux recently! Patch 6.81, which was active from April 2014 until a few weeks ago, devolved into a "deathball" meta which tended to end games early. (This was responsible for the rather anticlimactic conclusion of The International this past July.) Patch 6.82 introduced some new "comeback" mechanics, which generally enhance the ability of a team that's playing from behind to recover after taking a teamfight, or finding high-value pickoffs. The details are still being tweaked, but so far the result does seem to have been less one-sided games, both in amateur and pro games.
And yes, "team stealth" is referring to Smoke of Deceit. A really interesting item, available in limited quantities, which allows a group of players to become temporarily invisible and sneak around the map -- even under sentry wards, which usually reveal invisible heroes! -- until they close in on enemy heroes or towers.
maybe in the pro games, but even then that's a pretty iffy statement. a bad engage in a teamfight or surprise gank can swing things pretty hard if the teams know what they're doing. in anything below, say, high diamond there's just too many mistakes to make this claim.
For example, in all of the finals of the recent International, there was a leader at 10m who won the game. I'm not going to go through all the games and check, but I'd guess we'd find this to be true the vast majority of the time in professional DotA games, as it is in League games.
Not to make too sweeping of a generalization, but I expect it is inherent in games with an economy and relatively low random factors. On top of the normal expectation that who gains a lead is likely to be more skilled, economies that allow you to build additional strength via the dominance conferred by your current strength reinforce existing leads.
Comebacks are fairly frequent in dota.