Jump to content

Tosh: When to Ult


tukeykramer

Recommended Posts

I've been working on better understanding when to use Tosh's ult. As a starting point, I decided to assume Tosh is being built primarily around his ult (i.e. I'm giving Tosh's ult the greatest chance of being useful). Also, I'm basing this analysis on a late game, fully itemized Tosh. The build I chose for testing purposes was as follows:

 

Shadowmourne

Isomorphic Pyre

Force of Entropy

Darwin's Might

Bandits Artifice

Time Splitter

 

Shadowmourne, Pyre, FoE, and Time Splitter were chosen because they all have useful effects that proc on attack, which continue to be useful during Tosh's ult. Bandit's was chosen because Tosh's ult affects the resulting clone. Darwin's Might was chosen as an item that makes Tosh more of a tank while also providing some leech when Tosh isn't using his ult.

 

Importantly, this item build causes Tosh to reach his attack speed cap, meaning that Tosh does not gain any additional weapon speed when using his ult. While this simplifies my analysis, I didn't choose this item build to maximize Tosh's attack speed--I believe this is one of the better item builds for a Tosh player who wants to focus primarily on using Tosh's ult. However, this build generally ignores +time items that may improve the usefulness of Tosh's ult.

 

To perform my analysis, I calculated the total damage output of Tosh (at level 18), his clone from Bandit's Artifice, and his spectres (from 0 to 4 spectres) with and without Tosh's ult. I also considered the damage output with and without the clone from Bandit's. Note that I assume that when Tosh uses his ult, his clone does true damage equal to Tosh's true damage (this appeared to be the case in my in game tests with obelisks). I also ignore item-based effects that proc on Tosh's attacks as they are constant regardless of whether or not Tosh is using his ult (with one exception: Shadowmourne's increased crit chance). Note that this is only the case because Tosh's weapon speed is not increased by his ult when using this build.

 

As an example of my calculations, if Tosh has 1 clone and 1 spectre, he will have a total damage output of 200 true damage per attack when using his ult (80 from Tosh, 80 from the clone, and 40 from the spectre). If Tosh doesn't use his ult, he will have a total damage output 613 physical damage per attack (309 from Tosha, 304 from the clonea, and 20 from the spectreb). However, that physical damage will be partially mitigated by the armor of the hero being attacked. I therefore calculate the armor cutoff point at which Tosh does an equal amount of damage with or without using his ult. In this example, if the enemy hero has 37.8 armor, Tosh will do equal damage to that hero with or without Tosh's ult. If the enemy hero has less than 37.8 armor, Tosh is better off not using his ult. If the enemy hero has greater than 37.8 armor, Tosh is better off using his ult.

 

The remaining cutoff points are as follows (I've also converted the armor values into physical resistance percentages for ease of reference):

 

Cutoffs when Tosh has a clone:

Spectres: 0 1 2 3 4

Armor Cutoff 44.08 37.81 32.85 28.78 25.35

Physical Resist 73.9% 68.4% 63.2% 58.4% 53.8%

 

Cutoffs when Tosh does not have a clone:

Spectres: 0 1 2 3 4

Armor Cutoff 44.38 33.12 25.61 20.12 15.86

Physical Resist 74.1% 63.5% 54.2% 45.8% 38.3%

 

As demonstrated above, the utility of Tosh's ult in the late game is greatly dependent on the number of Tosh's spectres. Even so, when giving Tosh's ult the best chance of being useful, late game, Tosh is usually better off not using his ult against a non-agility heroc and he is usually better off using his ult against an agility hero,d regardless of the quantity of Tosh's spectres. When Tosh doesn't have any spectres, he is almost always better off not using his ult.

 

UPDATED: Here are the armor cutoffs with some potential early game builds. In each of the following cases Tosh does not have a clone because he would generally not have enough cash to buy a Bandit's Artifice. I performed these calculations similarly to those performed above, except I also adjusted for the weapon speed boost granted by Tosh's ult (since it actually increases weapon speed in these cases).

 

Level 11: Two Duran's Machetes, Phase Cloak, Pulse Hammer

 

Spectres: 0 1 2 3 4

Armor Cutoff 11.91 3.76 -1.37 -4.93 -7.57

Physical Resist 30.4% 10.8% -4.3% -16.2% -25.9%

 

Level 6: Two Duran's Machetes

 

Spectres: 0 1 2 3 4

Armor Cutoff -21.55 -29.63 -34.69 -38.20 -40.79

Physical Resist -92.8% -146.5% -187.7% -220.1% -246.4%

 

Basically, early and mid game Tosh is almost always better off using his ult, even with zero spectres.

 

One additional test, this time with Pyre:

 

Level 11: Two Duran's Machetes, Isomorphic Pyre (ignoring the spell damage from Pyre)

 

Spectres: 0 1 2 3 4

Armor Cutoff 10.29 2.35 -2.61 -6.05 -8.57

Physical Resist 26.9% 6.9% -8.3% -20.2% -29.8%

 

However, when you include the spell damage from Pyre, even if Pyre only did 30 damage per hit after enemy spell resist, the armor cutoff with zero spectres is 3.9. In other words, with this build, you are virtually always better off using your ult, even when you have zero spectres.

 

 

 

OVERALL SUMMARY: Tosh's ult is very useful early and mid game, but becomes much less useful late game. If you intend for Tosh's primary role to be the use of his ult and build Tosh consistent with that role, his ult is still highly situational. Instead of focusing your late game build primarily around Tosh's ult, you are better off building Tosh to fulfill a different role (e.g. AA or Support/Tank). If you build AA, Tosh's ult will be useful for damage output only when facing an enemy hero with a lot of armor or when you have 3+ spectres. If you build Tank, you will almost always be better off using Tosh's ult.

___________________________

 

Footnotes:

a. Tosh's weapon damage is 281. Shadowmourne adds 20% crit chance, meaning that on average Tosh will deal 309 weapon damage (281*0.80+281*1.5*0.20). Tosh's clone's damage can be calculated in a similar manner (276 weapon damage will deal, on average, 304 weapon damage when accounting for crit damage). Note that this analysis assumes that the mechanic for critical strike is fixed and no longer deals true damage instead of physical damage (while this is not currently the case, it sounds like the devs are working on fixing that mechanic to make it the case). This assumption means that the usefulness of Tosh's ult is currently overstated in my analysis based on the actual, current in-game mechanic.

b. Tosh's spectres have increased attack speed due to his ult. In particular, without Tosh's ult, his spectres have an attack speed of 1.8, and with Tosh's ult, his spectres have an attack speed of 0.6. Because Tosh's spectres are attacking 3x as often when under the influence of his ult, I divide their weapon damage (about 60) by 3 to get an equivalence of 20 physical damage per spectre.

c. Assuming late game armor values of around 20 and assuming Tosh has a clone or has fewer than 4 spectres without a clone.

d. Assuming late game armor values around 40.

Link to comment
Share on other sites

slim jim is broken and needs nerf

 

Any scrub who wants to play Tosh and doesn't take note of this thread is a retard.

 

It's irrelevant to the results, but I'm still curious; when you first decided to do this analysis what did you hypothesize the results would be and before you found out conclusively how were you building Tosh, for the ult or for general AA superiority? Congrats on greatly improving your understanding of the hero and especially your ability to accurately decide how to build and when to ult based on any given situation, and thanks for sharing it with the community.

Link to comment
Share on other sites

It's irrelevant to the results, but I'm still curious; when you first decided to do this analysis what did you hypothesize the results would be and before you found out conclusively how were you building Tosh, for the ult or for general AA superiority? Congrats on greatly improving your understanding of the hero and especially your ability to accurately decide how to build and when to ult based on any given situation, and thanks for sharing it with the community.

 

My hypothesis was that, late game, Tosh's ult would only be useful for outputting damage when Tosh had 3+ spectres. I was actually somewhat surprised that there are some circumstances where Tosh has four spectres and does more damage if he doesn't use his ult (e.g. against a hero with less than 38% physical damage resist or 53% damage resist if Tosh has a clone).

 

Since I tend to play Tosh against players who work very hard to kill my spectres, I have generally built Tosh to be versitile (i.e. good with or without his ult).

Link to comment
Share on other sites

I think your analysis is somewhat biased because your build maximally utllizes his ult rather than his regular AA.

 

I think it would be more prudent to compare a build like this with a different build that maximized non-ult DPS, in order to better compare which type of build one might like to strive for. With this particular build, it surprises you that ult DPS falls shorts in many cases, but with a different build, you could find that using his ult falls short in basically all cases.

Link to comment
Share on other sites

I think your analysis is somewhat biased because your build maximally utllizes his ult rather than his regular AA.

 

I think it would be more prudent to compare a build like this with a different build that maximized non-ult DPS, in order to better compare which type of build one might like to strive for. With this particular build, it surprises you that ult DPS falls shorts in many cases, but with a different build, you could find that using his ult falls short in basically all cases.

 

The main point of my first analysis is that even if you choose a build that will maximize the utility of Tosh's ult in late game, it's still highly situational (i.e. you aren't always better off using your ult). Therefore, late game, it may make sense to focus your build more on other roles (e.g. AA DPS or Tank/Support) and to only use Tosh's ult when you have 3+ spectres or when you need the movement speed. In other words, the analysis is intended to be biased in favor of his ult to show that even when you bias in favor of his ult, late game it isn't as useful as some might expect.

 

Also, while there are some builds where Tosh would always be better off using his ult, those builds generally don't maximize the damage output of Tosh's ult. For example, if Tosh builds only tank items, he will probably always be better off using his ult (even with zero spectres). However, I'm trying to focus on builds that try to make Tosh's ult his primary function (as opposed to tanking or providing CC).

Link to comment
Share on other sites

Very nice analysis kramer.

 

Do you think that a tank build for Tosh might be the best idea for a late game build then? Mid-game, pyre makes Tosh a beast regardless. And I'm still questioning whether an AA dps build is best on tosh because his ult gets ignored.

 

I was thinking:

 

Pyre

Force of Entropy

Eternal Overdrive

Yamato Reactor

Coat of Arms

Organic Carapace

 

thoughts anyone?

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.

×
×
  • Create New...