[GemCraft chapter 0] Perfect, Optimized Gemcraft Playing

37 posts

Flag Post
Part 0 (like Chapter 0)

The ideal start is basically the most mana-efficient technique to create a mana farm. The first thing one should focus on is creating a single mana gem, but even before that, they should create 3 traps, all with grade 4 RO gems by transmuting both grade 3 R and O gems onto your 6 other grade 3 gems; then combining these (max stats is assumed), and combine one particular RO gem with two other Reds and two other Oranges (which are created through gem transmutation) so there is a splash radius of greater than 15, so it hits 5 targets. Put these in a trap. Anger the first wave as many times as possible (without angering with your grade 2 R and O).

This guide will focus on how to create ideal start with the greatest accuracy possible. The important stuff for this guide is in bold, if you want to hear the mathematical reasoning and stuff like that, then read the whole guide

New Terminology Used

Pool limit - the amount that you could spend where you would break even on the mana increase of a mana pool.
Grade 3 mana gem - an grade 3 RO gem with 222 firing speed, 8.4 mana steal in a tower, and 30 splash in a tower.
Grade 6 mana gem - a grade 6 RO geme with 279 firing speed, 8.4 mana steal in a tower, and 30 splash in a tower.

Before you use this guide, it is strongly recommended that you have read the thread on 10 billion points made easy. Unless you have actually made 10 billion points, that is. This guide assumes that you are using 7000 mana per grade 3 mana gem, which is very close to the minimum amount necessary. We are using the last track because it has the greatest opportunity for mana earning, with 103 traps, about 50% more than any other level. This directly affects mana pool multiplier.

Part I - Grade 3 track

The formula for pool limit is ( 20 * #mana gems (for trap cost) ^ 2 + 7000 * # mana gems ) / 20. This reduces to:

m^2 + 350 m
Where m is the number of gems.
If this number is less than the cost of pooling, then you should create a grade 3 mana gem. If this number is greater than the cost of pooling, then you should mana pool.

Step 0 - Do necessary gem transmutations to create what is described above.
Step 1 - Create a maxed mana gem. It should be 10.4 mana steal in trap with at least 30 splash radius in a tower, so it hits 6 targets. It should have a 222 firing speed.
Step 2 - It is best to mana pool until you have 6000 max mana - you get a 5% bonus on all your mana-earning. Why 6000? 7020 / 20 = 351, so we will pool until the mana cost of pooling exceeds 351. Create another mana gem.
Step 3 - Mana pool to 11,000 max mana - Pool limit = 704. Create another mana gem.
Step 4 - Mana pool to 16,000 max mana - Pool limit = 1059. Create another mana gem.
Step 5 - Pool to 19K - the cost is greater than 1416. Create another mana gem.
Step N in this part - Follow this pattern by pooling up to where the pool cost exceeds 350 * # + #^2 of mana gems. Anger more than is sustainable to ensure that all your traps are being used at all times. Continue this until all the traps are covered in Grade 3 RO gems (except for Grade 3s in the beginning). You should have 83,000 mana pool at the end of this.

Part II - The Grade 6 track

We will now upgrade these gems to grade 6. This allows the cap to go from 533 firing speed to 670. This is 25% more efficient than before for another 4000 mana per upgrade. What does this mean? Each mana pool is worth 5.25 current mana gems, because there are about 105 places on the track. And by doing this, each of these are worth 0.25 mana gems, so we pool until the cost of pooling exceeds 84,000. This means you should pool until 93,000 mana.

Step II.N - Every upgrade adds 0.25 gems to the stakes, so this becomes 84,000 for pool limit, then 84,200 as you add another, then 84,400, and so on. .... You will be upgrading your gems multiple times before you mana pool. Keep track of the number of upgraded gems and your pool limit. You continue this until you have completely covered the map with Grade 6 mana gems. Pool limit for step 2 = 84,000 + 200*# of gems already upgraded to 670 firing speed

Build and place 3 R/B gems at the very start of the track to condense the monsters further and allow more monsters to be on the track. You might consider using prismatics (beware of lag) and purple gems in the beginning because of armor, but after a certain point, the extra damage is trivial and not worth the mana.

Part III - the maximum revenue

Now, we will begin replacing the traps with ROL (Red-Orange-Lime) gems. Why? Because this is the gem combination which ensures the maximum amount of mana steal. You may have heard of the ROY gem. How can this be better? It has 670 firing speed and a 70% chance of double hit, so it averages 1,139 hits. This is better than 1071 hits for a ROY gem. Furthermore, it can be constructed easier, because I have problems with the firing speed of ROY gems. It is not as mana efficient for construction as the grade 6 RO gem, but since we have run out of traps, this is the most efficient method of producing mana. It steals more than 45% more mana than the normal RO gem.

We will be doing two things here: We will be creating grade 10 ROL gems - these require 16 Grade 6 gems each, and with gem transmutations, that means a little less than 6000 per gem, so that's about 95,000 per gem for 45% more effectiveness of 1 gem.

The new pool limit formula: [ (number of mana gems RO)+(number of mana gems ROL)*1.447] * 95,000 (this is the cost) / [20 (mana pool increases the mana gain by 5%) * 0.447 (for the increased effectiveness) ].

Part IV. - Finally starting to build firepower

Here, we want to things - to build up your firepower, and also to build your multiplier.
Build a grade 15 prismatic gem. Drag a Red on it. Drag a Cyan on it. Drag a Lime on it. Drag a Yellow on it. You now have a correctly aligned prismatic. I would recommend building firepower with two super-gems. This allows you to alternate between building firepower with one and then the other.

Combine with multiple grade 6s. When you combine, drag it ONTO another gem, so it doesn't become misaligned. Each combination increases the firepower by 40-55. A whole sheet averages about 2000 max damage increased.

Part V - Placing your prismatics

Put your prismatics in a trap. This sounds stupid, but it actually deals 5x more damage in a trap. A few reasons why.
In a trap, you can hit up to 5 targets at once.
Triple damage is not shown, but you can have about a 30% chance of triple damage in a trap. It has been tested.
More range and firing speed means that you can fire about 3.5 x faster than in a tower.
A 66% chance of double hit means this is firing even faster. Note: there has been speculation that there is a possibility of more than one extra hit.

Before you put your gem in a trap, it is necessary to hue the gems - this allows the maximum damage to be dealt. On the last level, the optimum hue is 195, which can be achieved by dragging this gem onto several Cyan gems. Hue is dependent on the gems added. The Hue # corresponds to a number of degrees on a circle, ranging from 0 to 360 and once it reaches 360 it starts over at 0 (so a hue of 0 and a hue of 355 are very, very close together), and when two gems are combined, the Hue is the bisector of the angle created by the two radii of the original gems.

Part VI - On angering

Angering actually relates to the Pascal triangle, but that's completely different. Basically, it is a multinomial equation and an infinite product. It affects Hp geometrically:
(1+grade angered * 0.05)(current hp of wave)
It affects armor geometrically as well:
Initial armor * 1.25^#of angers OR Armor + 2. Whichever is higher
It affects Score in a infinite product (which can be reduced to a multinomial equation)
(1 + 0.15*grade) / (# of anger already done + 1).

If you anger with grade 7 or higher, then the score will keep on increasing by more than it increased the last time - any less, and it will have a limit. Keep this in mind when angering. Anger with as high grade gems as you can as long as you can be consistent - Angering with a Grade 15 and then Grade 8s is not as good as angering with Grade 11s consistently.

How much should I anger?

All that matters with angering is the Monster Health.
Your limit / wave is this: Each wave takes about 60 seconds, so using this, then ...
Total monster health capacity for a wave = 2,600 * active prismatic tower damage or 13,000 * active prismatic trap damage. This assumes that there can only be one extra hit. If there can be multiple extra hits, then the number is 3,000 * active prismatic tower damage or 15,000 * trap.
I anger boss waves to about 3-5x this, The wave preceding and 2 waves following this, I anger to half capacity.
Armored waves get 1.5 x this.
Normal waves get 1x this.
Runner waves get 0.5x this.
Swarm waves get enough so that they get through your traps, so that's about 10 million health of the total wave.

It is important to adjust this - if you find that monsters are leaking once or twice only, you can risk a greater score and higher banishment by angering more. If the banishment costs are high, then it's probably better to back off by 1 or 2 gems (any more would make things too easy. Overangering is not a huge problem.

Please note that the advice to anger is way off in this guide. Angering is the most difficult aspect of the game because you can anger with any grade, with different results for each grade angered. Furthermore, angering needs to be done on a case-by-case basis. I suggest "superangering" boss waves that give tons of score, especially those with only 2 monsters (twin bosses). As a rule of thumb: you get the most score with the least initial monsters - runners, monsters, and armored waves are equal in the score they give for the given wave. However, a wave of 8 armored gives roughly twice as much score as a wave of 16.

Other threads to see:

All 229 waves
10 billion points made easy
All 512 waves [if you really want a challenge]

Unrelated footnote because textile broke
Here is a savefile where I maxed stats with just one map.
<gc0save>22694092v2u4ndXVIV6Mh5dwmfB_TIp8HWkJ8Xp2HHQc6cgDcMau4nfQhcdEvx7GgkLFJ14URavjul6hhuLiab0bDcVECwt5pY32S5zXeC4+4kggkpMHNv8buHFH0Nxo1ECaj7GWmXSe4k2dxZvn3L/Kk29V0V+ltPOQzrY_sspSc7tQAoSTb1A8xf8WFyn5XR3dS3PW0Kzy9bOHHvL9uxvfOQuHeb4Q6l8LJjBrRmuhDn2nky7vLZqMgrYwMRz/xa83mbdaiyAUzCNW7tHKpEMz19xh6Buuuqx3/jamD00QOfqbLbAwxK+5s6REvxCoyaBsQMGGmSFI2nmlZ+evTnwPjMImzYhL9Etu5MDpO1AqpixNPNg_SJPz5EhEi1ABj3hggyvjD1TiwVMujX/58aWvQ3nR2PgNwtka6V3pFXTxffAJbd3++nIWql1Cb7Mf5iZrtW9ZwrhIcVwWaJ7rhK8KyTCJUVBm8W8_TJPYFiMG+bXxT9lTg3BbdZOuCAKHHoQkh0zxvsyBr+DEO4ojeyBxwHfEcUFaWeJ8XBMp1LMtNjvYPNS8RZQzf/vnSfJSYAqQgzf/G6unFQIclFGLdur0d4XKgybhWjiEH2HZOcrAud0JC5KLxRteP4bW+RvO9lXsBtEcvsZKuC84hJy9he3mk_m+D9u+4Juz3kcHLi1DB31gOJNKzqLfwUFRBliZAK79xC4nq0IhfiEr3rzMDBEZvJrotwplmckkbnAFrhtln4sxu0dRBGrNMtpqrWSygcQAxR2mXWCev125Wb/_+v09MlMvEYpS52mZI5s3xqJBF8PVn+h0ij3/QzE9xrULcsfg73zI1za299Th/napUFODdkUx5B/6gE/1L55EmMVAnGdmqu76rBWOdBHyjRNfmVikjk_UsehFodrJzUi6fQnZtl9meMmm6LNFGeGkdDKOP+SQrX8qAC3D8DQ53_jyU8P/YfVWfi++KmuPPZ0Wiki+zykNbeSUIaNULR0H91DyLZQFcwXNgH1sRR1SgC7Ow7LXhQwM_j7Hits1Xr1qVPlxhBWThdrehxmbLx1D7S9u4los1ruB2AS9uCw4scSLbeWwHEzg6StyS1Fy3zLqAG1pujrp/AlvlEh0ZRZpKAhNSerRLuytw/bp2d7LoGQnuhkjmtrhJ_w_WNldMLdNbsAe9O9NXw3equD8Ejcxy2q2Ze5Xo7gUQgS5WbC+BDwQjkrpgCz/Zzo7mB7zQTctgc/hNd1cnBQVl2CYvxZQoN6us_vWd4XY8CTad_PL+9FIIoscfaL9UhtztQ276mn6UBzJXzbr7BmQzFvQONkG5ac4+hagEGfmlRq7qxV6UO4YBW275D93i86_x4v_Ko6NkE/vbVQEJDbbuDs_cHzU4VnjSbzP7nOR7naCZZT5LJEmkM7h0uYqP6KIgJG6j25fLULcYNA7VvviSmPy/6k1khAacmdKVsSb/JuaqYzQXRTQzp+XSmcvEvN_5iAvLCos1sjsgcXQ+0__Jt2_1k+LfTd42bgBDDIuAy794nPjian+L3xgJWyPh0de8xgsAdi3u12r/pOAZQ3TQOo2mTZnMkDu1b2e3tF6+SOkOonVhiqllHAB1b75Rxll1mV+HuGlSUbICm9ZTEQsgPpVpgMeJhujGVOHh7wD_X7CJ9jb1Bu75pY2UQaM7CFZeKI9Ps+TXfM7F2hKPbfdZGNi2JfLV2c32SgCGGdLNEMgL9kZCK0fHGKCbwekFw63FS2Pzqcg82nf1F9tEkhJVt0HUrJ/h+kkbFZXfZLy5HS0YKduIY8iqov1mzkr8wZ4zMOBRZ6owTYZtfxs0cuJMOqUtW6g45T9D3pQ_gj/KoRwGmCegXptLOZVBjACsVmUFhqybKLVbTZmQd4z+2nkNP6HCEm4HXs3BGyYm9G99eu9ohvPB7o9MrnxMUSiCWvwHN5k+Df9Uzui7hAV1yIB335ndW5u8zsDAs2k1DMpFSRRxvWbW_w+j+CVSLXtq0A95K+rw3rS8h2kl2C7ruyFMZgb8F2VvZ7ggcLb4STQX7dYDZ838vZ5nOnnklGN8bFOEHBRJBx2F2QBCN+yRbHV+fbnKH2qZRKkHFkcaRai2DOycMs3IrdfG4_jaw6Mo_cxflkt6xZ+rqKfXS/kxojgrxwSN7ktUlge5PeikH/3069P8CHf9GZAv1wQ7HPdnbQ0chXZbS9StqqVcz1zlEsm8nFzjJ7FVBM6kSYhoMMy87WnIq9SInShpt0QncnUaHBtj68TLPM9/54ucpdnC22sL3otsNDJlWNL1uN/l2tQF5VPaKA9ZJSerRaEhKtMtR0OOQNrfrvBAnSnWdCF/z1kHoGv4NH2/25uYtI7fADuhsUCHVCrl1cd+YHByoLOGNLKizblOlWN7++jCQOERpWh7_5nfs54+VXc7RegRD4h6pBYMtkip9/tO/OQ+e2OqOAznmxoJypTqRJpbDL9Vj6eXWSKmjo29KTVQ_1jwhTkrD0X3MhkHhxWA1jqcKAYKB_qTj_gLJeD4poPATNh/p9xDwi2G_vHYBT/fkdo600ZGFLRO4tB69fxi69NZ0N1WZ7Rkr9nc8wxnXzdsQQetwKlZpyFQw6eg5mLpQjTAZj80gygohxqpfjYcf9F5t8tCVQIHOiERF2nSFbXBwNjA/gjr8SLlqXaHnZBPV8HKaO13fAFu1p_p/4perojvBcbmxSvIyIVmu0fAGfjYSsnQBzswEb1TAvv0NpBwi_Grv63sL1pebWQd2IeLMpKIGjXgGnCZ4xRzZpgwVwFoTSmwWURPITI3HKWfD28JMWyDIxivNQkZEJmiXbKLcPGUIfoQe7J/kECDwYQjHXHLwjB8CQOqbQNv81YXnh43e2qyeU7Wd8nmEO5kD_jyaHJXYVKf2lUeT1eYQUBnlUJqq9LtrFiDz_y+UApBZ1C_JpLlJ5gn9KZ6Kv1eNazmFG3Nr/v8oMgTeuNbsEFp4hvXH/qBaAz+usbcn6DS0jSpb4g6nJvn/VxZcKoxbIuNhu0lJGnUGZTUAkP2eWXt+uoJYQFzkQ/reIqghUW5E/w_3Ft_Rpn1XoPJv4rVNcWh+glvlKiHiiba14fWWTMUaRkuWJc1wyO18Iv2bJgRf1B+HhqJwYFwISWy7praRl1W63xOMH2xvFYLMxea86p0frP3hBsiGH1Av9iqgbjJfZOvJ0vW7nCeumgG3fzGxv8rzdNeOjsE6CXDzG08/IS32gkor24uKNTaPW0Smc31jRy+arKKTsCAku9u37WxS4pIIs/mxwosvTV5NMphu_/bf8mhhm3+D7CboJGX42d0b3jgUZoTu5KistDJdLG0S2E1Xjcf+bYNrmkqGk7ktvm_BjHg28t+gzLuMobASV6v/rzLQYI8FFbmXIGTypmCbV_aG3kceX1gHrAcmpG5Ot1gDSRhYCRVaNL1diXofucd4O8nvHKWzX_qm6ScW/w7jkkaq8VifWKQLBtNE0EPU_ad+n6XWu+kxEsoQGko+1oOrJI7+0/OygRHJnemyuS/HX+sIfP7E_XCZJSgo1paPrpXrZXqkpOQ06G+Njv2x+QKof_cAurT4gf8PXPAZkI8dU8mXGA0DqcsS3UCpCdFPXI0rVv+mSx8gRNHIxmQxCtnk8cBbihPZbcOglNzeBpI9l4EQY85sn+eH9p6cvnkm2Uvtlns7GVO/hykspB/diMEVk1d_KT0vY7/_2v9UzLqsoMfCxiPq5kiQ97Ug2cUkCa++cc5KcMNzDDL5mrQzThu4/C2jQ8ixNGho5H6px7Ga8LJSiddCJWf8BmBbCTkZOCn3R8eAxtRCz_E9LyIxWNtO2f_sL57ncPC6BGtdxdXsAl4hzBBEkDp6j/2z/uEkK_LXqT6Llk2t_N1NnamWlLmaYxAzrF2eExj5R/b7h5tvm8Ddi8ymvDOKrdPuW2vHP7b72dC_EfY1n+jI2u_LYIUjjsVyW392kE2dioAfJbUV2X6ImQNtpNjpnrdvTZ/c3qlUdy04LEKQEWMzNQ+YYJu_Erg0OYFl6j/JyKNTHnTVLns2mjTSIS7AEXNTijJl68TdFvBbyt1SElDJRaNKmOaRVsjEOt9TzoGByPxyU6fqUGfQ8UtyBH4JeFIYlEiGKfQjoxUxP3kT42foWVUcVDFflencqpU1OxchHRMelVikCwnNX4LZXsCSc3t0kAeoW6I3EaAWUd_mZlHvTpGGpSquq3o7t_XTJ3uvH02voqtZkdnrTS5dnO7YDe31rbJ1LFlAPLpIqMIkipxHAWasWxG_y7Dz9PTz1Lj6_fVIZ/JlqdXGCAxFC3PC3H7OL4ldzPPpFUQpPY8iXLfEsp_Qp5rJFix4DZJ2T_55FArJtkFghcxMs0JvRyWke4BM4+5yLZFQlsO29w1YwViySMmj2lHy3kReigGOy0qenQB5VY881HDcQggx2kiAtx7/sBamiINXHXLObF6YhpUFAj05yGI/kruMNmuHPX12YRI6vUErZ6HWaGXaDOag+3AbOZ8i3ziLXtLFrfU7LoBvD_XoXGcA1ecW1rXNXzdiwoFVAvlMFyvbSjSVXSfZFXHuIPirSh/a066EqCa_HsBb8MvsmmjVOpo8t4+q/b_+wx6_KZEMcUn5sGiwq2iiKGbst4bZ4mZ/8cE2qgBwYYpG3rD5zpJSHP7kFdG_41uV5XR+VEYCNye1heq2403e5n1x+KkFvvbAwaYzi/3fxNqh8mWyVLOTdgC+4uCNqLJwhszEv9gM21aTqQjsvGlYy_TvmT3ZCdyLvbSwyAGOo07vp56byWFHUvHmCpoMHHY1krtZG+fUxJr_ZE/Xg6NxVDRSN4ODE_V0gTVWhFboQcmPqBB4XM01kc+jdUsQIZN8Mnn8uzsTgweKvWw_l/h3SpZhXwD_aG4JR0+m+NNaqyj84hxtgTndIxo1zcT4Cdc0N6aK1bzPVDSZPTRfKRh1pYYcLW9Cjl9LqSuAgfvLfq1sTJW+UyqgcZYNHO8OdfrKfUIdeNoRNHaz1UlD59+ke3rV+Ev3JLPlxnFo3I_nNG2JrW5pe3bLQ7sbYlPxG9/3jolHqTIsrxlj7lxlL92mSK9qMe60_XQDkvzW/wzoAqLpUEhYdOJtrtH7fhlkHBA8z7R_+vRMSq3Ntb9vGdwahHjJ_kYZ2jBFUI8V9L3heiiMfpZawgd/kTjBnonlvyNn08l8baJkGiy0r4l_7gcNyL/LMgW7QQ3qxyQWpoJzU+1eDvY/bJFBgloNNvOCh0hKuA8ve2IRt6buk9hbnuKZiYNMouLP_CGgnBdmJhbM5qnAdu6DpfP_02uwE2nN0omeGplIZhmAQ/f7nygvsKIZjKWQA_4DlGDbN2WB7xhmQNUhhGLHOa+5U2wCFHZ2mQlSg_h0IlOegxLdYOY3ZbSD8bRhSDKJZSJhTHGQX4NfAPUhuNT4zROIUta/wQiBLaoQrygA8955nS/aw4wyY6wGU6lxvkmv677cMKRR8YX30SgxYSFVtEaQQU6bHS+_jxXNfsT96EAnfDXMgZzfItETtdOXvhZv3iHn8mrfE_Pn9o0cUZFpp9nzQ5xZvgy_chy3Y3oGAPG53SoYmmdZMMBgDjRNbg/q1IaRGEVy95H6YJpI3QUSquCcwnOb1Kn8p3_7DlOGCH8Jq/5E3J1_89LjGIbjLNnKQbZHTp9K5kN6Q13J_P3VdBEGKXHar/7Kjigmovwnnu8Wu/HrnDWc38yWPIF/0L__ReH2JtjVs7bdjoYfhro9jLqK/CsCxtj_Bil9TuIpgfGN147E9mZNE+0o5zpooiL6spQ7cbcd74f6gN1Mb6Ckb_3du98ROiSoT7OsrnjA0+_yyHKzfxRbNrHmWYtZ+bsI9zaI/+WrTEqSqS4fzzlU4QTPosKhqC1oWB+T_x/PMl3cYxozJgH7sbPZm2x1_+wj3NeKSS2OaU1UeUO+P+sOubEbXNJ/fB27Ohpia7LKNd3RwP9qH+WFY7qDupIpcdOCYADB_7_ROjBOaM_GvW5fGSm7UUNFD5mSyRBX7n+T6B1BQ3P/G29YtiX1PmLAP1Ju6VpQTmSNtqJHD83Nk7vGlENIpcoAMxVMDzMjMmTuYY9DdrBLHQGuR73M_fMquubMGBwu/rkGaorl2UN1Sc7AO5Z7/zrCyT_BaAR2qKbp80/zR1XM6VwEwaeP8QtX2PyRTizD1DXbVbDmmm8Rn8f_lyyN5xou/xR3pdC8Dk7VXePUabMhrkE3CgCZAHwsIhXHWy4aqIwO0Lt/uq_naKJR8lmspzU4gIvaJb0EP7J_sUwbQ/Q4hLiYKqMde5+yQY_a1RADRScqAGqow0Nn+M2mlGZk9a93XUN64GQ6c6QwhKNBI2URm7q6XIcrO7NftxFMURYQIdTXtZDRtiFbmBXsOg/3vzPw71O/_Xrz3WnIfPV_jmo0nFyJrTsqgjBcutHR8Bhb_x6APzzuwaCj2oGG2Ar8je4GKk/zaNU33vUCYdU5IeTm37AX25OEVygqsFL6uD4fX3sRQMLSJY1L9SClEgKfEn6k8XZ3gIa5CIXFNiHFQ0WnMvgETOd9YkemtPpOPXtHGpRY5SGIvA0cEl18C7JGx5PGTBoE_ujnAXi7H/OXRKtn4BXAgGC2McFFXOKlRn_RB4aMc4Z27ezpxyA/dbuzP40e/FbAjVHdn3d0F/jRE8CczVYMKQT7EpBzePL72wdu+mE2eG8k9XsAw0JL3oo2Gu2__aRsJL8vTTwAFvgSBYRxHdqEwmWHPhmRSUneL7_fzsKFwnlZalwnofw6TFLFwPh9HhkC7PfJnMc5+bW_0JvibdhWzg1IXR3dzcNvokP1X4ieYSQleqwUITqWRuqWi_UCtB7inHPbFNOfvmuNYrFp/gjjr30_DzZd3byiIRJG8O7qUOGWQ8kF4EkhMRJCPGelrtFxih2ew6fKEd3Z43HvcqryR/jsNECItpag32nU78Uz0zpd+4B7YfMDZIvcN2t2EDbfs3Wt8iE7fgFpunogV0dbIu4H+oivMWpHBUxIoaoALjrj6_LGp0M61VZq4s1xsBZoVsSumQ65LhdNoShWszIDSUC1iZAJf6V3SgXgGC+Yr5QEP2_TpvH8V0vsZBO9zJ2WouOFZ0t3ucPjsO0RwF_RxozZ4I7MhTmwpPjRZTfOKg/MjGZ46qaqW/xtiwr/whz2Lmv74IXAcPPQ+e2bITGtBapJvUNFs2sl_1+ed8D2qo1_cHw+dxXMBNNygjucDupIpP7dd0QmS2ivfP/+NkbAALPqCKXOFMPH7F7218Oz3Xzm7WvyM7Hx1ZcpVSuEW7zGJcnAszoS+li3h0XQM_Zdp2v5Zp3S9EJfpKc_u6qpeGyP/Nkw/c/8xKLqR6VYG2LCuxGMbh</gc0save>
 
Flag Post

Part 2 note – if you combine those grade 3 maxed mana gems with each other, you receive a grade 4 mana gem that is as good as maxed grade 6 mana gem, it will have firing speed of 279. This will save you mana so you won’t have to pay for those extra 50 grade 6 mana gems. The same approach could apply if you create a grade 6 R/O/L gem, then combine it with a grade 6 R/O gem, it might reach firing speed of 279 – though I haven’t tested this one. If not, it will still have firespeed above the normal limits for triples at grade 7, so you can use it as an intermediate solution.

Part 6 note about hue alignment – “by dragging this gem onto several CYAN gems”. However, if the prismatic’s hue is 130, you can use a single blue gem with hue of 280 to align your prismatic.

 
Flag Post

@einfach ROL (Red-Orange-Lime) gems the best manafarm in trap? Math is sweet, but are you sure? Did you do real test in game? Are you sure that L-part of gen can suck MANA? I don´t think so ….

I know definitely (becouse of testing): that in towers maxed OL -lime part of gem CAN´T suck mama. Why it should work in trap? …

Is need to make test.

Your advice to anger is wrong…
Right way is:
(in row according to the most anger to less)

bose wave
runners
armored
normal
swarm

RUNERS: - you shouldn´t summone them less then normal. Then can give you much more EXP… There is no need to scare of their speed. You have extremly manaincome for around 15rounds, which is really enought to shoot up them.

for 228 heavy loaded high gems anger.:

Your ratio is also wrong. (becouse of EXP) These ratio is too low. I was wathing HP and EXPs for many days.

swarns: anger for 10-20times less HP then bosses (just to keep your trap´s path full of monsters)
bosses: around 8-10times more HP then normal.
runners around 4-5 times more HP then normal
armored around 3-4 times more HP then normal

Bosses anger minimal 1-2 upper gems then anything else. (If you have time for it)
Swarns can be anger with lower gems then usually for save time.

 
Flag Post

L-part of a gem does help sucking mana. In fact, L-part in a tower does transfer all specials to the monster that suffers the chained hit, and in traps L does a chance at the NEXT hit (in fact it’s more than one hit, it has a chance to do another hit after the initial 70%), and with R giving the trap 6 SEPARATE hits, this gives a good chance to land 12-15 hits per trap shot. It’s somewhat random which monster will catch any extra hits, but law of big numbers says they will get average number of hits, and if those 70% chain hits are not capped (improbable, but could be), this average will be 1/(1-0.7)=3.33, and abit less if capped, and only 1.7 if I’m not right about recursion in chain hits in the trap. And I have a proof that R/O/L traps work better than plain R/O, I had my last epic battle using them, and they give more mana/frame than equal amount of R/O traps. Note that triples have lower specials percentage.

About runners – these buggers are the ones that can kill an unprepared adventurer, they have best chances to avoid stun/slow traps and amass banishment cost that’s enough to oneshot the wizard, that’s why there’s such caution.

 
Flag Post

MANAearn -— Well… I did in-game test.
It works L-part of gem in TRAP really can such muna, but can´t do it in TOWER. (quite strange for me)

It really seems to be ROL the best manafarm gem.
I tried to find out better combination even with prismatics and conclusion is that ROL should be the best.

in numbers: all maxed.. combination, Firespeed, manamulit. targets, result per minute
ROY: 1071 8.6 6 = 55 263
RO: 670 10.1 6 = 40 602
ROL: (670 8.6 6) +70% .. 34 572 + 70% = 58 772
prism.: (974 7.5 5) + 56% .. 36 525 + 56% = 56 979

1 O(t): = 12.1 RO = 16.5 ROY = 17.55 ROL

 
Flag Post

Wait … there’s a chance at NEXT hit?? If it’s a geometric sequence, the sum converges to – 2 1/3 hits per hit * 6 = 15 hits!!!
So if that’s true then Lime is a more important skill on a prismatic then yellow. If this is so – then all prismatics should have a Lime in the front.
And a superior Prismatic Mana advantage would be O-L-Y-R … instead of O-Y-L-R because the chain hit improves from 56 to 66%, especially if there is an opportunity for multiple hits.

The only part of the guide that needs improvement is the Angering. The rest is perfect – the mana pooling actually creates more mana than building another mana gem. Hmm… combining 2 grade 3 mana gems actually causes the extra supergemming of a grade 6 to cost about as much mana if not more.

 
Flag Post

@ honza
lime reaplies orange in towers.
Make a lime/orange dual super, and compare it with a purple/orange super tower if you don’t believe me.

 
Flag Post

@fractalman yes reaplies, but don´t earn any extra mana.

I tested combinations in towers: maxed: RO, O, OL, ORL …. significantly pure O has the highest manaincome and another colors didn´t make any changes.

@einfach. Prismatics gems are unusable for manaincome, even if then could be the best, becouse of lags.
YOR instead of OR did significant lag different… immagine prismatics.

 
Flag Post

I am currently testing this method, and unfortunately, I have messed up several times. The biggest mistake was forgetting to anger wave 4 – I also contaminated an RO gem as well, another 1000 mana loss, but that is more severe. I have underangered and I have 3 grade 3 mana gems by wave 7.5 and a mana pool of 16K. With perfect play, I can imagine one having a similar position by wave 6.5. To beat honza, I need to triple the number of mana gems every 2 rounds basically.

Edit 1: In about 1/3 of a round, I made another gem from scratch and pooled to 19K.
Edit 2: Now at 5 gems before round 8. 22K pool.
Edit 3: 7 gems right at the start of round 8. 28K pool and another error – created a red-orange-blue by accident. However, at this rate, I will create my gems and fill the track by round 10.
Edit 4: 9 gems at 8.2 or so. 32K pool. The major thing that slows me down is the smoke from transmuting.

 
Flag Post

I’m wondering, in part I step II, you say:

Why 5000?

Where did you get that number? Or is it supposed to be 6000?

 
Flag Post

Sorry, it was supposed to be 6000 – just I found a better estimate for the cost of a grade 3 mana gem and forgot to change it.

 
Flag Post

OK – Here is a screenshot of the strategy before Wave 9 – NO SPEEDHACK OR MACROS

I am definitely going to superanger wave 18 (twin giants) to have something along the lines of 1 mil. hp per monster (maybe more). Suggestions? Actually, what’s the most I can anger it so that I kill it before wave 42?

EDIT: Not shortly after this pic, I get a message saying “Flash Player Quit Unexpectedly” First time I ever got the message. Oh well, not wholly unexpected, considering that I was running two games at once, and one had had 300 1.2 mil. creatures and a prismatic mana farm :) for lag, although I’m kinda bummed because that was my best start ever.

Honza, you could do even better: USE THE STRATEGY!!!

 
Flag Post

@einfach all gems in traps, which we can see are RO 670 ? Anyway… very very NICE result !!! 9 wave… hmm

AND .. which one wave did you anger to let monsters to reach your wizard point?
and.. On which wave did you start to use R in traps and how many traps did you have in that time?

2 used traps next to wizadr point are gems for killing monsters? Or they´re just like garbage here? (fail during combination)

Oooo non of tower. !!! I tried it without any tower before and I wasn´t more succesfull then with them.

flasplayer—— I let autoupdate verstion of flashplayer and with CURRENT versions is 228challenge UNPLAYABLE !!!!
It crash to me in Opera in Chrome. and in FF during 30minutes consumed 1.2GB of RAM and still did go up. !!!

I downloaded 8-9th version of flash pl. and these are ok… I used them for 19days challenge. It consumed max 280M RAM.

…. Did you use strategy, which is up? Which one gem was you first 222FS? 6-7th?

 
Flag Post

Honza, all the gems are only 533 fs. It is better to do that rather than to upgrade one gem to 670 – more mana efficient according to the guide. It’s only a 25% difference, but a little more mana efficiency is huge when it comes to exponents, and the mana build is exponential growth.

Once you have 3 mana gems, you WANT leaks. After they leak, when they come back around, they will be subject to more targeting (the RO traps will be hitting 5-6 monsters in their area, whereas if there were no leaks, they would only be able to target 3 or so monsters in the area), so you will gain more mana.

I kind of messed up two times – with all the smoke from transmuting, a gem can look like it is Orange or Red, and with all the lag (from the prismatic farm in a different tab), I don’t bother to check. And when I combine them, I realize that it was actually blue or lime.

So … basically, I guess I should elaborate more on the actual strategy – what is up is just calculations – only math and #s for mana pooling and such… I followed this, but I start by having three original mana gems:

So to start, I build 3 traps right at the start, and I transmute my inventory so I have 3 grade 2 Red / Orange and 3 grade 3 Red / Orange. I combine each grade 3 Red with a grade 3 Orange. I select a particular grade 4 combo and combine with 2 more reds (to get 5 max targets) and 2 oranges. I put this one in the third trap. So we have three traps that are filled with grade 4 RO. I use the rest of the mana to create grade 1s and anger wave 1 4 times (you will have to wait a bit for you to get enough mana for the last grade 1).

Make sure you anger each wave 4-5 times, then 6 times until you have a grade 3 mana gem (533 fs in a trap), at which point you will kick it up a notch, and start angering 8-9 times. Don’t be conservative about your angering – just don’t under-anger. If you’re at wave 4-5 or so, you’re right where I was. Then, you get another mana gem … 11-12 angers. Another 13-14 angers – now, you can’t kill any (armor is too high), which is good because your mana income will keep rising as you get more and more monsters. Seriously no need to worry about banishment – your mana income is too high. From this point, you can max all your traps with 533 fs within 3-4 waves.

Follow the guide from here-on. Superanger wave 18 for score (just ‘cause – it’ll gather a ton of mana). You can create a prismatic once you have a farm and put it in a trap / tower to kill the monsters if you’re worried about banishment, but there really is no need.

 
Flag Post

When exactly did you start to improve 1st your gem to 533FS??? How many MAX mana and how many RO gems did you have?
Probably I did mistake, when I started with 5 O traps instead of 3RO. 5 - 15targest.
And second my mistake was UNDERarguing (which I immproved but I could even more)

 
Flag Post

I would start by getting up to maybe 500-600 mana, then buying gems, transmuting orange (and sometimes red), and then combining, about 2-3 monsters would leak through at a time – nothing major. I boosted this to 1000 after 3 gems, then 2000, then 6000-7000 because that is about the cost of a grade 3 gem (533 firing speed). So even though you had 25% more mana income in your picture because you had grade 6, that difference would have been compensated for by the start of wave 9.

Also, I got the gems up to 8.4 and 6 targets before moving on – it’s better that way.

It is only possible to get 1 gem up to 5 targets – this is your “favorite” gem, the others are at 4 (they target multiple targets even if it doesn’t say 5 targets or 6 targets).

I suppose it’s possible to get a complete mana farm by just about wave 10.

EDIT: I am working on maxing stats with just 1 map. I am at level 11 – ~8000 xp. and in an xp. stalemate. Replaying battles – about to get the 10xSudden Death amulet and level 12.

 
Flag Post

OK – I thought of one more thing:
Having Points spent in, for example, Orange Mastery, drastically improves the chance of getting orange gems. So if you have fewer than maximum points spent in Purple and Green Mastery (which are unnecessary skills), and slightly more but still fewer in Blue, Cyan, and Yellow (which you won’t need to get a ton of experience on 228), you could build your map with less transmuting required and build it faster. Possible experimentation on this would help. There is 1 major drawback – Purple and Cyan give you the most damage for your prismatics compared to the other gems, so 10-20% more gem combining may be necessary to boost the prismatic to get to however high damage you want.

EDIT: I am now level 14 with just 1 map – just a few thousand xp to go!

 
Flag Post

Even better
I let 3 initial RO gems be. I waited to 7K max MANA. Then I created 4th RO and started to maxed him. After that I maxed out rest 3 RO gems. and then just I created another one and so on.

Summoing from 5th wave 3grade gems, from 7th 5gr… for 8th .. 6grade.., 9th: with 7gra (13times usually)
I realized it should be even more !!

@einfach I didn´t use any of your math results just your advice. I found another mistake with your “optimalization”. You´ll get maxed-ORL on 10th grade…. I don´t know why, but I get maxed ROL on 9grade which turn into 11grade. ???

Means I´ll not wait for so big max mana to change RO to ROL. I agree is not good to start with this on 300K max MANA. (This max mana I suppose I´ll have when I´ll reach RO in all trap.)

Actually I don´t agree with your equils so much, becouse you don´t have “increasing” element here.. You just count how many cost something, but you FORGET to put into this one..
Easy explanation.. Earlier higher manaincome, better for future. (and everything has “lower” in this “future” loads for everything is “lower”, becouse of higher manaincome.

I´m still definitely persuaded that is easy to reach maxed: ROL, O(towers), 1M FP gem. before 30th wave run out.

10.1- 6targ. 533 everywhere

 
Flag Post

Can anybody count or find out how many “frames” is per minute of real game? speed: 1x (not 3x)
I´m counting many exact equils for 228. I can´t finish it without this variable.

EDIT: well is easy to know it.. :) just do test (cheat in game.) create maxed O put into tower 12*1.4*279*(multip let set up 1-and freeze it) for to easy see increasing of MANA. and let set up MAX mana for 100K

1.4multip. change for 1.. in skills for easier counting

12*1*279*1

Then argue all neccesary waves a lot to let the be invulnerable for don´t get any MANA for dieing of monsters…

Result of frames per minutes we can count from this manaincome.

12*1*279*1= 3348 per minute.. per minute..

Put this gem into tower and take stop clock empty MANA (by cheat) disable all efects… and let it run 1x speed. When 1st shoot reach monster let run stop clock.. after one minute hold the game…
It should increace mana by 3348… This is first look if we have everything OK….
If NOT somewhere is problem.

then empty MANA drag gem ot of tower… and then put back.. before gem start to shoot slowed game by engine cheat to neccesary speed for easy counting of “frames” …. and watch increasing of MANA
when mana increase by 1674 we can stop this counting of “frames” let say.. we have 200frames.. means 200 per 30sec..

then just multiply by 2 and here we have nummber of “frames” per 1 GAME-minute.

Or is possible to make quater of minute.

 
Flag Post

WHEN EXACTLY START TO TURN -RO- to ROL

@einfach I really don´t know how did you could when ROL should replace RO, but is completaly WRONG, becouse you forgot for nummber of manaearning gems. Same situation for RO 533 to 670…

There we have count when should be…. -- In these equils MUST be nummbers of RO gems.. (working gems) becouse everyone of them bring mana…

1 O(t): = 12.1 RO = 16.5 ROY = 17.55 ROL

ROL has for 45.0413% higher income for one gem. means 1.450413times more MANA per time per on gem.
100traps. + 0.45% higher income.. means… 1.0045times more. change ONE gem make…(mana pooling make 1.05times more)

Cost for turn RO to ROL- (10grade gem + combinateing + 1grade gems.) = 16times 6gems+rest= 3897*16+rest.62350 +let say 3000= 65 000MANA..

cost 65K ma. for 1.0045multi RO-ROL
turn one RO to ROL- increase multipl. 11.11times LOWER then MANApooling..

MEANS RO should be start to change for ROL when MANApooling cost 721 500MANA…

For me is definitely long time before 20th wave. means 1M FP before 30th. wave is so easy.

EDIT (mistake before) Every 3rd 6grade gem will LOST for change of color.
RO to ROL- cost: cca 85K mana

ROL should be start to change at 920K mana cost for manapooling.

 
Flag Post

Oh, oops… so the math is a little wrong, but the idea is solid.

You have more accurate numbers on this, I guess. I thought there were 2 more spaces on the track than there actually were, and this messed up my numbers a little. There are 103 spaces on the track, but for the number of working mana gems, you probably have 100, since that’s an easier number to work with.

Hmm…are you sure it only takes 65 000 Mana to create an ROL gem? For the creation, you might be able to achieve a faster firing speed by having more Orange than Red, because a grade 6 Orange will increase the firing speed twice as fast as a grade 6 Red.

About the “frames per minute” – I didn’t count every single one, but when I was Consistently mana pooling every 11 frames (couldn’t find one where it was going 10 or 12 frames to mana pool), I got to EXACTLY 80 mana pools in a round (not one more or less). So my estimate is about 870-900 per round. How long is the round time?

WOW! – Round … 8.5 for half a field – I bet you could get the entire field by wave 9! OK, well, the guide was only an estimate and an algorithm for calculating pool limit and thus perfect play for the first few round.

EDIT: Realizing that the numbers are off, I will be refining the guide soon…

 
Flag Post

Yes, I´m sure about 65K.. becouse I tried it in “tester” this´s create ROL FROM RO not create a new gem. …

All you have to do is create full bar of 6gr.L stones and take maxed RO combine him with 6grade… then 7 till 10.. then you have to just take a few 1grade orange gems and red gems and this´s it. nothing more.

Grade 10th is enought to has maxed ROL.

traps: 1RO + 4FP gems… this´s why just 100 (for easier counting… is 100 in real is 98 I thing.)
There is no neccesary to be so exact, becouse everyone will loose around MINIMAL 15% per EVERY manapooling.

118K. my currnet max mana. 10th wave on the way and “just” little bit more then 1 half of path RO 670bring mana and this gives just 10frames to manapooling.

I counted then neccesary quantity of “frames” for manapooling WILL only 5. !!! (at the begin) when I´ll has MAXED all manaearn gems everywhere. !!!! around 20wave. (probably)


look at this.. - everymanapooling wide next neccesary mana for pooling 1000mana, but multipl. increase by 5% (base 5% !! )
Can you see it?? Neccesary quantity of mana for manapooling HAS LINEAR requirements, BUT

income of mana HAS polynomial(parabolic) time realtion. !!

1K mana at 5M is nothing, but still we get 5% (base)

means… This will couse that longer and higher maxmana you will get, less “frames” will neccesary to next manapooling !! Lol
= more and more EXP.

 
Flag Post

Grade 10 requires 16 grade 6 gems.

But you’re forgetting about gem transmutation! This is 66% as efficient as creating raw 6’s, so that turns ~3800 per gem to 5700. So about 95,000 mana for maxed ROL – which is still 2x as mana efficient as creating grade 11s.

 
Flag Post

I didn´t forget for anything…. just read: my: WHEN EXACTLY START TO TURN RO to ROL

Read what I explaned and describe exactly… Overthere is guide to get ROL maxed during game.

I wrote 16times 6grade gem for 10grade gem… cost of transmutation?? — I didn´t forget. Is there: +let say 3000: this is for cost of trasnsmutaion + 1gr gems..

Don´t forget we allready have maxed RO… This´s why. in real is less then 3000, but I use this nummber for easier counting.

Definitely it cost very near to 65K mana.

 
Flag Post

No – what I mean is that – ok. you need 16 grade 6 R, O, or L gems (and some grade 1s) to create an ROL.

But, when you create them, you won’t get all ROL, because about 62.5% of the gems will be a different color. For these 62.5%, we will transmute, which is:
1 O,R,L + 3 off color = 3 O,R,L
So, really, we are creating 2 O,R,L for 3 off-color, making the process 2/3 as efficient as creating raw 6s – we will need 131.25% * 16 * 3897 (the cost per grade 6) = 81,837 mana.

And there are 15 combinations required to create a grade 10. Plus some more for the grade 1s, which cost 150 to create and combine (plus-or-minus 1 mana and taking into account transmuting). So this is maybe 83-85K mana per maxed ROL.


So, about the linear / quadratic. because the cost of mana pooling increases, the cumulative increase becomes a quadratic function, so this balances – you do not get more and more mana pooling as you progress. Because the cumulative mana pool cost is a quadratic f(x). The lim as the number of pools approaches infinity of the amount of base mana needed to be gained is 19,000. This means that if you have ROL working to max capacity, you could get ~309 mana pools per minute. So the estimated maximum mana one could possibly have in the 228 wave challenge is 55 million. So, realistically, I would estimate that 30-40 mil. is possible.