Dice roller - Java-Gaming.org

Roquen

JGO Ninja

Medals: 66

Dice roller

« Posted 2012-07-28 15:04:13 »

Everyone's first program:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357

import java.util.concurrent.atomic.AtomicLong;


/**
 * <b>Coding comments</b>
 * <p>
 * Threading. All methods are reentrant.  Technically a given instance is
 * not thread safe. From a practical standpoint, calling from multiple
 * threads will simply make the instance non-deterministic. Heavy access
 * from multiple threads is undesirable as each call will cause cache
 * line invalidations in all other actively accessing threads (assuming
 * on different cores). Specifically it's possible for the sequence to
 * appear to rewind some number of positions.
 */

public class Dice
{ 
  private static AtomicLong mix = new AtomicLong();
  
  private int seed;
 
  // these are proven reasonable number from L'Ecuyer
  private static final int RNG_M = 0xac549d55;
  private static final int RNG_A = 54321;
  
  public Dice(int seed)
  {
    this.seed = seed;
  }
  
  public Dice()
  {
    this((int)(mix.getAndDecrement() ^ System.nanoTime()));
  }
 
  /**
   * Simple 32-bit LCG.  Roughly the top 14-16 bits are
   * reasonable.
   */
  private final int rng()
  {
    int r = seed;
     
    r    *= RNG_M;
    r    += RNG_A;
    seed = r;
   
    return seed;
  }
 
  /**
   * almost uniform random on [0,0xFFFF]
   */
  private final int rng16()
  { 
    return rng() >>> 16;
  }
  
  /**
   * almost uniform random on [-,]
   */
  @SuppressWarnings("unused")
  private final int rng16s()
  {
    return rng() >> 16;
  }
  
  // 32-bit random number
  private final int rng32()
  {
    int r;
    
    r  = (rng() >>> 16);
    r |= (rng() & 0xFFFF0000);
    
    return r;
  }
  
  
  /**
   * Generates a (nearly) uniform random number on the
   * range [0, range).
   * <p>
   */
  private final int rng(int range)
  {
    return ((rng()>>>15) * range) >>> 17;
  }
  
  /**
   * 
   */
  public final void setSeed(int seed)
  {
    this.seed = seed;
  }
  
  public final int getSeed()
  {
    return seed;
  }
  
  
  /**
   * Simulates flipping 'n' coins.
   * <p>
   * The legal range of 'n' is [1, 16]
   */
  public final int flip16(int n)
  {
    return Integer.bitCount(rng() >>> (32-n));
  }
  
  /**
   * Simulates flipping 'n' coins.
   * <p>
   * The legal range of 'n' is [1, 32].
   */
  public final int flip32(int n)
  {
    return Integer.bitCount(rng32() >>> (32-n));
  }
  
  /**
   * Simulates flipping 16 coins.
   * <p>
   * Returns a random number on the range [0, 16].
   * <p>
   * @see #flip16(int)
   * @see #flip32()
   */
  public final int flip16()
  {
    return Integer.bitCount(rng16());
  }
  
  /**
   * Simulates flipping 32 coins.
   * <p>
   * Returns a random number on the range [0, 32]
   */
  
  public final int flip32()
  {
    return Integer.bitCount(rng32());
  }
  
  /** Flips a coin. Result is 1 or 2. */
  public final int d2()
  {
    return 1 + (rng() >>> 31);
  }
  
  /** Simulates the sum of <i>n</i> coin tosses. */
  public final int d2(int n)
  {
    int r = n;

    while (n >= 32) {
      r += flip32();
      n -= 32;
    }
    
    if (n >= 16) {
      r += flip16();
      n -= 16;
    }
    
    if (n != 0)
      r += flip16(n);
    
    return r;
  }
  
  /** Simulates a roll of a four sided die. */
  public final int d4()
  {
    return  1 + (rng() >>> 30);
  }
  
  /** Simulates the sum of <i>n</i> four sided dice rolls.*/
  public final int d4(int n)
  {
    int r = n;

    for(int i=0; i<n; i++)
      r += (rng() >>> 30);

    return r;
  }
  
  /** Simulates a roll of a six sided die. */
  public final int d6()
  {
    // 6 = 3*2 = (2+1)*2
    int r = rng16();
    
    r = ((r<<1)+r) >>> 15;
    
    return r+1;
  }
  
  /** Simulates the sum of <i>n</i> six sided dice rolls.*/
  public final int d6(int n)
  {
    int r = 0;

    for(int i=0; i<n; i++)
      r += d6();
    
    return r;
  }
  
  /** Simulates a roll of an eight sided die. */
  public final int d8()
  {
    return 1 + (rng() >>> 29);
  }
  
  /** Simulates the sum of <i>n</i> eight sided dice rolls.*/
  public final int d8(int n)
  {
    int r = n;

    for(int i=0; i<n; i++)
      r += (rng() >>> 29);
    
    return r;
  }
  
  /** Simulates a roll of a ten sided die. */
  public final int d10()
  {
    // 10 = 5*2 = (4+1)*2
    int r = rng16();
    
    r = ((r<<2)+r) >>> 15;
    
    return r+1;
  }
  
  /** Simulates the sum of <i>n</i> ten sided dice rolls.*/
  public final int d10(int n)
  {
    int r = 0;

    for(int i=0; i<n; i++)
      r += d10();
    
    return r;
  }
  
  /** Simulates a roll of a twelve sided die. */
  public final int d12()
  {
    // 12 = 3*4 = (2+1)*4
    int r = rng16();
    
    r = ((r<<1)+r) >>> 14;
    
    return r+1;
  }
  
  /** Simulates the sum of <i>n</i> twelve sided dice rolls.*/
  public final int d12(int n)
  {
    int r = 0;

    for(int i=0; i<n; i++)
      r += d12();
    
    return r;
  }
  
  /** Simulates a roll of a twenty sided die. */
  public final int d20()
  {
    // 20 = 5*4 = (4+1)*4
    int r = rng16();
    
    r = ((r<<2)+r) >>> 14;
    
    return r+1;
  }
  
  /** Simulates the sum of <i>n</i> twenty sided dice rolls.*/
  public final int d20(int n)
  {
    int r = 0;

    for(int i=0; i<n; i++)
      r += d20();
    
    return r;
  }
  
  /** Simulates a roll of a one-hundred sided die. */
  public final int d100()
  {
    // 100 = 25*4 = (3*8+1)*4 
    int r = rng() >>> 14;
    int t = (r << 1)+r;
    
    r = ((t<<3)+r) >>> 16;
    
    return r+1;
  }
  
  /** Simulates the sum of <i>n</i> one-hundred sided dice rolls.*/
  public final int d100(int n)
  {
    int r = 0;

    for(int i=0; i<n; i++)
      r += d100();
    
    return r;
  }
  
  /**
   * Returns a random number on [0, 0xFFFF] with a
   * triangular distribution.
   */
  public final int triangle()
  {
    int r0 = rng16();
    int r1 = rng16();
    
    return (r0+r1+1)>>1;
  }
  
  /**
   * Returns a random number on [0, range) with a
   * triangular distribution.
   */
  public final int triangle(int range)
  {
    return (triangle() * range) >>> 16;
  }
  
  /**
   * Returns a random number with a triangle distribution.
   * <p>
   * [0, a+b)
   * <p>
   * a 'steps' of linear increase/decreasing on the edges
   */
  public final int triangle(int a, int b)
  {
    // complete comments
    // isn't it really a+b-1
    a = rng(a);
    b = rng(b);
    
    return a+b;
  }
}

Roquen

JGO Ninja

Medals: 66

Re: Dice roller

« Reply #2 - Posted 2012-07-28 20:52:48 »

This nerd shit is that most people get it horrible wrong so they change to using something like the Mersenne twister and their results are just as wrong but they're to ignorant to know better. Questions? BTW: All of computer science is nerd shit, so maybe you should take up something like "shin kicking" or something as a hobby.

Seriously this is about as fast as you're going to get in Java and still statistically correct. Fast for small number of dice per roll (say 10ish). It gets a little more complicated for rolling many dice assuming you still want to be statistically correct.

Roquen

JGO Ninja

Medals: 66

Re: Dice roller

« Reply #5 - Posted 2012-07-28 21:57:37 »

@Danny02: There's already a thread that covers some basics in performance tuning. Basically the core generator here and in Random the same family and perfectly fine for most AI usages, dice rolling, etc. What people screw up for dice rolling is that the distribution changes based on the number of dice rolled. Uniform for 1 die (equal probability). Triangular for 2 and each additional moves it closer and closer to gaussian.

@Rorkien: The first part of my reply was a "your mother is sooo stupid" kinda response...the "shin kicking" and seriously part were intended to give that away.