Death Ball so hard

The past international saw Newbee dominate with the 5 man death ball. Get a pushing hero to 6, group up, take towers, get ahead, push more, get further ahead, and win. This is the death ball way. It became so prevalent that it pushed other strategies out of the meta and wrapped how people played the game leading to a structural change for the tower glyph and gold and hero balancing (sorry, DP.) Re-balancing the game in this way reduced the incentive to employ this strategy, but it's still powerful.

The poster heroes for taking towers are Pugna, Death Prophet (DP), Shadow Shaman (SS), Nature's Prophet (NP) and the big bad wolf (Lycan). Since the nerf stick hit the DP and Lycan directly, they haven't been first pick/first ban in pro games like they were during The International 2014 but are making appearances in the hands of capable teams and players, like EG's Arteezy and Cloud 9's Eternal Envy. Pugna and the SS have fallen out of favor as a direct result of the meta shift even though they are still viable picks although more niche. The NP is a bit of an oddity here because he's picked to globally gank and fight early, evident by the popularity of the Null Talisman and Blademail build.

The strategy was too oppressive, limiting the number of viable strategies and heroes. Teams with a death ball could take two towers in quick succession by winning a team fight at a tower, executing a successful gank, or slow sieging a tower with a counter initiation edge (think Dark Seer) giving them a huge lead. This lead gave more map control the the death ball leading to Roshan, more towers, securing the enemy jungle, et al. While the lead wasn't always insurmountable, it gave significant percentage points to teams who could reliably leverage it into a win.

Why doesn't the death ball bring all the boys to the yard like it used to? Mostly it's the nerf stick, let's say 50% can be attributed to direct nerfing; Lycan, DP, and NP see play, but NP has evolved so he only counts as half leading to 2.5 out of my 5 poster children.. What about the remaining 50%? I have a few theories, and they go as such:

1. Team theory - certain teams do better at certain strategies because of their competent hero pool (who they pick and can play at a professional level) and decision making processes (why they pick certain heroes and how they use them). For tier 1 teams, on an average day, they are all within a few percentage points of each other and a few more percentage points above tier 1.5 and more yet over tier 2 teams. As the meta shifts, these percentages also shift in favor of the team which can leverage those shifts into wins. The examples of this are Alliance and Navi who were so dominant up to the last International, but derailed and haven't found much success since. Thus, because there are narrow margins separating pro teams, shifts in the meta create an environment for teams to capitalize on them based on who they are, dooming teams who relied on strategies which have been pushed out. 
2. Power Curve theory - with each update and introduction of new heroes, it shifts the aggregate power of all heroes and re-balances the relative power level of all heroes. All you have to do is watch the Beyond the Summit 2 game between Team Tinker and C9 where Lifestealer went up against Terrorblase and it was ugly. TB raised the power curve, pushing comparable heroes out - sorry, naix -  and raising the stock of counter picks. A similar phenomena happens with an update, but the aggregate power level may not grow or might even shrink. Since the international, the updates and introduction of more heroes lowered the stock of most death balling heroes or pushed them out of the meta completely; Pugna, Dragon Knight, SS, Enchatress, Chen, Clockwerk, and Lone Druid see basically no play as attributed to the Powe Curve. As one commentator put it, Chen creeps run toward TB and die before they get there.

Whatever the case may be, the death ball is gone for now. Teams have turned toward line ups oriented toward team fighting or brute forcing their way to victory. They no longer pressure the map in terms of towers but take an active approach toward securing an advantage through directly pressuring key heroes. This is where the current meta sits.

Foreseer's Contract Update and Nemesis Assassin Event Thoughts

The Foreseer's Contract has been issued. After a cool digital comic which offered a peak at the newest hero to sign their name in blood for ancient's struggle, the Oracle was made live. I'll highlight some of the additions, positive and negative.

• The new Phantom Assassin arcana looks awesome, but disappointing that it's only a weapon. The model change, alternate animations, and in-game icon are well done. My disappoint with it only being a weapon stems from wanting to use a whole set, not just 4 of the 5 piece or whatever the case may be. Maybe Valve should add an "augment hero" equip slot for occations such as this.
• The Oracle as a hero seems underwhelming in power level, but I don't mind. At least for the moment, the Oracle isn't breaking pubs the way some previous heroes have (Terrorblade I'm looking at you.) I would much rather have the Oracle start a little under-powered and get tuned once it's better understood how he interacts with heroes than the alternative.
• Buying the prediction item every time seems a little greedy for Valve. If it were me, buying the PA arcana would give you one everyday for the duration of the event (as long as you used it up the previous day.)
• Playing with the in game contract idea with reward is super fun. It introduces an objective that doesn't create an advantage for either team. Creation of such an objective with a sought after reward (an awesome set)  adds dimensions of freshness and fun to the game.

Drafting the PA because of the event and Oracle because of new hero infatuation at higher rates is warping the pub metagame. Some pub analysis: Level 1 is the Oracle and PA. They are bad against level 2 heroes and sharing the Razor and Sand King. Level 3 heroes, then, are those good against both the Razor and SK, which are the Bloodseeker and Sniper. However, the Sniper is not a true level 3 hero because of his inherent weakness to PA. This leaves only the Bloodseeker seeing an increase in popularity. It should also be noted that SK is strong against Razor making him the most likely to see more play and an increase in win percentage. It will be interesting to see what tomorrows number from DOTABuff bring.

Level 1	PA:	Oracle
Level 2	Razor	Razor
	Omniknight	Earth Spirit
	Sand King	Phantom Assassin
	Centaur Warrunner	Magnus
	Ogre Magi	Sand King
	Lion	Leshrac
	Axe	Pugna
	Crystal Maiden	Undying
Level 2	Razor:	Sand King:
	Drow Ranger	Zeus
	Sniper	Techies
	Techies	Spectre
	Viper	Bloodseeker
	Meepo	Juggernaut
	Bloodseek	Phoenix
	Sand King	Sniper
	Ember Spirit	Elder Titan

The Top 5 support picks in pro DOTA

Gone are the days of the brown boots supports. Thank goodness. Not only are the games more exciting with supports who have their force staff and blink dagger at the ready, but their team contributions to these games can tip the scales of a close game in their favor. In no particular order, my top five supports for current pro play are: Skywrath Mage, Ogre Magi, Mirana, Elder Titan, and Vengeful Spirit.

The Skywrath Mage (DOTA buff profile) is probably the best support in CM currently. His instant silence in the hands of a pro player makes Pandas look silly as they blink on to shake hands with Ancient Seal. Seal also makes heros like Storm, Weaver, Puck, and most supports over glorified creeps. Concussive shot is the other spell which needs mentioning here. The range on this spell is a ludicrous 1600 units, which makes it perfect for setting up ganks or zoning heroes during the laning stage and beyond. This spell, like Mirana's sacred arrow, takes advantage of being cast from the fog of war; mid heroes beware.

The bra-less wonder. Ogre McGee. Our favorite melee support, the Ogre Magi (DOTA buff), received a buff of last patch (6.82) to his 2X multicast, allowing him to make an impact on the game earlier, and base HP regen to 2.5, increasing his ability to harass. These seemly small buffs create a larger window of time for this hero to maintain relevance, propelling him to a tier 1 support. The Magi signature spell is his fireblast, which also received a decrease in damage but an increase in the duration between each proc thus stunning a target for longer. This reliable, instant cast stun provides the burst damage necessary to gank and blow up other heroes. The other reason to pick up this hero is to boost a right-clicker's dps and movement speed with bloodlust. At level 4, the 50% increased attack speed and 16% movement speed is like having 1.5 yashas equipped (with out the bonus states).

Mirana (DOTA buff) is the most polar hero in this pool. She fluctuates from unbeatable to a waste of space, all depending on how prepared people are for the arrow. A team constantly on guard and placing wards will make a roaming Mirana look embarrassing. However, if those arrow are hitting and translating into kills, the early game lead, and probably the game itself, is yours. Mirana also pairs well with other supports who can set up arrows like Ogre, Venge, Shadow Demon, and Rubbik to name a few. Lastly on the ability front, Moonlight shadow adds a unique dimension to team fighting and ganking. The other thing about Mirana that is unique among these 5 supports is that she can play any position, from farming carry to brown boots support so picking her early or late isn't a liability and reveals very little about your strategy.

Team Secret's Big Daddy has been playing out of his mind on the Elder Titan (DOTA buff)recently in the support role rather than as an offlaner. This hero offers unique effects which only get better as the game goes on, namely Natural Order. Removing 100% base armor cripples stat based carries, like Morphling and Drow, as well as amplifying your teams damage. This hero also scales well with items, but is not dependent on them because of Astral Spirit. His stomp offers a long range initiation that sleeps units for a full 5 seconds, allowing your team and an earthsplitter to set up for a complete blow out. The only weakness of this hero is that his laning presence early is poor because stomp takes time to channel and he is melee so he has trouble harassing and zoning enemy heroes.

The Vengeful Spirit (DOTA buff) is well positioned right now because she works well with farming heroes that rely on physical damage - like Razor, Death Prophet, and Lycan, offers a bat rider-like ability to remove an enemy hero from play at instant speed via swap, increases the teams damage and has a reliable stun. Nether Swap and the increased damage through vengeance aura and wave of terror are the reasons to pick the Venge. She also facilitates Roshan with all the minus armor she's packing, especially if you get a medallion of courage. This item seems to have the DOTA communities consensus on being a core item for her. Shendelzare (yes, that's her name) can also play more of a core role, but operates just fine without items.

These 5 supports the best at what they do. Items aren't a necessity and they all add a dimension to your team which creates space for the team's cores to farm. Next time you play have to play a support, keep these 5 in mind; they won't steer you wrong. 

More Midas, More Problems

                Is midas really worth it? This item embodies many of the great aspect of DOTA. Are you behind in farm? Grab a midas.  Are you really far ahead from the 6-15 minute mark? Grab a midas. Your game has degenerated to a farming war? Grab a midas. The opposing team is going to pressure your towers from the 10 minute mark on? Don’t grab a midas. The item itself represents an investment of time and gold which you hope will pay back dividends in the future. Once purchased, the repayment period for the midas is 11 uses or 16 minutes & 40 seconds. Assuming that a player is averaging 400 gold per minutes (gpm) prior to purchasing it, the initial time investment 2050 gold / 400 gpm = 5 minutes 7 seconds to farm it. Let’s assume that our hero continues to farm 400 gpm plus their newly acquire midas - who’s added gpm is 2050 gold/16.66 minutes = 123 gpm. This makes for a grand farming total of 523 gpm. This changes the 16 minute and 40 second payback period to a true repayment period of 2050 gold / 523 gmp = 3 minutes and 55 seconds. To sum it up, the time it would take to save enough gold to purchase a midas and then farm up enough gold to where it paid for itself would be 9 minutes and 3 seconds. A player gains or loses about a minute for each 100 gpm difference from the baseline case of 400 gpm.

                With all these numbers, let’s try to make sense of what’s going on. The initial time investment is the period of time when the player is working toward getting their item. This is the point when the player is most vulnerable to having their farm disrupted. Each minute of decreased gpm, from deaths for example, during this period adds another minute that the game must go on in order for the midas to generate value. The repayment period is the amount of time that the game has to go on for the midas to break even. It needs another use after that (100 seconds) to generate value. Therefore the game must go on for 18 minutes and 20 seconds for the midas to turn a net profit in isolation. The true repayment period of the midas represents the aggregated affect the item has on a players farming with a players gpm. Finally there is a roughly 9 minutes of cost to break even on time investment that is, to have 2050 gold again.

All these tangibles are important to understand when thinking about purchasing a midas. If you won’t have about 5 minutes of uninterrupted farm, it doesn’t make much sense to purchase a midas because it will push back the true repayment period, delaying your relevance in the game giving. The repayment period gives us an idea of how long the game needs to last from point of purchase it to have a positive return on investment. If the game ends before the 16 minute and 40 seconds window is up, you would have been much better off purchasing a different item which would have improved your hero in some way other than gpm and slight attack speed increase. If the game does go the distance of 18 minutes and 20 seconds from point of purchase, then the true cost of the midas is what we need to worry about. This tells us how long after the purchase of the golden hand we can expect to start getting other items and begin impacting the game in a more noticeable way.

Many of the things discussed here are tangible and easy to quantify. There are still many aspects of the midas worthy of discussion, like the impact of the extra experience or the increase in the value of a creep wave, but weren’t touched on here. The inspiration for this post is to get a rough idea on the worthiness of purchasing the golden hand and a starting point for making sound decisions. Obviously I can’t factor in every variable; however a baseline has been established and will be built upon. Below I have provided a table and graph for quick reference when deciding to purchase a midas.

Player GPM:	Initial Investment	Player GPM with Midas	True Repayment time	Total Time invested
100	20:30	223	09:12	29:42
150	13:40	273	07:31	21:11
200	10:15	323	06:21	16:36
250	08:12	373	05:30	13:42
300	06:50	423	04:51	11:41
350	05:51	473	04:20	10:11
400	05:07	523	03:55	09:03
450	04:33	573	03:35	08:08
500	04:06	623	03:17	07:23
550	03:44	673	03:03	06:46
600	03:25	723	02:50	06:15
650	03:09	773	02:39	05:48